deepbox 0.1.0

package/dist/chunk-MLBMYKCG.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/ml/index.ts","../src/ml/_validation.ts","../src/ml/clustering/DBSCAN.ts","../src/ml/clustering/KMeans.ts","../src/ml/decomposition/index.ts","../src/ml/tree/DecisionTree.ts","../src/ml/ensemble/GradientBoosting.ts","../src/ml/linear/Lasso.ts","../src/ml/linear/LinearRegression.ts","../src/ml/linear/LogisticRegression.ts","../src/ml/linear/Ridge.ts","../src/ml/manifold/TSNE.ts","../src/ml/naive_bayes/index.ts","../src/ml/neighbors/index.ts","../src/ml/svm/SVM.ts","../src/ml/tree/RandomForest.ts"],"names":["invM","delta","ez","sumExp","mean"],"mappings":";;;;;AAAA,IAAA,UAAA,GAAA;AAAA,QAAA,CAAA,UAAA,EAAA;AAAA,EAAA,MAAA,EAAA,MAAA,MAAA;AAAA,EAAA,sBAAA,EAAA,MAAA,sBAAA;AAAA,EAAA,qBAAA,EAAA,MAAA,qBAAA;AAAA,EAAA,UAAA,EAAA,MAAA,UAAA;AAAA,EAAA,0BAAA,EAAA,MAAA,0BAAA;AAAA,EAAA,yBAAA,EAAA,MAAA,yBAAA;AAAA,EAAA,MAAA,EAAA,MAAA,MAAA;AAAA,EAAA,oBAAA,EAAA,MAAA,oBAAA;AAAA,EAAA,mBAAA,EAAA,MAAA,mBAAA;AAAA,EAAA,KAAA,EAAA,MAAA,KAAA;AAAA,EAAA,gBAAA,EAAA,MAAA,gBAAA;AAAA,EAAA,SAAA,EAAA,MAAA,SAAA;AAAA,EAAA,SAAA,EAAA,MAAA,SAAA;AAAA,EAAA,kBAAA,EAAA,MAAA,kBAAA;AAAA,EAAA,GAAA,EAAA,MAAA,GAAA;AAAA,EAAA,sBAAA,EAAA,MAAA,sBAAA;AAAA,EAAA,qBAAA,EAAA,MAAA,qBAAA;AAAA,EAAA,KAAA,EAAA,MAAA,KAAA;AAAA,EAAA,IAAA,EAAA,MAAA;AAAA,CAAA,CAAA;;;ACWO,SAAS,gBAAA,CAAiB,GAAW,IAAA,EAAoB;AAC9D,EAAA,IAAI,CAAC,YAAA,CAAa,CAAA,CAAE,KAAA,EAAO,CAAA,CAAE,OAAO,CAAA,EAAG;AACrC,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR,GAAG,IAAI,CAAA,qGAAA;AAAA,KACT;AAAA,EACF;AACF;AAkBO,SAAS,iBAAA,CAAkB,GAAW,CAAA,EAAiB;AAE5D,EAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,EACpE;AACA,EAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,EACpE;AACA,EAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,EAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAGvB,EAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,EAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,EAAA,IAAI,aAAa,CAAA,EAAG;AAClB,IAAA,MAAM,IAAI,oBAAoB,iCAAiC,CAAA;AAAA,EACjE;AACA,EAAA,IAAI,cAAc,CAAA,EAAG;AACnB,IAAA,MAAM,IAAI,oBAAoB,kCAAkC,CAAA;AAAA,EAClE;AAGA,EAAA,IAAI,QAAA,KAAa,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,EAAG;AAC3B,IAAA,MAAM,IAAI,UAAA;AAAA,MACR,gEAAgE,QAAQ,CAAA,aAAA,EAAgB,CAAA,CAAE,KAAA,CAAM,CAAC,CAAC,CAAA;AAAA,KACpG;AAAA,EACF;AAGA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,IAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,IAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,MAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,IAC3E;AAAA,EACF;AAGA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,IAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,IAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,MAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,IAC3E;AAAA,EACF;AACF;AAgBO,SAAS,8BAA8B,CAAA,EAAiB;AAE7D,EAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,EACpE;AACA,EAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAGvB,EAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,EAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,EAAA,IAAI,aAAa,CAAA,EAAG;AAClB,IAAA,MAAM,IAAI,oBAAoB,iCAAiC,CAAA;AAAA,EACjE;AACA,EAAA,IAAI,cAAc,CAAA,EAAG;AACnB,IAAA,MAAM,IAAI,oBAAoB,kCAAkC,CAAA;AAAA,EAClE;AAGA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,IAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,IAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,MAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,IAC3E;AAAA,EACF;AACF;AAkBO,SAAS,qBAAA,CACd,CAAA,EACA,iBAAA,EACA,SAAA,EACM;AAEN,EAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,EACpE;AACA,EAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAGvB,EAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,EAAA,IAAI,cAAc,iBAAA,EAAmB;AACnC,IAAA,MAAM,IAAI,UAAA;AAAA,MACR,CAAA,MAAA,EAAS,SAAS,CAAA,cAAA,EAAiB,SAAS,oBAAoB,iBAAiB,CAAA,SAAA;AAAA,KACnF;AAAA,EACF;AAGA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,IAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,IAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,MAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,IAC3E;AAAA,EACF;AACF;;;ACjIO,IAAM,SAAN,MAAkC;AAAA,EAC/B,GAAA;AAAA,EACA,UAAA;AAAA,EACA,MAAA;AAAA,EAEA,OAAA;AAAA,EACA,YAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAII,EAAC,EACL;AACA,IAAA,IAAA,CAAK,GAAA,GAAM,QAAQ,GAAA,IAAO,GAAA;AAC1B,IAAA,IAAA,CAAK,UAAA,GAAa,QAAQ,UAAA,IAAc,CAAA;AACxC,IAAA,IAAA,CAAK,MAAA,GAAS,QAAQ,MAAA,IAAU,WAAA;AAEhC,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,GAAG,CAAA,IAAK,IAAA,CAAK,OAAO,CAAA,EAAG;AAC/C,MAAA,MAAM,IAAI,qBAAA,CAAsB,iCAAA,EAAmC,KAAA,EAAO,KAAK,GAAG,CAAA;AAAA,IACpF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,UAAU,CAAA,IAAK,IAAA,CAAK,aAAa,CAAA,EAAG;AAC7D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,oCAAA;AAAA,QACA,YAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,IAAA,CAAK,MAAA,KAAW,WAAA,IAAe,IAAA,CAAK,WAAW,WAAA,EAAa;AAC9D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,yCAAA,CAAA;AAAA,QACA,QAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA,EAKQ,UAAA,CAAW,GAAa,CAAA,EAAsB;AACpD,IAAA,IAAI,IAAA,CAAK,WAAW,WAAA,EAAa;AAC/B,MAAA,IAAI,GAAA,GAAM,CAAA;AACV,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,QAAQ,CAAA,EAAA,EAAK;AACjC,QAAA,GAAA,IAAO,IAAA,CAAK,KAAK,CAAA,CAAE,CAAC,KAAK,CAAA,KAAM,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAE,CAAA;AACzC,QAAA,IAAI,GAAA,GAAM,IAAA,CAAK,GAAA,EAAK,OAAO,KAAA;AAAA,MAC7B;AACA,MAAA,OAAO,OAAO,IAAA,CAAK,GAAA;AAAA,IACrB;AAGA,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,GAAA,GAAM,IAAA,CAAK,GAAA;AAC9B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,QAAQ,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,QAAQ,CAAA,CAAE,CAAC,KAAK,CAAA,KAAM,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AACpC,MAAA,KAAA,IAAS,IAAA,GAAO,IAAA;AAChB,MAAA,IAAI,KAAA,GAAQ,OAAO,OAAO,KAAA;AAAA,IAC5B;AACA,IAAA,OAAO,KAAA,IAAS,KAAA;AAAA,EAClB;AAAA;AAAA;AAAA;AAAA,EAKQ,YAAA,CAAa,MAAkB,QAAA,EAA4B;AACjE,IAAA,MAAM,YAAsB,EAAC;AAC7B,IAAA,MAAM,KAAA,GAAQ,KAAK,QAAQ,CAAA;AAC3B,IAAA,IAAI,CAAC,OAAO,OAAO,SAAA;AAEnB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,QAAQ,CAAA,EAAA,EAAK;AACpC,MAAA,MAAM,KAAA,GAAQ,KAAK,CAAC,CAAA;AACpB,MAAA,IAAI,KAAA,IAAS,IAAA,CAAK,UAAA,CAAW,KAAA,EAAO,KAAK,CAAA,EAAG;AAC1C,QAAA,SAAA,CAAU,KAAK,CAAC,CAAA;AAAA,MAClB;AAAA,IACF;AAEA,IAAA,OAAO,SAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,GAAA,CAAI,GAAW,EAAA,EAAmB;AAChC,IAAA,6BAAA,CAA8B,CAAC,CAAA;AAE/B,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAGhC,IAAA,MAAM,OAAmB,EAAC;AAC1B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,IAAA,CAAK,KAAK,GAAG,CAAA;AAAA,IACf;AAGA,IAAA,MAAM,SAAmB,IAAI,KAAA,CAAM,QAAQ,CAAA,CAAE,KAAK,EAAE,CAAA;AACpD,IAAA,MAAM,cAAwB,EAAC;AAE/B,IAAA,IAAI,SAAA,GAAY,CAAA;AAEhB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AAEjC,MAAA,IAAI,MAAA,CAAO,CAAC,CAAA,KAAM,EAAA,EAAI;AAEtB,MAAA,MAAM,SAAA,GAAY,IAAA,CAAK,YAAA,CAAa,IAAA,EAAM,CAAC,CAAA;AAE3C,MAAA,IAAI,SAAA,CAAU,MAAA,GAAS,IAAA,CAAK,UAAA,EAAY;AAEtC,QAAA,MAAA,CAAO,CAAC,CAAA,GAAI,EAAA;AACZ,QAAA;AAAA,MACF;AAGA,MAAA,WAAA,CAAY,KAAK,CAAC,CAAA;AAClB,MAAA,MAAA,CAAO,CAAC,CAAA,GAAI,SAAA;AAGZ,MAAA,MAAM,OAAA,GAAU,IAAI,GAAA,CAAI,SAAS,CAAA;AAEjC,MAAA,KAAA,MAAW,KAAK,OAAA,EAAS;AAEvB,QAAA,IAAI,MAAA,CAAO,CAAC,CAAA,KAAM,EAAA,EAAI;AACpB,UAAA,MAAA,CAAO,CAAC,CAAA,GAAI,SAAA;AAAA,QACd;AAGA,QAAA,IAAI,MAAA,CAAO,CAAC,CAAA,KAAM,EAAA,EAAI;AACpB,UAAA,MAAA,CAAO,CAAC,CAAA,GAAI,SAAA;AAEZ,UAAA,MAAM,UAAA,GAAa,IAAA,CAAK,YAAA,CAAa,IAAA,EAAM,CAAC,CAAA;AAE5C,UAAA,IAAI,UAAA,CAAW,MAAA,IAAU,IAAA,CAAK,UAAA,EAAY;AACxC,YAAA,WAAA,CAAY,KAAK,CAAC,CAAA;AAElB,YAAA,KAAA,MAAW,KAAK,UAAA,EAAY;AAC1B,cAAA,IAAI,OAAO,CAAC,CAAA,KAAM,MAAM,MAAA,CAAO,CAAC,MAAM,EAAA,EAAI;AACxC,gBAAA,OAAA,CAAQ,IAAI,CAAC,CAAA;AAAA,cACf;AAAA,YACF;AAAA,UACF;AAAA,QACF;AAAA,MACF;AAEA,MAAA,SAAA,EAAA;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,UAAU,MAAA,CAAO,MAAA,EAAQ,EAAE,KAAA,EAAO,SAAS,CAAA;AAChD,IAAA,IAAA,CAAK,YAAA,GAAe,WAAA;AACpB,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,QAAQ,EAAA,EAAoB;AAC1B,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,UAAA,CAAW,GAAW,EAAA,EAAqB;AACzC,IAAA,IAAA,CAAK,IAAI,CAAC,CAAA;AACV,IAAA,IAAI,CAAC,KAAK,OAAA,EAAS;AACjB,MAAA,MAAM,IAAI,eAAe,mCAAmC,CAAA;AAAA,IAC9D;AACA,IAAA,OAAO,IAAA,CAAK,OAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,MAAA,GAAiB;AACnB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,OAAA,EAAS;AACjC,MAAA,MAAM,IAAI,eAAe,wCAAwC,CAAA;AAAA,IACnE;AACA,IAAA,OAAO,IAAA,CAAK,OAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAUA,IAAI,WAAA,GAAwB;AAC1B,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,YAAA,EAAc;AACtC,MAAA,MAAM,IAAI,eAAe,8CAA8C,CAAA;AAAA,IACzE;AACA,IAAA,OAAO,CAAC,GAAG,IAAA,CAAK,YAAY,CAAA;AAAA,EAC9B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,KAAK,IAAA,CAAK,GAAA;AAAA,MACV,YAAY,IAAA,CAAK,UAAA;AAAA,MACjB,QAAQ,IAAA,CAAK;AAAA,KACf;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,UAAU,MAAA,EAAuC;AAC/C,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,KAAK,KAAK,MAAA,CAAO,OAAA,CAAQ,MAAM,CAAA,EAAG;AACjD,MAAA,QAAQ,GAAA;AAAK,QACX,KAAK,KAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,KAAA,IAAS,CAAA,EAAG;AAC3C,YAAA,MAAM,IAAI,qBAAA,CAAsB,iBAAA,EAAmB,KAAA,EAAO,KAAK,CAAA;AAAA,UACjE;AACA,UAAA,IAAA,CAAK,GAAA,GAAM,KAAA;AACX,UAAA;AAAA,QACF,KAAK,YAAA;AACH,UAAA,IAAI,OAAO,UAAU,QAAA,IAAY,CAAC,OAAO,SAAA,CAAU,KAAK,CAAA,IAAK,KAAA,GAAQ,CAAA,EAAG;AACtE,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,oCAAA;AAAA,cACA,YAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,UAAA,GAAa,KAAA;AAClB,UAAA;AAAA,QACF,KAAK,QAAA;AACH,UAAA,IAAI,KAAA,KAAU,WAAA,IAAe,KAAA,KAAU,WAAA,EAAa;AAClD,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,yCAAA,CAAA;AAAA,cACA,QAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,MAAA,GAAS,KAAA;AACd,UAAA;AAAA,QACF;AACE,UAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,mBAAA,EAAsB,GAAG,CAAA,CAAA,EAAI,KAAK,KAAK,CAAA;AAAA;AAC3E,IACF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AACF;;;ACjRO,IAAM,SAAN,MAAkC;AAAA,EAC/B,SAAA;AAAA,EACA,OAAA;AAAA,EACA,GAAA;AAAA,EACA,IAAA;AAAA,EACA,WAAA;AAAA,EAEA,eAAA;AAAA,EACA,OAAA;AAAA,EACA,QAAA;AAAA,EACA,MAAA;AAAA,EACA,YAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYjB,WAAA,CACE,OAAA,GAMI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,SAAA,GAAY,QAAQ,SAAA,IAAa,CAAA;AACtC,IAAA,IAAA,CAAK,OAAA,GAAU,QAAQ,OAAA,IAAW,GAAA;AAClC,IAAA,IAAA,CAAK,GAAA,GAAM,QAAQ,GAAA,IAAO,IAAA;AAC1B,IAAA,IAAA,CAAK,IAAA,GAAO,QAAQ,IAAA,IAAQ,UAAA;AAC5B,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AAEA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,SAAS,CAAA,IAAK,IAAA,CAAK,YAAY,CAAA,EAAG;AAC3D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,mCAAA;AAAA,QACA,WAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,OAAO,CAAA,IAAK,IAAA,CAAK,UAAU,CAAA,EAAG;AACvD,MAAA,MAAM,IAAI,qBAAA,CAAsB,iCAAA,EAAmC,SAAA,EAAW,KAAK,OAAO,CAAA;AAAA,IAC5F;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,GAAG,CAAA,IAAK,IAAA,CAAK,MAAM,CAAA,EAAG;AAC9C,MAAA,MAAM,IAAI,qBAAA,CAAsB,kCAAA,EAAoC,KAAA,EAAO,KAAK,GAAG,CAAA;AAAA,IACrF;AACA,IAAA,IAAI,IAAA,CAAK,IAAA,KAAS,QAAA,IAAY,IAAA,CAAK,SAAS,UAAA,EAAY;AACtD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,IAAA,CAAK,IAAI,CAAC,CAAA,CAAA;AAAA,QAClE,MAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,IAAa,CAAC,OAAO,QAAA,CAAS,OAAA,CAAQ,WAAW,CAAA,EAAG;AAC9E,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,OAAA,CAAQ,WAAW,CAAC,CAAA,CAAA;AAAA,QAC5E,aAAA;AAAA,QACA,OAAA,CAAQ;AAAA,OACV;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,GAAA,CAAI,GAAW,EAAA,EAAmB;AAChC,IAAA,6BAAA,CAA8B,CAAC,CAAA;AAE/B,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAA,CAAK,YAAA,GAAe,SAAA;AAEpB,IAAA,IAAI,QAAA,GAAW,KAAK,SAAA,EAAW;AAC7B,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,UAAA,EAAa,QAAQ,CAAA,yBAAA,EAA4B,IAAA,CAAK,SAAS,CAAA,CAAA;AAAA,QAC/D,WAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAGA,IAAA,IAAI,SAAA,GAAY,IAAA,CAAK,mBAAA,CAAoB,CAAC,CAAA;AAE1C,IAAA,IAAI,cAAc,MAAA,CAAO,iBAAA;AAEzB,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,IAAA,CAAK,SAAS,IAAA,EAAA,EAAQ;AAE9C,MAAA,MAAM,MAAA,GAAS,IAAA,CAAK,cAAA,CAAe,CAAA,EAAG,SAAS,CAAA;AAG/C,MAAA,MAAM,eAA2B,EAAC;AAClC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,WAAW,CAAA,EAAA,EAAK;AACvC,QAAA,MAAM,gBAA4B,EAAC;AACnC,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,IAAI,MAAA,CAAO,OAAO,IAAA,CAAK,MAAA,CAAO,SAAS,CAAC,CAAC,MAAM,CAAA,EAAG;AAChD,YAAA,MAAM,QAAkB,EAAC;AACzB,YAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,cAAA,KAAA,CAAM,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,YACzD;AACA,YAAA,aAAA,CAAc,KAAK,KAAK,CAAA;AAAA,UAC1B;AAAA,QACF;AAEA,QAAA,IAAI,aAAA,CAAc,SAAS,CAAA,EAAG;AAC5B,UAAA,MAAM,WAAqB,EAAC;AAC5B,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,IAAI,GAAA,GAAM,CAAA;AACV,YAAA,KAAA,MAAW,SAAS,aAAA,EAAe;AACjC,cAAA,GAAA,IAAO,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAAA,YACrB;AACA,YAAA,QAAA,CAAS,IAAA,CAAK,GAAA,GAAM,aAAA,CAAc,MAAM,CAAA;AAAA,UAC1C;AACA,UAAA,YAAA,CAAa,KAAK,QAAQ,CAAA;AAAA,QAC5B,CAAA,MAAO;AAEL,UAAA,MAAM,cAAwB,EAAC;AAC/B,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,WAAA,CAAY,IAAA,CAAK,MAAA,CAAO,SAAA,CAAU,IAAA,CAAK,SAAA,CAAU,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,UAC/E;AACA,UAAA,YAAA,CAAa,KAAK,WAAW,CAAA;AAAA,QAC/B;AAAA,MACF;AAEA,MAAA,SAAA,GAAY,OAAO,YAAY,CAAA;AAG/B,MAAA,MAAM,OAAA,GAAU,IAAA,CAAK,gBAAA,CAAiB,CAAA,EAAG,WAAW,MAAM,CAAA;AAG1D,MAAA,IAAI,KAAK,GAAA,CAAI,WAAA,GAAc,OAAO,CAAA,GAAI,KAAK,GAAA,EAAK;AAC9C,QAAA,IAAA,CAAK,SAAS,IAAA,GAAO,CAAA;AACrB,QAAA;AAAA,MACF;AAEA,MAAA,WAAA,GAAc,OAAA;AACd,MAAA,IAAA,CAAK,SAAS,IAAA,GAAO,CAAA;AAAA,IACvB;AAEA,IAAA,IAAA,CAAK,eAAA,GAAkB,SAAA;AACvB,IAAA,IAAA,CAAK,OAAA,GAAU,IAAA,CAAK,cAAA,CAAe,CAAA,EAAG,SAAS,CAAA;AAC/C,IAAA,IAAA,CAAK,WAAW,IAAA,CAAK,gBAAA,CAAiB,CAAA,EAAG,SAAA,EAAW,KAAK,OAAO,CAAA;AAChE,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,eAAA,EAAiB;AACzC,MAAA,MAAM,IAAI,eAAe,yCAAyC,CAAA;AAAA,IACpE;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,QAAQ,CAAA;AAEzD,IAAA,OAAO,IAAA,CAAK,cAAA,CAAe,CAAA,EAAG,IAAA,CAAK,eAAe,CAAA;AAAA,EACpD;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,UAAA,CAAW,GAAW,EAAA,EAAqB;AACzC,IAAA,IAAA,CAAK,IAAI,CAAC,CAAA;AACV,IAAA,IAAI,CAAC,KAAK,OAAA,EAAS;AACjB,MAAA,MAAM,IAAI,eAAe,mCAAmC,CAAA;AAAA,IAC9D;AACA,IAAA,OAAO,IAAA,CAAK,OAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA,EAKQ,oBAAoB,CAAA,EAAmB;AAC7C,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAI,IAAA,CAAK,SAAS,QAAA,EAAU;AAE1B,MAAA,MAAM,OAAA,uBAAc,GAAA,EAAY;AAChC,MAAA,MAAM,GAAA,GAAM,KAAK,SAAA,EAAU;AAE3B,MAAA,OAAO,OAAA,CAAQ,IAAA,GAAO,IAAA,CAAK,SAAA,EAAW;AACpC,QAAA,OAAA,CAAQ,IAAI,IAAA,CAAK,KAAA,CAAM,GAAA,EAAI,GAAI,QAAQ,CAAC,CAAA;AAAA,MAC1C;AAEA,MAAA,MAAM,YAAwB,EAAC;AAC/B,MAAA,KAAA,MAAW,OAAO,OAAA,EAAS;AACzB,QAAA,MAAM,WAAqB,EAAC;AAC5B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,QAAA,CAAS,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,GAAA,GAAM,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,QAC9D;AACA,QAAA,SAAA,CAAU,KAAK,QAAQ,CAAA;AAAA,MACzB;AAEA,MAAA,OAAO,OAAO,SAAS,CAAA;AAAA,IACzB,CAAA,MAAO;AAEL,MAAA,MAAM,GAAA,GAAM,KAAK,SAAA,EAAU;AAC3B,MAAA,MAAM,YAAwB,EAAC;AAC/B,MAAA,MAAM,YAAY,IAAI,YAAA,CAAa,QAAQ,CAAA,CAAE,KAAK,QAAQ,CAAA;AAG1D,MAAA,MAAM,QAAA,GAAW,IAAA,CAAK,KAAA,CAAM,GAAA,KAAQ,QAAQ,CAAA;AAC5C,MAAA,MAAM,gBAA0B,EAAC;AACjC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,aAAA,CAAc,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,QAAA,GAAW,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACxE;AACA,MAAA,SAAA,CAAU,KAAK,aAAa,CAAA;AAG5B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,IAAI,IAAA,GAAO,CAAA;AACX,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,MAAM,IAAA,GAAO,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA,IAAK,aAAA,CAAc,CAAC,CAAA,IAAK,CAAA,CAAA;AACjF,UAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,QACjB;AACA,QAAA,SAAA,CAAU,CAAC,CAAA,GAAI,IAAA;AAAA,MACjB;AAGA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,WAAW,CAAA,EAAA,EAAK;AAEvC,QAAA,MAAM,SAAA,GAAY,UAAU,MAAA,CAAO,CAAC,GAAG,CAAA,KAAM,CAAA,GAAI,GAAG,CAAC,CAAA;AACrD,QAAA,IAAI,CAAA,GAAI,KAAI,GAAI,SAAA;AAChB,QAAA,IAAI,OAAA,GAAU,CAAA;AAEd,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,CAAA,IAAK,SAAA,CAAU,CAAC,CAAA,IAAK,CAAA;AACrB,UAAA,IAAI,KAAK,CAAA,EAAG;AACV,YAAA,OAAA,GAAU,CAAA;AACV,YAAA;AAAA,UACF;AAAA,QACF;AAEA,QAAA,MAAM,eAAyB,EAAC;AAChC,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,YAAA,CAAa,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,OAAA,GAAU,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,QACtE;AACA,QAAA,SAAA,CAAU,KAAK,YAAY,CAAA;AAG3B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,IAAI,IAAA,GAAO,CAAA;AACX,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,MAAM,IAAA,GAAO,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA,IAAK,YAAA,CAAa,CAAC,CAAA,IAAK,CAAA,CAAA;AAChF,YAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,UACjB;AACA,UAAA,IAAI,IAAA,IAAQ,SAAA,CAAU,CAAC,CAAA,IAAK,QAAA,CAAA,EAAW;AACrC,YAAA,SAAA,CAAU,CAAC,CAAA,GAAI,IAAA;AAAA,UACjB;AAAA,QACF;AAAA,MACF;AAEA,MAAA,OAAO,OAAO,SAAS,CAAA;AAAA,IACzB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA,EAKQ,cAAA,CAAe,GAAW,SAAA,EAA2B;AAC3D,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,SAAmB,EAAC;AAE1B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,IAAI,UAAU,MAAA,CAAO,iBAAA;AACrB,MAAA,IAAI,QAAA,GAAW,CAAA;AAEf,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,WAAW,CAAA,EAAA,EAAK;AACvC,QAAA,IAAI,IAAA,GAAO,CAAA;AACX,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,MAAM,OACJ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,IAAI,SAAA,GAAY,CAAC,CAAC,CAAA,GAC3C,MAAA,CAAO,UAAU,IAAA,CAAK,SAAA,CAAU,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AAC7D,UAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,QACjB;AAEA,QAAA,IAAI,OAAO,OAAA,EAAS;AAClB,UAAA,OAAA,GAAU,IAAA;AACV,UAAA,QAAA,GAAW,CAAA;AAAA,QACb;AAAA,MACF;AAEA,MAAA,MAAA,CAAO,KAAK,QAAQ,CAAA;AAAA,IACtB;AAEA,IAAA,OAAO,MAAA,CAAO,MAAA,EAAQ,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EAC1C;AAAA;AAAA;AAAA;AAAA,EAKQ,gBAAA,CAAiB,CAAA,EAAW,SAAA,EAAmB,MAAA,EAAwB;AAC7E,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAI,OAAA,GAAU,CAAA;AAEd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,QAAQ,MAAA,CAAO,MAAA,CAAO,KAAK,MAAA,CAAO,MAAA,GAAS,CAAC,CAAC,CAAA;AACnD,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,OACJ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,IAAI,SAAA,GAAY,CAAC,CAAC,CAAA,GAC3C,MAAA,CAAO,UAAU,IAAA,CAAK,SAAA,CAAU,SAAS,KAAA,GAAQ,SAAA,GAAY,CAAC,CAAC,CAAA;AACjE,QAAA,OAAA,IAAW,IAAA,GAAO,IAAA;AAAA,MACpB;AAAA,IACF;AAEA,IAAA,OAAO,OAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKQ,SAAA,GAA0B;AAChC,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,MAAA,IAAI,OAAO,IAAA,CAAK,WAAA;AAChB,MAAA,OAAO,MAAM;AACX,QAAA,IAAA,GAAA,CAAQ,IAAA,GAAO,OAAO,KAAA,IAAS,MAAA;AAC/B,QAAA,OAAO,IAAA,GAAO,MAAA;AAAA,MAChB,CAAA;AAAA,IACF;AACA,IAAA,OAAO,IAAA,CAAK,MAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,cAAA,GAAyB;AAC3B,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,eAAA,EAAiB;AACzC,MAAA,MAAM,IAAI,eAAe,iDAAiD,CAAA;AAAA,IAC5E;AACA,IAAA,OAAO,IAAA,CAAK,eAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,MAAA,GAAiB;AACnB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,OAAA,EAAS;AACjC,MAAA,MAAM,IAAI,eAAe,wCAAwC,CAAA;AAAA,IACnE;AACA,IAAA,OAAO,IAAA,CAAK,OAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,OAAA,GAAkB;AACpB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,IAAA,CAAK,aAAa,MAAA,EAAW;AAC/C,MAAA,MAAM,IAAI,eAAe,yCAAyC,CAAA;AAAA,IACpE;AACA,IAAA,OAAO,IAAA,CAAK,QAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,KAAA,GAAgB;AAClB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,IAAA,CAAK,WAAW,MAAA,EAAW;AAC7C,MAAA,MAAM,IAAI,eAAe,wCAAwC,CAAA;AAAA,IACnE;AACA,IAAA,OAAO,IAAA,CAAK,MAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,WAAW,IAAA,CAAK,SAAA;AAAA,MAChB,SAAS,IAAA,CAAK,OAAA;AAAA,MACd,KAAK,IAAA,CAAK,GAAA;AAAA,MACV,MAAM,IAAA,CAAK,IAAA;AAAA,MACX,aAAa,IAAA,CAAK;AAAA,KACpB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,UAAU,MAAA,EAAuC;AAC/C,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,KAAK,KAAK,MAAA,CAAO,OAAA,CAAQ,MAAM,CAAA,EAAG;AACjD,MAAA,QAAQ,GAAA;AAAK,QACX,KAAK,WAAA;AACH,UAAA,IAAI,OAAO,UAAU,QAAA,IAAY,CAAC,OAAO,SAAA,CAAU,KAAK,CAAA,IAAK,KAAA,GAAQ,CAAA,EAAG;AACtE,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,mCAAA;AAAA,cACA,WAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,SAAA,GAAY,KAAA;AACjB,UAAA;AAAA,QACF,KAAK,SAAA;AACH,UAAA,IAAI,OAAO,UAAU,QAAA,IAAY,CAAC,OAAO,SAAA,CAAU,KAAK,CAAA,IAAK,KAAA,GAAQ,CAAA,EAAG;AACtE,YAAA,MAAM,IAAI,qBAAA,CAAsB,iCAAA,EAAmC,SAAA,EAAW,KAAK,CAAA;AAAA,UACrF;AACA,UAAA,IAAA,CAAK,OAAA,GAAU,KAAA;AACf,UAAA;AAAA,QACF,KAAK,KAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,KAAA,GAAQ,CAAA,EAAG;AAC1C,YAAA,MAAM,IAAI,qBAAA,CAAsB,kBAAA,EAAoB,KAAA,EAAO,KAAK,CAAA;AAAA,UAClE;AACA,UAAA,IAAA,CAAK,GAAA,GAAM,KAAA;AACX,UAAA;AAAA,QACF,KAAK,MAAA;AACH,UAAA,IAAI,KAAA,KAAU,QAAA,IAAY,KAAA,KAAU,UAAA,EAAY;AAC9C,YAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,mCAAA,CAAA,EAAuC,MAAA,EAAQ,KAAK,CAAA;AAAA,UACtF;AACA,UAAA,IAAA,CAAK,IAAA,GAAO,KAAA;AACZ,UAAA;AAAA,QACF,KAAK,aAAA;AACH,UAAA,IAAI,KAAA,KAAU,WAAc,OAAO,KAAA,KAAU,YAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,CAAA,EAAI;AACjF,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,qCAAA;AAAA,cACA,aAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,WAAA,GAAc,KAAA,KAAU,MAAA,GAAY,MAAA,GAAY,KAAA;AACrD,UAAA;AAAA,QACF;AACE,UAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,mBAAA,EAAsB,GAAG,CAAA,CAAA,EAAI,KAAK,KAAK,CAAA;AAAA;AAC3E,IACF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AACF;;;AChcO,IAAM,MAAN,MAAiC;AAAA,EACrB,WAAA;AAAA,EACA,MAAA;AAAA,EAET,WAAA;AAAA,EACA,kBAAA;AAAA,EACA,uBAAA;AAAA,EACA,KAAA;AAAA,EACA,kBAAA;AAAA,EACA,YAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASjB,WAAA,CACE,OAAA,GAGI,EAAC,EACL;AACA,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AACA,IAAA,IAAA,CAAK,MAAA,GAAS,QAAQ,MAAA,IAAU,KAAA;AAEhC,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,MAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,cAAc,CAAA,EAAG;AAC/D,QAAA,MAAM,IAAI,qBAAA;AAAA,UACR,qCAAA;AAAA,UACA,aAAA;AAAA,UACA,IAAA,CAAK;AAAA,SACP;AAAA,MACF;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,GAAA,CAAI,GAAW,EAAA,EAAmB;AAChC,IAAA,6BAAA,CAA8B,CAAC,CAAA;AAE/B,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAA,CAAK,YAAA,GAAe,SAAA;AAEpB,IAAA,IAAI,WAAW,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,oBAAoB,wCAAwC,CAAA;AAAA,IACxE;AAGA,IAAA,MAAM,oBAAoB,IAAA,CAAK,WAAA,IAAe,IAAA,CAAK,GAAA,CAAI,UAAU,SAAS,CAAA;AAC1E,IAAA,IAAI,iBAAA,GAAoB,IAAA,CAAK,GAAA,CAAI,QAAA,EAAU,SAAS,CAAA,EAAG;AACrD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,eAAe,iBAAiB,CAAA,uCAAA,EAA0C,KAAK,GAAA,CAAI,QAAA,EAAU,SAAS,CAAC,CAAA,CAAA;AAAA,QACvG,aAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AAGA,IAAA,MAAM,OAAA,GAAU,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACzB,IAAA,IAAA,CAAK,KAAA,GAAQ,OAAA;AAEb,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,UAAA,CAAW,CAAA,EAAG,OAAO,CAAA;AAG5C,IAAA,MAAM,CAAC,EAAA,EAAI,CAAA,EAAG,EAAE,CAAA,GAAI,GAAA,CAAI,WAAW,KAAK,CAAA;AAGxC,IAAA,MAAM,aAAyB,EAAC;AAChC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,iBAAA,EAAmB,CAAA,EAAA,EAAK;AAC1C,MAAA,MAAM,YAAsB,EAAC;AAC7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,SAAA,CAAU,IAAA,CAAK,MAAA,CAAO,EAAA,CAAG,IAAA,CAAK,EAAA,CAAG,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MAC/D;AACA,MAAA,UAAA,CAAW,KAAK,SAAS,CAAA;AAAA,IAC3B;AACA,IAAA,IAAA,CAAK,WAAA,GAAc,OAAO,UAAU,CAAA;AAGpC,IAAA,MAAM,oBAA8B,EAAC;AACrC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,iBAAA,EAAmB,CAAA,EAAA,EAAK;AAC1C,MAAA,MAAM,KAAK,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AACtC,MAAA,iBAAA,CAAkB,IAAA,CAAM,EAAA,GAAK,EAAA,IAAO,QAAA,GAAW,CAAA,CAAE,CAAA;AAAA,IACnD;AACA,IAAA,IAAA,CAAK,kBAAA,GAAqB,OAAO,iBAAiB,CAAA;AAGlD,IAAA,IAAI,aAAA,GAAgB,CAAA;AACpB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,KAAK,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AACtC,MAAA,aAAA,IAAkB,EAAA,GAAK,MAAO,QAAA,GAAW,CAAA,CAAA;AAAA,IAC3C;AACA,IAAA,MAAM,sBAAA,GACJ,aAAA,KAAkB,CAAA,GACd,iBAAA,CAAkB,GAAA,CAAI,MAAM,CAAC,CAAA,GAC7B,iBAAA,CAAkB,GAAA,CAAI,CAAC,CAAA,KAAM,IAAI,aAAa,CAAA;AACpD,IAAA,IAAA,CAAK,uBAAA,GAA0B,OAAO,sBAAsB,CAAA;AAE5D,IAAA,IAAA,CAAK,kBAAA,GAAqB,iBAAA;AAC1B,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,CAAA,EAAmB;AAC3B,IAAA,IAAI,CAAC,KAAK,MAAA,IAAU,CAAC,KAAK,WAAA,IAAe,CAAC,KAAK,KAAA,EAAO;AACpD,MAAA,MAAM,IAAI,eAAe,qCAAqC,CAAA;AAAA,IAChE;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,KAAK,CAAA;AAEtD,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,WAAA,GAAc,KAAK,kBAAA,IAAsB,CAAA;AAG/C,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,UAAA,CAAW,CAAA,EAAG,KAAK,KAAK,CAAA;AAG/C,IAAA,MAAM,cAA0B,EAAC;AACjC,IAAA,MAAM,WAAA,GAAc,KAAA;AACpB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,WAAA,EAAa,CAAA,EAAA,EAAK;AACpC,QAAA,IAAI,GAAA,GAAM,CAAA;AACV,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,GAAA,IACE,MAAA,CAAO,UAAU,IAAA,CAAK,SAAA,CAAU,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA,GAC3D,OAAO,IAAA,CAAK,WAAA,CAAY,KAAK,IAAA,CAAK,WAAA,CAAY,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AAAA,QAC7E;AAEA,QAAA,IAAI,KAAK,MAAA,EAAQ;AACf,UAAA,MAAM,QAAA,GAAW,MAAA;AAAA,YACf,KAAK,kBAAA,EAAoB,IAAA,CAAK,KAAK,kBAAA,CAAmB,MAAA,GAAS,CAAC,CAAA,IAAK;AAAA,WACvE;AACA,UAAA,GAAA,CAAI,KAAK,GAAA,GAAM,IAAA,CAAK,IAAA,CAAK,QAAA,GAAW,WAAW,CAAC,CAAA;AAAA,QAClD,CAAA,MAAO;AACL,UAAA,GAAA,CAAI,KAAK,GAAG,CAAA;AAAA,QACd;AAAA,MACF;AACA,MAAA,WAAA,CAAY,KAAK,GAAG,CAAA;AAAA,IACtB;AAEA,IAAA,OAAO,OAAO,WAAW,CAAA;AAAA,EAC3B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,YAAA,CAAa,GAAW,EAAA,EAAqB;AAC3C,IAAA,IAAA,CAAK,IAAI,CAAC,CAAA;AACV,IAAA,OAAO,IAAA,CAAK,UAAU,CAAC,CAAA;AAAA,EACzB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,iBAAiB,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAC,KAAK,MAAA,IAAU,CAAC,KAAK,WAAA,IAAe,CAAC,KAAK,KAAA,EAAO;AACpD,MAAA,MAAM,IAAI,eAAe,6CAA6C,CAAA;AAAA,IACxE;AAEA,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAEvB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,WAAA,GAAc,KAAK,kBAAA,IAAsB,CAAA;AAC/C,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,WAAA,CAAY,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/C,IAAA,IAAA,CAAK,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,OAAO,WAAA,EAAa;AACrC,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,YAAA,EAAe,WAAW,CAAA,iBAAA,EAAA,CAAqB,CAAA,CAAE,MAAM,CAAC,CAAA,IAAK,CAAA,EAAG,QAAA,EAAU,CAAA;AAAA,OAC5E;AAAA,IACF;AAEA,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AAGA,IAAA,MAAM,gBAA4B,EAAC;AACnC,IAAA,MAAM,WAAA,GAAc,KAAA;AACpB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,IAAI,GAAA,GAAM,CAAA;AACV,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,WAAA,EAAa,CAAA,EAAA,EAAK;AACpC,UAAA,MAAM,IAAA,GAAO,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAA,GAAI,WAAA,GAAc,CAAC,CAAC,CAAA;AAC1D,UAAA,MAAM,QAAA,GAAW,MAAA;AAAA,YACf,KAAK,kBAAA,EAAoB,IAAA,CAAK,KAAK,kBAAA,CAAmB,MAAA,GAAS,CAAC,CAAA,IAAK;AAAA,WACvE;AAEA,UAAA,MAAM,MAAA,GAAS,KAAK,MAAA,GAAS,IAAA,GAAO,KAAK,IAAA,CAAK,QAAA,GAAW,WAAW,CAAA,GAAI,IAAA;AACxE,UAAA,GAAA,IACE,MAAA,GAAS,MAAA,CAAO,IAAA,CAAK,WAAA,CAAY,IAAA,CAAK,IAAA,CAAK,WAAA,CAAY,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AAAA,QACtF;AAEA,QAAA,GAAA,IAAO,MAAA,CAAO,KAAK,KAAA,CAAM,IAAA,CAAK,KAAK,KAAA,CAAM,MAAA,GAAS,CAAC,CAAC,CAAA;AACpD,QAAA,GAAA,CAAI,KAAK,GAAG,CAAA;AAAA,MACd;AACA,MAAA,aAAA,CAAc,KAAK,GAAG,CAAA;AAAA,IACxB;AAEA,IAAA,OAAO,OAAO,aAAa,CAAA;AAAA,EAC7B;AAAA;AAAA;AAAA;AAAA,EAKQ,UAAA,CAAW,GAAW,OAAA,EAAyB;AACrD,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,WAAuB,EAAC;AAE9B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,GAAA,GAAM,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AACvD,QAAA,MAAM,UAAU,MAAA,CAAO,OAAA,CAAQ,KAAK,OAAA,CAAQ,MAAA,GAAS,CAAC,CAAC,CAAA;AACvD,QAAA,GAAA,CAAI,IAAA,CAAK,MAAM,OAAO,CAAA;AAAA,MACxB;AACA,MAAA,QAAA,CAAS,KAAK,GAAG,CAAA;AAAA,IACnB;AAEA,IAAA,OAAO,OAAO,QAAQ,CAAA;AAAA,EACxB;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,UAAA,GAAqB;AACvB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,WAAA,EAAa;AACrC,MAAA,MAAM,IAAI,eAAe,yCAAyC,CAAA;AAAA,IACpE;AACA,IAAA,OAAO,IAAA,CAAK,WAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,iBAAA,GAA4B;AAC9B,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,kBAAA,EAAoB;AAC5C,MAAA,MAAM,IAAI,eAAe,iDAAiD,CAAA;AAAA,IAC5E;AACA,IAAA,OAAO,IAAA,CAAK,kBAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,sBAAA,GAAiC;AACnC,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,uBAAA,EAAyB;AACjD,MAAA,MAAM,IAAI,eAAe,uDAAuD,CAAA;AAAA,IAClF;AACA,IAAA,OAAO,IAAA,CAAK,uBAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,QAAQ,IAAA,CAAK;AAAA,KACf;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,oBAAoB,mDAAmD,CAAA;AAAA,EACnF;AACF;;;AChTO,IAAM,yBAAN,MAAmD;AAAA,EAChD,QAAA;AAAA,EACA,eAAA;AAAA,EACA,cAAA;AAAA,EACA,WAAA;AAAA,EACA,WAAA;AAAA,EAEA,IAAA;AAAA,EACA,SAAA;AAAA,EACA,WAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAMI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,QAAA,GAAW,QAAQ,QAAA,IAAY,EAAA;AACpC,IAAA,IAAA,CAAK,eAAA,GAAkB,QAAQ,eAAA,IAAmB,CAAA;AAClD,IAAA,IAAA,CAAK,cAAA,GAAiB,QAAQ,cAAA,IAAkB,CAAA;AAChD,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AACA,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AAEA,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,IAAa,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,WAAW,CAAA,EAAG;AACxE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,IAAA,CAAK,WAAW,CAAC,CAAA,CAAA;AAAA,QACzE,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAEA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,QAAQ,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACzD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,2CAAA,EAA8C,KAAK,QAAQ,CAAA,CAAA;AAAA,QAC3D,UAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,kBAAkB,CAAA,EAAG;AACvE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,kDAAA,EAAqD,KAAK,eAAe,CAAA,CAAA;AAAA,QACzE,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,cAAc,CAAA,IAAK,IAAA,CAAK,iBAAiB,CAAA,EAAG;AACrE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,iDAAA,EAAoD,KAAK,cAAc,CAAA,CAAA;AAAA,QACvE,gBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IACE,IAAA,CAAK,WAAA,KAAgB,MAAA,KACpB,CAAC,MAAA,CAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,WAAA,GAAc,CAAA,CAAA,EAC3D;AACA,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,KAAK,WAAW,CAAA,CAAA;AAAA,QACjE,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAAA,EACF;AAAA,EAEQ,MAAA,GAAuB;AAC7B,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,MAAA,OAAO,IAAA,CAAK,MAAA;AAAA,IACd;AACA,IAAA,IAAI,OAAO,IAAA,CAAK,WAAA;AAChB,IAAA,OAAO,MAAM;AACX,MAAA,IAAA,GAAQ,IAAA,GAAO,UAAU,UAAA,KAAgB,CAAA;AACzC,MAAA,OAAO,IAAA,GAAO,UAAA;AAAA,IAChB,CAAA;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AACtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAGjB,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,MAAM,QAAkB,EAAC;AAEzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AACd,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAGA,IAAA,IAAA,CAAK,WAAA,GAAc,CAAC,GAAG,IAAI,GAAA,CAAI,KAAK,CAAC,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,KAAM,IAAI,CAAC,CAAA;AAG3D,IAAA,MAAM,OAAA,GAAU,KAAA,CAAM,IAAA,CAAK,EAAE,MAAA,EAAQ,UAAS,EAAG,CAAC,CAAA,EAAG,CAAA,KAAM,CAAC,CAAA;AAC5D,IAAA,IAAA,CAAK,OAAO,IAAA,CAAK,SAAA,CAAU,KAAA,EAAO,KAAA,EAAO,SAAS,CAAC,CAAA;AACnD,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA,EAEQ,SAAA,CACN,KAAA,EACA,KAAA,EACA,OAAA,EACA,KAAA,EACU;AACV,IAAA,MAAM,IAAI,OAAA,CAAQ,MAAA;AAGlB,IAAA,IAAI,KAAA,IAAS,KAAK,QAAA,IAAY,CAAA,GAAI,KAAK,eAAA,IAAmB,CAAA,GAAI,KAAK,cAAA,EAAgB;AACjF,MAAA,OAAO;AAAA,QACL,MAAA,EAAQ,IAAA;AAAA,QACR,UAAA,EAAY,IAAA,CAAK,gBAAA,CAAiB,KAAA,EAAO,OAAO,CAAA;AAAA,QAChD,kBAAA,EAAoB,IAAA,CAAK,qBAAA,CAAsB,KAAA,EAAO,OAAO;AAAA,OAC/D;AAAA,IACF;AAGA,IAAA,MAAM,OAAA,GAAU,IAAI,GAAA,CAAI,OAAA,CAAQ,GAAA,CAAI,CAAC,CAAA,KAAM,KAAA,CAAM,CAAC,CAAC,CAAC,CAAA;AACpD,IAAA,IAAI,OAAA,CAAQ,SAAS,CAAA,EAAG;AACtB,MAAA,MAAM,QAAA,GAAW,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA;AAC/B,MAAA,OAAO;AAAA,QACL,MAAA,EAAQ,IAAA;AAAA,QACR,UAAA,EAAY,KAAA,CAAM,QAAQ,CAAA,IAAK,CAAA;AAAA,QAC/B,kBAAA,EAAoB,IAAA,CAAK,qBAAA,CAAsB,KAAA,EAAO,OAAO;AAAA,OAC/D;AAAA,IACF;AAGA,IAAA,MAAM,EAAE,YAAA,EAAc,SAAA,EAAW,WAAA,EAAa,YAAA,KAAiB,IAAA,CAAK,aAAA;AAAA,MAClE,KAAA;AAAA,MACA,KAAA;AAAA,MACA;AAAA,KACF;AAEA,IAAA,IAAI,WAAA,CAAY,MAAA,KAAW,CAAA,IAAK,YAAA,CAAa,WAAW,CAAA,EAAG;AACzD,MAAA,OAAO;AAAA,QACL,MAAA,EAAQ,IAAA;AAAA,QACR,UAAA,EAAY,IAAA,CAAK,gBAAA,CAAiB,KAAA,EAAO,OAAO,CAAA;AAAA,QAChD,kBAAA,EAAoB,IAAA,CAAK,qBAAA,CAAsB,KAAA,EAAO,OAAO;AAAA,OAC/D;AAAA,IACF;AAGA,IAAA,MAAM,OAAO,IAAA,CAAK,SAAA,CAAU,OAAO,KAAA,EAAO,WAAA,EAAa,QAAQ,CAAC,CAAA;AAChE,IAAA,MAAM,QAAQ,IAAA,CAAK,SAAA,CAAU,OAAO,KAAA,EAAO,YAAA,EAAc,QAAQ,CAAC,CAAA;AAElE,IAAA,OAAO;AAAA,MACL,MAAA,EAAQ,KAAA;AAAA,MACR,YAAA;AAAA,MACA,SAAA;AAAA,MACA,IAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAAA,EAEQ,gBAAA,CAAiB,OAAiB,OAAA,EAA2B;AACnE,IAAA,MAAM,MAAA,uBAAa,GAAA,EAAoB;AACvC,IAAA,KAAA,MAAW,KAAK,OAAA,EAAS;AACvB,MAAA,MAAM,KAAA,GAAQ,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC1B,MAAA,MAAA,CAAO,IAAI,KAAA,EAAA,CAAQ,MAAA,CAAO,IAAI,KAAK,CAAA,IAAK,KAAK,CAAC,CAAA;AAAA,IAChD;AAEA,IAAA,IAAI,QAAA,GAAW,CAAA;AACf,IAAA,IAAI,QAAA,GAAW,CAAA;AACf,IAAA,KAAA,MAAW,CAAC,KAAA,EAAO,KAAK,CAAA,IAAK,MAAA,EAAQ;AACnC,MAAA,IAAI,QAAQ,QAAA,EAAU;AACpB,QAAA,QAAA,GAAW,KAAA;AACX,QAAA,QAAA,GAAW,KAAA;AAAA,MACb;AAAA,IACF;AACA,IAAA,OAAO,QAAA;AAAA,EACT;AAAA,EAEQ,qBAAA,CAAsB,OAAiB,OAAA,EAA6B;AAC1E,IAAA,MAAM,MAAA,GAAS,IAAA,CAAK,WAAA,IAAe,EAAC;AACpC,IAAA,IAAI,MAAA,CAAO,MAAA,KAAW,CAAA,IAAK,OAAA,CAAQ,WAAW,CAAA,EAAG;AAC/C,MAAA,OAAO,EAAC;AAAA,IACV;AAEA,IAAA,MAAM,UAAA,uBAAiB,GAAA,EAAoB;AAC3C,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,MAAA,CAAO,QAAQ,CAAA,EAAA,EAAK;AACtC,MAAA,MAAM,CAAA,GAAI,OAAO,CAAC,CAAA;AAClB,MAAA,IAAI,CAAA,KAAM,MAAA,EAAW,UAAA,CAAW,GAAA,CAAI,GAAG,CAAC,CAAA;AAAA,IAC1C;AAEA,IAAA,MAAM,SAAS,IAAI,KAAA,CAAc,OAAO,MAAM,CAAA,CAAE,KAAK,CAAC,CAAA;AACtD,IAAA,KAAA,MAAW,SAAS,OAAA,EAAS;AAC3B,MAAA,MAAM,KAAA,GAAQ,KAAA,CAAM,KAAK,CAAA,IAAK,CAAA;AAC9B,MAAA,MAAM,GAAA,GAAM,UAAA,CAAW,GAAA,CAAI,KAAK,CAAA;AAChC,MAAA,IAAI,GAAA,KAAQ,QAAW,MAAA,CAAO,GAAG,KAAK,MAAA,CAAO,GAAG,KAAK,CAAA,IAAK,CAAA;AAAA,IAC5D;AAEA,IAAA,MAAM,IAAA,GAAO,IAAI,OAAA,CAAQ,MAAA;AACzB,IAAA,OAAO,MAAA,CAAO,GAAA,CAAI,CAAC,CAAA,KAAM,IAAI,IAAI,CAAA;AAAA,EACnC;AAAA,EAEQ,aAAA,CACN,KAAA,EACA,KAAA,EACA,OAAA,EAMA;AACA,IAAA,IAAI,QAAA,GAAW,QAAA;AACf,IAAA,IAAI,WAAA,GAAc,CAAA;AAClB,IAAA,IAAI,aAAA,GAAgB,CAAA;AACpB,IAAA,IAAI,WAAqB,EAAC;AAC1B,IAAA,IAAI,YAAsB,EAAC;AAE3B,IAAA,MAAM,SAAA,GAAY,KAAA,CAAM,CAAC,CAAA,EAAG,MAAA,IAAU,CAAA;AACtC,IAAA,IAAI,cAAA,GAAiB,KAAA,CAAM,IAAA,CAAK,EAAE,MAAA,EAAQ,WAAU,EAAG,CAAC,CAAA,EAAG,CAAA,KAAM,CAAC,CAAA;AAElE,IAAA,IAAI,IAAA,CAAK,WAAA,KAAgB,MAAA,IAAa,IAAA,CAAK,cAAc,SAAA,EAAW;AAClE,MAAA,MAAM,GAAA,GAAM,KAAK,MAAA,EAAO;AAExB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,QAAA,MAAM,IAAI,CAAA,GAAI,IAAA,CAAK,MAAM,GAAA,EAAI,IAAK,YAAY,CAAA,CAAE,CAAA;AAChD,QAAA,MAAM,IAAA,GAAO,eAAe,CAAC,CAAA;AAC7B,QAAA,IAAI,SAAS,MAAA,EAAW;AACtB,UAAA,MAAM,OAAA,GAAU,eAAe,CAAC,CAAA;AAChC,UAAA,IAAI,YAAY,MAAA,EAAW;AACzB,YAAA,cAAA,CAAe,CAAC,CAAA,GAAI,OAAA;AACpB,YAAA,cAAA,CAAe,CAAC,CAAA,GAAI,IAAA;AAAA,UACtB;AAAA,QACF;AAAA,MACF;AACA,MAAA,cAAA,GAAiB,cAAA,CAAe,KAAA,CAAM,CAAA,EAAG,IAAA,CAAK,WAAW,CAAA;AAAA,IAC3D;AAEA,IAAA,MAAM,IAAI,OAAA,CAAQ,MAAA;AAElB,IAAA,MAAM,WAAA,uBAAkB,GAAA,EAAoB;AAC5C,IAAA,KAAA,MAAW,KAAK,OAAA,EAAS;AACvB,MAAA,MAAM,KAAA,GAAQ,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC1B,MAAA,WAAA,CAAY,IAAI,KAAA,EAAA,CAAQ,WAAA,CAAY,IAAI,KAAK,CAAA,IAAK,KAAK,CAAC,CAAA;AAAA,IAC1D;AAEA,IAAA,KAAA,MAAW,KAAK,cAAA,EAAgB;AAG9B,MAAA,MAAM,aAAA,GAAgB,CAAC,GAAG,OAAO,CAAA,CAAE,IAAA;AAAA,QACjC,CAAC,CAAA,EAAG,CAAA,KAAA,CAAO,KAAA,CAAM,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA,KAAM,KAAA,CAAM,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAAA,OACrD;AAEA,MAAA,MAAM,UAAA,uBAAiB,GAAA,EAAoB;AAC3C,MAAA,MAAM,WAAA,GAAc,IAAI,GAAA,CAAoB,WAAW,CAAA;AACvD,MAAA,IAAI,QAAA,GAAW,CAAA;AACf,MAAA,IAAI,SAAA,GAAY,CAAA;AAEhB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,GAAI,GAAG,CAAA,EAAA,EAAK;AAC9B,QAAA,MAAM,GAAA,GAAM,cAAc,CAAC,CAAA;AAC3B,QAAA,IAAI,QAAQ,MAAA,EAAW;AACvB,QAAA,MAAM,KAAA,GAAQ,KAAA,CAAM,GAAG,CAAA,IAAK,CAAA;AAC5B,QAAA,MAAM,GAAA,GAAM,KAAA,CAAM,GAAG,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAC/B,QAAA,MAAM,OAAA,GAAU,aAAA,CAAc,CAAA,GAAI,CAAC,CAAA;AACnC,QAAA,IAAI,YAAY,MAAA,EAAW;AAC3B,QAAA,MAAM,OAAA,GAAU,KAAA,CAAM,OAAO,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAGvC,QAAA,MAAM,YAAA,GAAe,WAAA,CAAY,GAAA,CAAI,KAAK,CAAA,IAAK,CAAA;AAC/C,QAAA,IAAI,YAAA,IAAgB,CAAA,EAAG,WAAA,CAAY,MAAA,CAAO,KAAK,CAAA;AAAA,aAC1C,WAAA,CAAY,GAAA,CAAI,KAAA,EAAO,YAAA,GAAe,CAAC,CAAA;AAC5C,QAAA,SAAA,EAAA;AAEA,QAAA,UAAA,CAAW,IAAI,KAAA,EAAA,CAAQ,UAAA,CAAW,IAAI,KAAK,CAAA,IAAK,KAAK,CAAC,CAAA;AACtD,QAAA,QAAA,EAAA;AAEA,QAAA,IAAI,QAAQ,OAAA,EAAS;AAErB,QAAA,IAAI,QAAA,GAAW,IAAA,CAAK,cAAA,IAAkB,SAAA,GAAY,KAAK,cAAA,EAAgB;AAGvE,QAAA,MAAM,QAAA,GAAW,IAAA,CAAK,cAAA,CAAe,UAAA,EAAY,QAAQ,CAAA;AACzD,QAAA,MAAM,SAAA,GAAY,IAAA,CAAK,cAAA,CAAe,WAAA,EAAa,SAAS,CAAA;AAC5D,QAAA,MAAM,YAAA,GAAA,CAAgB,QAAA,GAAW,QAAA,GAAW,SAAA,GAAY,SAAA,IAAa,CAAA;AAErE,QAAA,IAAI,eAAe,QAAA,EAAU;AAC3B,UAAA,QAAA,GAAW,YAAA;AACX,UAAA,WAAA,GAAc,CAAA;AACd,UAAA,aAAA,GAAA,CAAiB,MAAM,OAAA,IAAW,CAAA;AAClC,UAAA,QAAA,GAAW,aAAA,CAAc,KAAA,CAAM,CAAA,EAAG,CAAA,GAAI,CAAC,CAAA;AACvC,UAAA,SAAA,GAAY,aAAA,CAAc,KAAA,CAAM,CAAA,GAAI,CAAC,CAAA;AAAA,QACvC;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO;AAAA,MACL,YAAA,EAAc,WAAA;AAAA,MACd,SAAA,EAAW,aAAA;AAAA,MACX,WAAA,EAAa,QAAA;AAAA,MACb,YAAA,EAAc;AAAA,KAChB;AAAA,EACF;AAAA,EAEQ,cAAA,CAAe,QAA6B,CAAA,EAAmB;AACrE,IAAA,IAAI,QAAA,GAAW,CAAA;AACf,IAAA,KAAA,MAAW,KAAA,IAAS,MAAA,CAAO,MAAA,EAAO,EAAG;AACnC,MAAA,MAAM,IAAI,KAAA,GAAQ,CAAA;AAClB,MAAA,QAAA,IAAY,CAAA,GAAI,CAAA;AAAA,IAClB;AACA,IAAA,OAAO,QAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,IAAA,EAAM;AAC9B,MAAA,MAAM,IAAI,eAAe,yDAAyD,CAAA;AAAA,IACpF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,wBAAwB,CAAA;AAEtE,IAAA,MAAM,cAAwB,EAAC;AAC/B,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MAC1D;AACA,MAAA,WAAA,CAAY,KAAK,IAAA,CAAK,aAAA,CAAc,MAAA,EAAQ,IAAA,CAAK,IAAI,CAAC,CAAA;AAAA,IACxD;AAEA,IAAA,OAAO,MAAA,CAAO,WAAA,EAAa,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EAC/C;AAAA,EAEQ,aAAA,CAAc,QAAkB,IAAA,EAAwB;AAC9D,IAAA,IAAI,OAAA,GAAU,IAAA;AACd,IAAA,OAAO,CAAC,QAAQ,MAAA,EAAQ;AACtB,MAAA,MAAM,YAAA,GAAe,MAAA,CAAO,OAAA,CAAQ,YAAA,IAAgB,CAAC,CAAA,IAAK,CAAA;AAC1D,MAAA,IAAI,YAAA,KAAiB,OAAA,CAAQ,SAAA,IAAa,CAAA,CAAA,EAAI;AAC5C,QAAA,IAAI,CAAC,OAAA,CAAQ,IAAA;AACX,UAAA,MAAM,IAAI,aAAa,kDAAkD,CAAA;AAC3E,QAAA,OAAA,GAAU,OAAA,CAAQ,IAAA;AAAA,MACpB,CAAA,MAAO;AACL,QAAA,IAAI,CAAC,OAAA,CAAQ,KAAA;AACX,UAAA,MAAM,IAAI,aAAa,mDAAmD,CAAA;AAC5E,QAAA,OAAA,GAAU,OAAA,CAAQ,KAAA;AAAA,MACpB;AAAA,IACF;AACA,IAAA,OAAO,QAAQ,UAAA,IAAc,CAAA;AAAA,EAC/B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,aAAa,CAAA,EAAmB;AAC9B,IAAA,IAAI,CAAC,KAAK,MAAA,IAAU,CAAC,KAAK,IAAA,IAAQ,CAAC,KAAK,WAAA,EAAa;AACnD,MAAA,MAAM,IAAI,eAAe,yDAAyD,CAAA;AAAA,IACpF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,wBAAwB,CAAA;AACtE,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAEvB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,QAAA,GAAW,KAAK,WAAA,CAAY,MAAA;AAElC,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MAC1D;AAEA,MAAA,MAAM,IAAA,GAAO,IAAA,CAAK,WAAA,CAAY,MAAA,EAAQ,KAAK,IAAI,CAAA;AAC/C,MAAA,MAAM,GAAA,GAAM,IAAA,CAAK,kBAAA,GACb,CAAC,GAAG,IAAA,CAAK,kBAAkB,CAAA,GAC3B,IAAI,KAAA,CAAM,QAAQ,CAAA,CAAE,KAAK,CAAC,CAAA;AAC9B,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AAAA,IAChB;AAEA,IAAA,OAAO,OAAO,KAAK,CAAA;AAAA,EACrB;AAAA,EAEQ,WAAA,CAAY,QAAkB,IAAA,EAA0B;AAC9D,IAAA,IAAI,KAAK,MAAA,EAAQ;AACf,MAAA,OAAO,IAAA;AAAA,IACT;AAEA,IAAA,MAAM,YAAA,GAAe,MAAA,CAAO,IAAA,CAAK,YAAA,IAAgB,CAAC,CAAA,IAAK,CAAA;AACvD,IAAA,IAAI,YAAA,KAAiB,IAAA,CAAK,SAAA,IAAa,CAAA,CAAA,EAAI;AACzC,MAAA,IAAI,CAAC,KAAK,IAAA,EAAM;AACd,QAAA,OAAO,IAAA;AAAA,MACT;AACA,MAAA,OAAO,IAAA,CAAK,WAAA,CAAY,MAAA,EAAQ,IAAA,CAAK,IAAI,CAAA;AAAA,IAC3C;AACA,IAAA,IAAI,CAAC,KAAK,KAAA,EAAO;AACf,MAAA,OAAO,IAAA;AAAA,IACT;AACA,IAAA,OAAO,IAAA,CAAK,WAAA,CAAY,MAAA,EAAQ,IAAA,CAAK,KAAK,CAAA;AAAA,EAC5C;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AACA,IAAA,IAAI,OAAA,GAAU,CAAA;AACd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,IAAI,MAAA,CAAO,WAAA,CAAY,IAAA,CAAK,WAAA,CAAY,SAAS,CAAC,CAAC,CAAA,KAAM,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA,EAAG;AACrF,QAAA,OAAA,EAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,OAAO,UAAU,CAAA,CAAE,IAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,IAAI,OAAA,GAA8B;AAChC,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,WAAA,EAAa;AACrC,MAAA,OAAO,MAAA;AAAA,IACT;AACA,IAAA,OAAO,OAAO,IAAA,CAAK,WAAA,EAAa,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EACpD;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,UAAU,IAAA,CAAK,QAAA;AAAA,MACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,MACtB,gBAAgB,IAAA,CAAK,cAAA;AAAA,MACrB,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,aAAa,IAAA,CAAK;AAAA,KACpB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AACF;AASO,IAAM,wBAAN,MAAiD;AAAA,EAC9C,QAAA;AAAA,EACA,eAAA;AAAA,EACA,cAAA;AAAA,EACA,WAAA;AAAA,EACA,WAAA;AAAA,EAEA,IAAA;AAAA,EACA,SAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAMI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,QAAA,GAAW,QAAQ,QAAA,IAAY,EAAA;AACpC,IAAA,IAAA,CAAK,eAAA,GAAkB,QAAQ,eAAA,IAAmB,CAAA;AAClD,IAAA,IAAA,CAAK,cAAA,GAAiB,QAAQ,cAAA,IAAkB,CAAA;AAChD,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AACA,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AAEA,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,IAAa,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,WAAW,CAAA,EAAG;AACxE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,IAAA,CAAK,WAAW,CAAC,CAAA,CAAA;AAAA,QACzE,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAEA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,QAAQ,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACzD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,2CAAA,EAA8C,KAAK,QAAQ,CAAA,CAAA;AAAA,QAC3D,UAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,kBAAkB,CAAA,EAAG;AACvE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,kDAAA,EAAqD,KAAK,eAAe,CAAA,CAAA;AAAA,QACzE,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,cAAc,CAAA,IAAK,IAAA,CAAK,iBAAiB,CAAA,EAAG;AACrE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,iDAAA,EAAoD,KAAK,cAAc,CAAA,CAAA;AAAA,QACvE,gBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IACE,IAAA,CAAK,WAAA,KAAgB,MAAA,KACpB,CAAC,MAAA,CAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,WAAA,GAAc,CAAA,CAAA,EAC3D;AACA,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,KAAK,WAAW,CAAA,CAAA;AAAA,QACjE,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAAA,EACF;AAAA,EAEQ,MAAA,GAAuB;AAC7B,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,MAAA,OAAO,IAAA,CAAK,MAAA;AAAA,IACd;AACA,IAAA,IAAI,OAAO,IAAA,CAAK,WAAA;AAChB,IAAA,OAAO,MAAM;AACX,MAAA,IAAA,GAAQ,IAAA,GAAO,UAAU,UAAA,KAAgB,CAAA;AACzC,MAAA,OAAO,IAAA,GAAO,UAAA;AAAA,IAChB,CAAA;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AACtB,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAEvB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAEjB,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,MAAM,QAAkB,EAAC;AAEzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AACd,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAEA,IAAA,MAAM,OAAA,GAAU,KAAA,CAAM,IAAA,CAAK,EAAE,MAAA,EAAQ,UAAS,EAAG,CAAC,CAAA,EAAG,CAAA,KAAM,CAAC,CAAA;AAC5D,IAAA,IAAA,CAAK,OAAO,IAAA,CAAK,SAAA,CAAU,KAAA,EAAO,KAAA,EAAO,SAAS,CAAC,CAAA;AACnD,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA,EAEQ,SAAA,CACN,KAAA,EACA,KAAA,EACA,OAAA,EACA,KAAA,EACU;AACV,IAAA,MAAM,IAAI,OAAA,CAAQ,MAAA;AAClB,IAAA,IAAI,MAAM,CAAA,EAAG;AACX,MAAA,MAAM,IAAI,oBAAoB,oDAAoD,CAAA;AAAA,IACpF;AAEA,IAAA,IAAI,KAAA,IAAS,KAAK,QAAA,IAAY,CAAA,GAAI,KAAK,eAAA,IAAmB,CAAA,GAAI,KAAK,cAAA,EAAgB;AACjF,MAAA,OAAO,EAAE,QAAQ,IAAA,EAAM,UAAA,EAAY,KAAK,OAAA,CAAQ,KAAA,EAAO,OAAO,CAAA,EAAE;AAAA,IAClE;AAEA,IAAA,MAAM,EAAE,YAAA,EAAc,SAAA,EAAW,WAAA,EAAa,YAAA,KAAiB,IAAA,CAAK,aAAA;AAAA,MAClE,KAAA;AAAA,MACA,KAAA;AAAA,MACA;AAAA,KACF;AAEA,IAAA,IAAI,WAAA,CAAY,MAAA,KAAW,CAAA,IAAK,YAAA,CAAa,WAAW,CAAA,EAAG;AACzD,MAAA,OAAO,EAAE,QAAQ,IAAA,EAAM,UAAA,EAAY,KAAK,OAAA,CAAQ,KAAA,EAAO,OAAO,CAAA,EAAE;AAAA,IAClE;AAEA,IAAA,MAAM,OAAO,IAAA,CAAK,SAAA,CAAU,OAAO,KAAA,EAAO,WAAA,EAAa,QAAQ,CAAC,CAAA;AAChE,IAAA,MAAM,QAAQ,IAAA,CAAK,SAAA,CAAU,OAAO,KAAA,EAAO,YAAA,EAAc,QAAQ,CAAC,CAAA;AAElE,IAAA,OAAO;AAAA,MACL,MAAA,EAAQ,KAAA;AAAA,MACR,YAAA;AAAA,MACA,SAAA;AAAA,MACA,IAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAAA,EAEQ,OAAA,CAAQ,OAAiB,OAAA,EAA2B;AAC1D,IAAA,IAAI,GAAA,GAAM,CAAA;AACV,IAAA,KAAA,MAAW,KAAK,OAAA,EAAS;AACvB,MAAA,GAAA,IAAO,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAAA,IACrB;AACA,IAAA,OAAO,MAAM,OAAA,CAAQ,MAAA;AAAA,EACvB;AAAA,EAEQ,aAAA,CACN,KAAA,EACA,KAAA,EACA,OAAA,EAMA;AACA,IAAA,IAAI,SAAA,GAAY,CAAA,QAAA;AAChB,IAAA,IAAI,WAAA,GAAc,CAAA;AAClB,IAAA,IAAI,aAAA,GAAgB,CAAA;AACpB,IAAA,IAAI,WAAqB,EAAC;AAC1B,IAAA,IAAI,YAAsB,EAAC;AAE3B,IAAA,MAAM,SAAA,GAAY,KAAA,CAAM,CAAC,CAAA,EAAG,MAAA,IAAU,CAAA;AACtC,IAAA,IAAI,cAAA,GAAiB,KAAA,CAAM,IAAA,CAAK,EAAE,MAAA,EAAQ,WAAU,EAAG,CAAC,CAAA,EAAG,CAAA,KAAM,CAAC,CAAA;AAElE,IAAA,IAAI,IAAA,CAAK,WAAA,KAAgB,MAAA,IAAa,IAAA,CAAK,cAAc,SAAA,EAAW;AAClE,MAAA,MAAM,GAAA,GAAM,KAAK,MAAA,EAAO;AACxB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,QAAA,MAAM,IAAI,CAAA,GAAI,IAAA,CAAK,MAAM,GAAA,EAAI,IAAK,YAAY,CAAA,CAAE,CAAA;AAChD,QAAA,MAAM,CAAA,GAAI,eAAe,CAAC,CAAA;AAC1B,QAAA,MAAM,CAAA,GAAI,eAAe,CAAC,CAAA;AAC1B,QAAA,IAAI,CAAA,KAAM,MAAA,IAAa,CAAA,KAAM,MAAA,EAAW;AACtC,UAAA,MAAM,IAAI,YAAA,CAAa,CAAA,gDAAA,EAAmD,CAAC,CAAA,IAAA,EAAO,CAAC,CAAA,CAAE,CAAA;AAAA,QACvF;AACA,QAAA,cAAA,CAAe,CAAC,CAAA,GAAI,CAAA;AACpB,QAAA,cAAA,CAAe,CAAC,CAAA,GAAI,CAAA;AAAA,MACtB;AACA,MAAA,cAAA,GAAiB,cAAA,CAAe,KAAA,CAAM,CAAA,EAAG,IAAA,CAAK,WAAW,CAAA;AAAA,IAC3D;AAEA,IAAA,MAAM,IAAI,OAAA,CAAQ,MAAA;AAClB,IAAA,IAAI,QAAA,GAAW,CAAA;AAEf,IAAA,KAAA,MAAW,KAAK,OAAA,EAAS;AACvB,MAAA,MAAM,IAAA,GAAO,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACzB,MAAA,QAAA,IAAY,IAAA;AAAA,IACd;AAEA,IAAA,KAAA,MAAW,KAAK,cAAA,EAAgB;AAE9B,MAAA,MAAM,aAAA,GAAgB,CAAC,GAAG,OAAO,CAAA,CAAE,IAAA;AAAA,QACjC,CAAC,CAAA,EAAG,CAAA,KAAA,CAAO,KAAA,CAAM,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA,KAAM,KAAA,CAAM,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAAA,OACrD;AAEA,MAAA,IAAI,OAAA,GAAU,CAAA;AACd,MAAA,IAAI,OAAA,GAAU,CAAA;AACd,MAAA,IAAI,QAAA,GAAW,QAAA;AACf,MAAA,IAAI,QAAA,GAAW,CAAA;AAEf,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,GAAI,GAAG,CAAA,EAAA,EAAK;AAC9B,QAAA,MAAM,GAAA,GAAM,cAAc,CAAC,CAAA;AAC3B,QAAA,IAAI,QAAQ,MAAA,EAAW;AACvB,QAAA,MAAM,GAAA,GAAM,KAAA,CAAM,GAAG,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAC/B,QAAA,MAAM,OAAA,GAAU,aAAA,CAAc,CAAA,GAAI,CAAC,CAAA;AACnC,QAAA,IAAI,YAAY,MAAA,EAAW;AAC3B,QAAA,MAAM,OAAA,GAAU,KAAA,CAAM,OAAO,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AACvC,QAAA,MAAM,IAAA,GAAO,KAAA,CAAM,GAAG,CAAA,IAAK,CAAA;AAG3B,QAAA,OAAA,IAAW,IAAA;AACX,QAAA,OAAA,EAAA;AACA,QAAA,QAAA,IAAY,IAAA;AACZ,QAAA,QAAA,EAAA;AAEA,QAAA,IAAI,QAAQ,OAAA,EAAS;AAErB,QAAA,IAAI,OAAA,GAAU,IAAA,CAAK,cAAA,IAAkB,QAAA,GAAW,KAAK,cAAA,EAAgB;AAIrE,QAAA,MAAM,KAAA,GAAS,OAAA,GAAU,OAAA,GAAW,OAAA,GAAW,WAAW,QAAA,GAAY,QAAA;AAEtE,QAAA,IAAI,QAAQ,SAAA,EAAW;AACrB,UAAA,SAAA,GAAY,KAAA;AACZ,UAAA,WAAA,GAAc,CAAA;AACd,UAAA,aAAA,GAAA,CAAiB,MAAM,OAAA,IAAW,CAAA;AAClC,UAAA,QAAA,GAAW,aAAA,CAAc,KAAA,CAAM,CAAA,EAAG,CAAA,GAAI,CAAC,CAAA;AACvC,UAAA,SAAA,GAAY,aAAA,CAAc,KAAA,CAAM,CAAA,GAAI,CAAC,CAAA;AAAA,QACvC;AAAA,MACF;AAAA,IACF;AAGA,IAAA,IAAI,cAAc,CAAA,QAAA,EAAW;AAC3B,MAAA,OAAO;AAAA,QACL,YAAA,EAAc,CAAA;AAAA,QACd,SAAA,EAAW,CAAA;AAAA,QACX,aAAa,EAAC;AAAA,QACd,cAAc;AAAC,OACjB;AAAA,IACF;AAEA,IAAA,OAAO;AAAA,MACL,YAAA,EAAc,WAAA;AAAA,MACd,SAAA,EAAW,aAAA;AAAA,MACX,WAAA,EAAa,QAAA;AAAA,MACb,YAAA,EAAc;AAAA,KAChB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,IAAA,EAAM;AAC9B,MAAA,MAAM,IAAI,eAAe,wDAAwD,CAAA;AAAA,IACnF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,uBAAuB,CAAA;AACrE,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAEvB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,MAAM,cAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MAC1D;AACA,MAAA,WAAA,CAAY,KAAK,IAAA,CAAK,aAAA,CAAc,MAAA,EAAQ,IAAA,CAAK,IAAI,CAAC,CAAA;AAAA,IACxD;AAEA,IAAA,OAAO,OAAO,WAAW,CAAA;AAAA,EAC3B;AAAA,EAEQ,aAAA,CAAc,QAAkB,IAAA,EAAwB;AAC9D,IAAA,IAAI,KAAK,MAAA,EAAQ;AACf,MAAA,OAAO,KAAK,UAAA,IAAc,CAAA;AAAA,IAC5B;AAEA,IAAA,MAAM,YAAA,GAAe,MAAA,CAAO,IAAA,CAAK,YAAA,IAAgB,CAAC,CAAA,IAAK,CAAA;AACvD,IAAA,IAAI,YAAA,KAAiB,IAAA,CAAK,SAAA,IAAa,CAAA,CAAA,EAAI;AACzC,MAAA,IAAI,CAAC,IAAA,CAAK,IAAA,EAAM,MAAM,IAAI,aAAa,kDAAkD,CAAA;AACzF,MAAA,OAAO,IAAA,CAAK,aAAA,CAAc,MAAA,EAAQ,IAAA,CAAK,IAAI,CAAA;AAAA,IAC7C,CAAA,MAAO;AACL,MAAA,IAAI,CAAC,IAAA,CAAK,KAAA,EAAO,MAAM,IAAI,aAAa,mDAAmD,CAAA;AAC3F,MAAA,OAAO,IAAA,CAAK,aAAA,CAAc,MAAA,EAAQ,IAAA,CAAK,KAAK,CAAA;AAAA,IAC9C;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAcA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AAGA,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,IACtC;AACA,IAAA,KAAA,IAAS,CAAA,CAAE,IAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,QAAQ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AACzC,MAAA,MAAM,QAAQ,MAAA,CAAO,WAAA,CAAY,KAAK,WAAA,CAAY,MAAA,GAAS,CAAC,CAAC,CAAA;AAC7D,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAC5B,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAAA,IAC9B;AAEA,IAAA,OAAO,UAAU,CAAA,GAAK,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA,GAAO,IAAI,KAAA,GAAQ,KAAA;AAAA,EAC/D;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,UAAU,IAAA,CAAK,QAAA;AAAA,MACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,MACtB,gBAAgB,IAAA,CAAK,cAAA;AAAA,MACrB,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,aAAa,IAAA,CAAK;AAAA,KACpB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AACF;;;AC35BO,IAAM,4BAAN,MAAqD;AAAA;AAAA,EAElD,WAAA;AAAA;AAAA,EAGA,YAAA;AAAA;AAAA,EAGA,QAAA;AAAA;AAAA,EAGA,eAAA;AAAA;AAAA,EAGA,aAAsC,EAAC;AAAA;AAAA,EAGvC,cAAA,GAAiB,CAAA;AAAA;AAAA,EAGjB,SAAA,GAAY,CAAA;AAAA;AAAA,EAGZ,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAKI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,GAAA;AAC1C,IAAA,IAAA,CAAK,YAAA,GAAe,QAAQ,YAAA,IAAgB,GAAA;AAC5C,IAAA,IAAA,CAAK,QAAA,GAAW,QAAQ,QAAA,IAAY,CAAA;AACpC,IAAA,IAAA,CAAK,eAAA,GAAkB,QAAQ,eAAA,IAAmB,CAAA;AAElD,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,eAAe,CAAA,EAAG;AAChE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,wCAAA;AAAA,QACA,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,YAAY,CAAA,IAAK,IAAA,CAAK,gBAAgB,CAAA,EAAG;AACjE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,+BAAA;AAAA,QACA,cAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,QAAQ,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACzD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,kCAAA;AAAA,QACA,UAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,kBAAkB,CAAA,EAAG;AACvE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,yCAAA;AAAA,QACA,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAeA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAEtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAEjB,IAAA,MAAM,QAAkB,EAAC;AACzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAGA,IAAA,IAAA,CAAK,cAAA,GAAiB,MAAM,MAAA,CAAO,CAAC,KAAK,GAAA,KAAQ,GAAA,GAAM,GAAA,EAAK,CAAC,CAAA,GAAI,QAAA;AAGjE,IAAA,MAAM,cAAc,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,IAAA,CAAK,KAAK,cAAc,CAAA;AAGxE,IAAA,IAAA,CAAK,aAAa,EAAC;AAEnB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AAEzC,MAAA,MAAM,YAAsB,EAAC;AAC7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,SAAA,CAAU,IAAA,CAAA,CAAM,MAAM,CAAC,CAAA,IAAK,MAAM,WAAA,CAAY,CAAC,KAAK,CAAA,CAAE,CAAA;AAAA,MACxD;AAGA,MAAA,MAAM,IAAA,GAAO,IAAI,qBAAA,CAAsB;AAAA,QACrC,UAAU,IAAA,CAAK,QAAA;AAAA,QACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,QACtB,cAAA,EAAgB;AAAA,OACjB,CAAA;AACD,MAAA,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,MAAA,CAAO,SAAS,CAAC,CAAA;AAC7B,MAAA,IAAA,CAAK,UAAA,CAAW,KAAK,IAAI,CAAA;AAGzB,MAAA,MAAM,QAAA,GAAW,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,WAAA,CAAY,CAAC,CAAA,GAAA,CACV,WAAA,CAAY,CAAC,KAAK,CAAA,IAAK,IAAA,CAAK,YAAA,GAAe,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,QAAA,CAAS,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,MACzF;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,4DAA4D,CAAA;AAAA,IACvF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,2BAA2B,CAAA;AAEzE,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,cAAc,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,IAAA,CAAK,KAAK,cAAc,CAAA;AAExE,IAAA,KAAA,MAAW,IAAA,IAAQ,KAAK,UAAA,EAAY;AAClC,MAAA,MAAM,QAAA,GAAW,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,WAAA,CAAY,CAAC,CAAA,GAAA,CACV,WAAA,CAAY,CAAC,KAAK,CAAA,IAAK,IAAA,CAAK,YAAA,GAAe,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,QAAA,CAAS,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,MACzF;AAAA,IACF;AAEA,IAAA,OAAO,OAAO,WAAW,CAAA;AAAA,EAC3B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AAEA,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,IACtC;AACA,IAAA,KAAA,IAAS,CAAA,CAAE,IAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,QAAQ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AACzC,MAAA,MAAM,QAAQ,MAAA,CAAO,WAAA,CAAY,KAAK,WAAA,CAAY,MAAA,GAAS,CAAC,CAAC,CAAA;AAC7D,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAC5B,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAAA,IAC9B;AAEA,IAAA,OAAO,UAAU,CAAA,GAAK,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA,GAAO,IAAI,KAAA,GAAQ,KAAA;AAAA,EAC/D;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,cAAc,IAAA,CAAK,YAAA;AAAA,MACnB,UAAU,IAAA,CAAK,QAAA;AAAA,MACf,iBAAiB,IAAA,CAAK;AAAA,KACxB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AACF;AAuBO,IAAM,6BAAN,MAAuD;AAAA;AAAA,EAEpD,WAAA;AAAA;AAAA,EAGA,YAAA;AAAA;AAAA,EAGA,QAAA;AAAA;AAAA,EAGA,eAAA;AAAA;AAAA,EAGA,aAAsC,EAAC;AAAA;AAAA,EAGvC,cAAA,GAAiB,CAAA;AAAA;AAAA,EAGjB,SAAA,GAAY,CAAA;AAAA;AAAA,EAGZ,cAAwB,EAAC;AAAA;AAAA,EAGzB,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAKI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,GAAA;AAC1C,IAAA,IAAA,CAAK,YAAA,GAAe,QAAQ,YAAA,IAAgB,GAAA;AAC5C,IAAA,IAAA,CAAK,QAAA,GAAW,QAAQ,QAAA,IAAY,CAAA;AACpC,IAAA,IAAA,CAAK,eAAA,GAAkB,QAAQ,eAAA,IAAmB,CAAA;AAElD,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,eAAe,CAAA,EAAG;AAChE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,wCAAA;AAAA,QACA,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,YAAY,CAAA,IAAK,IAAA,CAAK,gBAAgB,CAAA,EAAG;AACjE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,+BAAA;AAAA,QACA,cAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,QAAQ,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACzD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,kCAAA;AAAA,QACA,UAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,kBAAkB,CAAA,EAAG;AACvE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,yCAAA;AAAA,QACA,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAgBA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAEtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAEjB,IAAA,MAAM,QAAkB,EAAC;AACzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAGA,IAAA,IAAA,CAAK,WAAA,GAAc,CAAC,GAAG,IAAI,GAAA,CAAI,KAAK,CAAC,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,KAAM,IAAI,CAAC,CAAA;AAC3D,IAAA,IAAI,IAAA,CAAK,WAAA,CAAY,MAAA,KAAW,CAAA,EAAG;AACjC,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,uDAAA;AAAA,QACA,GAAA;AAAA,QACA,KAAK,WAAA,CAAY;AAAA,OACnB;AAAA,IACF;AAGA,IAAA,MAAM,OAAA,GAAU,KAAA,CAAM,GAAA,CAAI,CAAC,KAAA,KAAW,KAAA,KAAU,IAAA,CAAK,WAAA,CAAY,CAAC,CAAA,GAAI,CAAA,GAAI,CAAE,CAAA;AAG5E,IAAA,MAAM,WAAW,OAAA,CAAQ,MAAA,CAAO,CAAC,CAAA,KAAM,CAAA,KAAM,CAAC,CAAA,CAAE,MAAA;AAChD,IAAA,MAAM,WAAW,QAAA,GAAW,QAAA;AAC5B,IAAA,IAAA,CAAK,iBAAiB,IAAA,CAAK,GAAA,CAAA,CAAK,QAAA,GAAW,CAAA,KAAM,WAAW,CAAA,CAAE,CAAA;AAG9D,IAAA,MAAM,YAAY,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,IAAA,CAAK,KAAK,cAAc,CAAA;AAGtE,IAAA,IAAA,CAAK,aAAa,EAAC;AAEnB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AAEzC,MAAA,MAAM,YAAsB,EAAC;AAC7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,IAAA,GAAO,KAAK,CAAA,GAAI,IAAA,CAAK,IAAI,EAAE,SAAA,CAAU,CAAC,CAAA,IAAK,CAAA,CAAE,CAAA,CAAA;AACnD,QAAA,MAAM,GAAA,GAAM,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA;AAC1B,QAAA,SAAA,CAAU,IAAA,CAAK,MAAM,IAAI,CAAA;AAAA,MAC3B;AAGA,MAAA,MAAM,IAAA,GAAO,IAAI,qBAAA,CAAsB;AAAA,QACrC,UAAU,IAAA,CAAK,QAAA;AAAA,QACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,QACtB,cAAA,EAAgB;AAAA,OACjB,CAAA;AACD,MAAA,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,MAAA,CAAO,SAAS,CAAC,CAAA;AAC7B,MAAA,IAAA,CAAK,UAAA,CAAW,KAAK,IAAI,CAAA;AAGzB,MAAA,MAAM,QAAA,GAAW,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,SAAA,CAAU,CAAC,CAAA,GAAA,CACR,SAAA,CAAU,CAAC,KAAK,CAAA,IAAK,IAAA,CAAK,YAAA,GAAe,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,QAAA,CAAS,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,MACvF;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,6DAA6D,CAAA;AAAA,IACxF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,4BAA4B,CAAA;AAE1E,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAE/B,IAAA,MAAM,YAAY,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,IAAA,CAAK,KAAK,cAAc,CAAA;AACtE,IAAA,KAAA,MAAW,IAAA,IAAQ,KAAK,UAAA,EAAY;AAClC,MAAA,MAAM,QAAA,GAAW,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,SAAA,CAAU,CAAC,CAAA,GAAA,CACR,SAAA,CAAU,CAAC,KAAK,CAAA,IAAK,IAAA,CAAK,YAAA,GAAe,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,QAAA,CAAS,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,MACvF;AAAA,IACF;AAEA,IAAA,MAAM,cAAwB,EAAC;AAC/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,IAAA,GAAO,KAAK,CAAA,GAAI,IAAA,CAAK,IAAI,EAAE,SAAA,CAAU,CAAC,CAAA,IAAK,CAAA,CAAE,CAAA,CAAA;AACnD,MAAA,MAAM,cAAA,GAAiB,QAAQ,GAAA,GAAM,IAAA,CAAK,YAAY,CAAC,CAAA,GAAI,IAAA,CAAK,WAAA,CAAY,CAAC,CAAA;AAC7E,MAAA,WAAA,CAAY,IAAA,CAAK,kBAAkB,CAAC,CAAA;AAAA,IACtC;AAEA,IAAA,OAAO,MAAA,CAAO,WAAA,EAAa,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EAC/C;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAcA,aAAa,CAAA,EAAmB;AAC9B,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,6DAA6D,CAAA;AAAA,IACxF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,4BAA4B,CAAA;AAE1E,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,YAAY,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,IAAA,CAAK,KAAK,cAAc,CAAA;AACtE,IAAA,KAAA,MAAW,IAAA,IAAQ,KAAK,UAAA,EAAY;AAClC,MAAA,MAAM,QAAA,GAAW,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,SAAA,CAAU,CAAC,CAAA,GAAA,CACR,SAAA,CAAU,CAAC,KAAK,CAAA,IAAK,IAAA,CAAK,YAAA,GAAe,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,QAAA,CAAS,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,MACvF;AAAA,IACF;AAEA,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,IAAA,GAAO,KAAK,CAAA,GAAI,IAAA,CAAK,IAAI,EAAE,SAAA,CAAU,CAAC,CAAA,IAAK,CAAA,CAAE,CAAA,CAAA;AACnD,MAAA,KAAA,CAAM,IAAA,CAAK,CAAC,CAAA,GAAI,IAAA,EAAM,IAAI,CAAC,CAAA;AAAA,IAC7B;AAEA,IAAA,OAAO,OAAO,KAAK,CAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AACA,IAAA,IAAI,OAAA,GAAU,CAAA;AACd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,IAAI,MAAA,CAAO,WAAA,CAAY,IAAA,CAAK,WAAA,CAAY,SAAS,CAAC,CAAC,CAAA,KAAM,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA,EAAG;AACrF,QAAA,OAAA,EAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,OAAO,UAAU,CAAA,CAAE,IAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,cAAc,IAAA,CAAK,YAAA;AAAA,MACnB,UAAU,IAAA,CAAK,QAAA;AAAA,MACf,iBAAiB,IAAA,CAAK;AAAA,KACxB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AACF;;;AC5iBO,IAAM,QAAN,MAAiC;AAAA;AAAA,EAE9B,OAAA;AAAA;AAAA,EAaA,KAAA;AAAA;AAAA,EAGA,UAAA,GAAa,CAAA;AAAA;AAAA,EAGb,YAAA;AAAA;AAAA,EAGA,MAAA;AAAA;AAAA,EAGA,MAAA,GAAS,KAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAejB,WAAA,CACE,OAAA,GAUI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,OAAA,GAAU,EAAE,GAAG,OAAA,EAAQ;AAC5B,IAAA,IAAI,IAAA,CAAK,OAAA,CAAQ,WAAA,KAAgB,MAAA,IAAa,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,OAAA,CAAQ,WAAW,CAAA,EAAG;AACxF,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,IAAA,CAAK,OAAA,CAAQ,WAAW,CAAC,CAAA,CAAA;AAAA,QACjF,aAAA;AAAA,QACA,KAAK,OAAA,CAAQ;AAAA,OACf;AAAA,IACF;AAAA,EACF;AAAA,EAEQ,SAAA,GAA0B;AAChC,IAAA,IAAI,IAAA,CAAK,OAAA,CAAQ,WAAA,KAAgB,MAAA,EAAW;AAC1C,MAAA,IAAI,IAAA,GAAO,KAAK,OAAA,CAAQ,WAAA;AACxB,MAAA,OAAO,MAAM;AACX,QAAA,IAAA,GAAA,CAAQ,IAAA,GAAO,OAAO,KAAA,IAAS,MAAA;AAC/B,QAAA,OAAO,IAAA,GAAO,MAAA;AAAA,MAChB,CAAA;AAAA,IACF;AACA,IAAA,OAAO,IAAA,CAAK,MAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EA6BA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAE9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AACtB,IAAA,IAAA,CAAK,MAAA,GAAS,MAAA;AAGd,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,KAAA,IAAS,CAAA;AACpC,IAAA,IAAI,EAAE,SAAS,CAAA,CAAA,EAAI;AACjB,MAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,6BAAA,EAAgC,KAAK,CAAA,CAAA,EAAI,SAAS,KAAK,CAAA;AAAA,IACzF;AAGA,IAAA,MAAM,OAAA,GAAU,IAAA,CAAK,OAAA,CAAQ,OAAA,IAAW,GAAA;AACxC,IAAA,MAAM,GAAA,GAAM,IAAA,CAAK,OAAA,CAAQ,GAAA,IAAO,IAAA;AAChC,IAAA,MAAM,YAAA,GAAe,IAAA,CAAK,OAAA,CAAQ,YAAA,IAAgB,IAAA;AAClD,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,OAAA,CAAQ,SAAA,IAAa,KAAA;AAC5C,IAAA,MAAM,QAAA,GAAW,IAAA,CAAK,OAAA,CAAQ,QAAA,IAAY,KAAA;AAC1C,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,OAAA,CAAQ,SAAA,IAAa,QAAA;AAC5C,IAAA,MAAM,GAAA,GAAM,KAAK,SAAA,EAAU;AAG3B,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAGxB,IAAA,IAAA,CAAK,YAAA,GAAe,CAAA;AAIpB,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,MAAM,QAAQ,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAEzC,IAAA,IAAI,YAAA,EAAc;AAEhB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAAA,MAC3C;AAGA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA;AAC/B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,KAAA,CAAM,CAAC,CAAA,GAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,CAAA,IAAK,CAAC,CAAA;AAAA,QAC9D;AAAA,MACF;AAGA,MAAA,MAAMA,KAAAA,GAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA;AAC/B,MAAA,KAAA,IAASA,KAAAA;AACT,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,KAAA,CAAM,CAAC,CAAA,GAAA,CAAK,KAAA,CAAM,CAAC,KAAK,CAAA,IAAKA,KAAAA;AAAA,MAC/B;AAAA,IACF;AAEA,IAAA,IAAI,MAAA;AACJ,IAAA,IAAI,SAAA,EAAW;AACb,MAAA,MAAA,GAAS,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACpC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA;AAC/B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,MAAM,QAAA,GAAW,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,CAAA,IAAK,CAAC,CAAA,IAAK,YAAA,GAAgB,KAAA,CAAM,CAAC,KAAK,CAAA,GAAK,CAAA,CAAA;AACtF,UAAA,MAAA,CAAO,CAAC,CAAA,GAAA,CAAK,MAAA,CAAO,CAAC,CAAA,IAAK,KAAK,QAAA,GAAW,QAAA;AAAA,QAC5C;AAAA,MACF;AACA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAA,CAAO,CAAC,CAAA,GAAI,IAAA,CAAK,KAAK,MAAA,CAAO,CAAC,KAAK,CAAC,CAAA;AAAA,MACtC;AAAA,IACF;AAEA,IAAA,MAAM,IAAA,GAAO,CAAC,WAAA,EAAqB,YAAA,KAAiC;AAClE,MAAA,MAAM,GAAA,GAAM,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,WAAA,GAAc,CAAA,GAAI,YAAY,CAAA,IAAK,CAAC,CAAA;AACzE,MAAA,MAAM,WAAW,GAAA,IAAO,YAAA,GAAgB,KAAA,CAAM,YAAY,KAAK,CAAA,GAAK,CAAA,CAAA;AACpE,MAAA,IAAI,aAAa,MAAA,EAAQ;AACvB,QAAA,MAAM,CAAA,GAAI,MAAA,CAAO,YAAY,CAAA,IAAK,CAAA;AAClC,QAAA,OAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,QAAA,GAAW,CAAA;AAAA,MAClC;AACA,MAAA,OAAO,QAAA;AAAA,IACT,CAAA;AAKA,IAAA,MAAM,WAAW,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAC5C,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,IAAI,CAAA,GAAI,CAAA;AACR,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,GAAA,GAAM,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACrB,QAAA,CAAA,IAAK,GAAA,GAAM,GAAA;AAAA,MACb;AAEA,MAAA,QAAA,CAAS,CAAC,CAAA,GAAI,CAAA,KAAM,CAAA,GAAI,IAAI,CAAA,GAAI,CAAA;AAAA,IAClC;AAKA,IAAA,MAAM,IAAI,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACrC,IAAA,IAAI,IAAA,CAAK,OAAA,CAAQ,SAAA,IAAa,IAAA,CAAK,KAAA,IAAS,IAAA,CAAK,KAAA,CAAM,IAAA,KAAS,CAAA,IAAK,IAAA,CAAK,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG;AAE1F,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,CAAA,CAAE,CAAC,CAAA,GAAI,MAAA,CAAO,IAAA,CAAK,KAAA,CAAM,IAAA,CAAK,IAAA,CAAK,KAAA,CAAM,MAAA,GAAS,CAAC,CAAA,IAAK,CAAC,CAAA;AAAA,MAC3D;AAAA,IACF;AAKA,IAAA,MAAM,OAAO,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACxC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,IAAI,IAAA,GAAO,CAAA;AACX,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAA,IAAQ,KAAK,CAAA,EAAG,CAAC,CAAA,IAAK,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,MAChC;AACA,MAAA,IAAA,CAAK,CAAC,CAAA,GAAI,IAAA;AAAA,IACZ;AAGA,IAAA,MAAM,IAAA,GAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA;AAI/B,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,OAAA,EAAS,IAAA,EAAA,EAAQ;AAEzC,MAAA,IAAI,SAAA,GAAY,CAAA;AAKhB,MAAA,IAAI,OAAA,GAA2B,IAAA;AAC/B,MAAA,IAAI,cAAc,QAAA,EAAU;AAC1B,QAAA,OAAA,GAAU,KAAA,CAAM,KAAK,EAAE,MAAA,EAAQ,GAAE,EAAG,CAAC,CAAA,EAAG,CAAA,KAAM,CAAC,CAAA;AAE/C,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,GAAG,CAAA,EAAA,EAAK;AAC9B,UAAA,MAAM,IAAI,IAAA,CAAK,KAAA,CAAM,GAAA,EAAI,IAAK,IAAI,CAAA,CAAE,CAAA;AACpC,UAAA,MAAM,GAAA,GAAM,QAAQ,CAAC,CAAA;AACrB,UAAA,OAAA,CAAQ,CAAC,CAAA,GAAI,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA;AAC3B,UAAA,OAAA,CAAQ,CAAC,IAAI,GAAA,IAAO,CAAA;AAAA,QACtB;AAAA,MACF;AAGA,MAAA,MAAM,SAAA,GAAY,OAAA,IAAW,KAAA,CAAM,IAAA,CAAK,EAAE,MAAA,EAAQ,CAAA,EAAE,EAAG,CAAC,CAAA,EAAG,CAAA,KAAM,CAAC,CAAA;AAClE,MAAA,KAAA,MAAW,KAAK,SAAA,EAAW;AACzB,QAAA,MAAM,KAAA,GAAQ,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA;AAG7B,QAAA,IAAI,UAAU,CAAA,EAAG;AACf,UAAA,MAAM,KAAA,GAAQ,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA;AACtB,UAAA,IAAI,UAAU,CAAA,EAAG;AACf,YAAA,MAAMC,SAAQ,CAAC,KAAA;AACf,YAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,cAAA,IAAA,CAAK,CAAC,KAAK,IAAA,CAAK,CAAC,KAAK,CAAA,IAAKA,MAAAA,GAAQ,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AAAA,YAC9C;AACA,YAAA,SAAA,GAAY,KAAK,GAAA,CAAI,SAAA,EAAW,IAAA,CAAK,GAAA,CAAIA,MAAK,CAAC,CAAA;AAAA,UACjD;AACA,UAAA,CAAA,CAAE,CAAC,CAAA,GAAI,CAAA;AACP,UAAA;AAAA,QACF;AAKA,QAAA,IAAI,GAAA,GAAM,CAAA;AACV,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,MAAM,GAAA,GAAM,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACrB,UAAA,MAAM,EAAA,GAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAC,CAAA,IAAK,YAAA,GAAe,KAAA,GAAQ,CAAA,CAAA;AAEvE,UAAA,MAAM,CAAA,GAAI,MAAM,IAAA,CAAK,CAAC,KAAK,CAAA,CAAA,GAAA,CAAM,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,IAAK,GAAA;AAC9C,UAAA,GAAA,IAAO,GAAA,GAAM,CAAA;AAAA,QACf;AACA,QAAA,GAAA,IAAO,IAAA;AAKP,QAAA,IAAI,IAAA,GAAO,IAAA,CAAK,aAAA,CAAc,GAAA,EAAK,KAAK,CAAA,GAAI,KAAA;AAG5C,QAAA,IAAI,QAAA,IAAY,OAAO,CAAA,EAAG;AACxB,UAAA,IAAA,GAAO,CAAA;AAAA,QACT;AAGA,QAAA,MAAM,KAAA,GAAQ,IAAA,IAAQ,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAK9B,QAAA,IAAI,UAAU,CAAA,EAAG;AACf,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,YAAA,IAAA,CAAK,CAAC,KAAK,IAAA,CAAK,CAAC,KAAK,CAAA,IAAK,KAAA,GAAQ,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AAAA,UAC9C;AAAA,QACF;AAGA,QAAA,CAAA,CAAE,CAAC,CAAA,GAAI,IAAA;AAGP,QAAA,SAAA,GAAY,KAAK,GAAA,CAAI,SAAA,EAAW,IAAA,CAAK,GAAA,CAAI,KAAK,CAAC,CAAA;AAAA,MACjD;AAGA,MAAA,IAAI,YAAY,GAAA,EAAK;AACnB,QAAA,IAAA,CAAK,SAAS,IAAA,GAAO,CAAA;AACrB,QAAA;AAAA,MACF;AAAA,IACF;AAGA,IAAA,IAAI,IAAA,CAAK,WAAW,MAAA,EAAW;AAC7B,MAAA,IAAA,CAAK,MAAA,GAAS,OAAA;AAAA,IAChB;AAGA,IAAA,IAAI,aAAa,MAAA,EAAQ;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,CAAA,GAAI,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA;AACvB,QAAA,CAAA,CAAE,CAAC,IAAI,CAAA,KAAM,CAAA,GAAI,KAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAK,CAAA;AAAA,MACrC;AAAA,IACF;AAGA,IAAA,IAAA,CAAK,KAAA,GAAQ,OAAO,CAAC,CAAA;AAKrB,IAAA,IAAI,YAAA,EAAc;AAChB,MAAA,IAAI,SAAA,GAAY,CAAA;AAChB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,SAAA,IAAA,CAAc,MAAM,CAAC,CAAA,IAAK,CAAA,KAAM,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,MAC1C;AACA,MAAA,IAAA,CAAK,aAAa,KAAA,GAAQ,SAAA;AAAA,IAC5B,CAAA,MAAO;AACL,MAAA,IAAA,CAAK,UAAA,GAAa,CAAA;AAAA,IACpB;AAGA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAqBQ,aAAA,CAAc,GAAW,MAAA,EAAwB;AAEvD,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,CAAC,KAAK,CAAC,MAAA,CAAO,QAAA,CAAS,MAAM,CAAA,EAAG;AACnD,MAAA,MAAM,IAAI,oBAAoB,uDAAuD,CAAA;AAAA,IACvF;AACA,IAAA,IAAI,CAAA,GAAI,MAAA,EAAQ,OAAO,CAAA,GAAI,MAAA;AAC3B,IAAA,IAAI,CAAA,GAAI,CAAC,MAAA,EAAQ,OAAO,CAAA,GAAI,MAAA;AAC5B,IAAA,OAAO,CAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAUA,IAAI,IAAA,GAAe;AACjB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,KAAA,EAAO;AAC/B,MAAA,MAAM,IAAI,eAAe,6CAA6C,CAAA;AAAA,IACxE;AACA,IAAA,OAAO,IAAA,CAAK,KAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,SAAA,GAAoB;AACtB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,0CAA0C,CAAA;AAAA,IACrE;AACA,IAAA,OAAO,IAAA,CAAK,UAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,KAAA,GAA4B;AAC9B,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,sCAAsC,CAAA;AAAA,IACjE;AACA,IAAA,OAAO,IAAA,CAAK,MAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAgBA,QAAQ,CAAA,EAAmB;AAEzB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,KAAA,EAAO;AAC/B,MAAA,MAAM,IAAI,eAAe,wCAAwC,CAAA;AAAA,IACnE;AAGA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,OAAO,CAAA;AAExD,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,IAAA,GAAO,KAAA,CAAM,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAG5B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAE1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAA,CAAK,CAAC,KACJ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA,GAAI,CAAC,CAAA,IAAK,CAAC,CAAA,GACxC,MAAA,CAAO,KAAK,KAAA,CAAM,IAAA,CAAK,KAAK,KAAA,CAAM,MAAA,GAAS,CAAC,CAAA,IAAK,CAAC,CAAA;AAAA,MACtD;AAEA,MAAA,IAAA,CAAK,CAAC,KAAK,IAAA,CAAK,UAAA;AAAA,IAClB;AAEA,IAAA,OAAO,OAAO,IAAI,CAAA;AAAA,EACpB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,qCAAqC,CAAA;AAAA,IAChE;AACA,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,IAAA,GAAO,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC3B,IAAA,IAAI,IAAA,CAAK,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACxB,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,IAAA,CAAK,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OAC/E;AAAA,IACF;AACA,IAAA,IAAI,KAAA,GAAQ,GACV,KAAA,GAAQ,CAAA;AACV,IAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAAA,IAC3C;AACA,IAAA,KAAA,IAAS,CAAA,CAAE,IAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,IAAA,GAAO,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAC7C,MAAA,MAAM,OAAA,GAAU,OAAO,IAAA,CAAK,IAAA,CAAK,KAAK,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AACtD,MAAA,KAAA,IAAA,CAAU,OAAO,OAAA,KAAY,CAAA;AAC7B,MAAA,KAAA,IAAA,CAAU,OAAO,KAAA,KAAU,CAAA;AAAA,IAC7B;AAEA,IAAA,IAAI,UAAU,CAAA,EAAG;AACf,MAAA,OAAO,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA;AAAA,IAC7B;AAEA,IAAA,OAAO,IAAI,KAAA,GAAQ,KAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO,EAAE,GAAG,IAAA,CAAK,OAAA,EAAQ;AAAA,EAC3B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,UAAU,MAAA,EAAuC;AAC/C,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,KAAK,KAAK,MAAA,CAAO,OAAA,CAAQ,MAAM,CAAA,EAAG;AACjD,MAAA,QAAQ,GAAA;AAAK,QACX,KAAK,OAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AACxD,YAAA,MAAM,IAAI,qBAAA,CAAsB,+BAAA,EAAiC,OAAA,EAAS,KAAK,CAAA;AAAA,UACjF;AACA,UAAA,IAAA,CAAK,QAAQ,KAAA,GAAQ,KAAA;AACrB,UAAA;AAAA,QACF,KAAK,SAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AACxD,YAAA,MAAM,IAAI,qBAAA,CAAsB,iCAAA,EAAmC,SAAA,EAAW,KAAK,CAAA;AAAA,UACrF;AACA,UAAA,IAAA,CAAK,QAAQ,OAAA,GAAU,KAAA;AACvB,UAAA;AAAA,QACF,KAAK,KAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AACxD,YAAA,MAAM,IAAI,qBAAA,CAAsB,6BAAA,EAA+B,KAAA,EAAO,KAAK,CAAA;AAAA,UAC7E;AACA,UAAA,IAAA,CAAK,QAAQ,GAAA,GAAM,KAAA;AACnB,UAAA;AAAA,QACF,KAAK,cAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,gCAAA;AAAA,cACA,cAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,YAAA,GAAe,KAAA;AAC5B,UAAA;AAAA,QACF,KAAK,WAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA,CAAsB,6BAAA,EAA+B,WAAA,EAAa,KAAK,CAAA;AAAA,UACnF;AACA,UAAA,IAAA,CAAK,QAAQ,SAAA,GAAY,KAAA;AACzB,UAAA;AAAA,QACF,KAAK,WAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA,CAAsB,6BAAA,EAA+B,WAAA,EAAa,KAAK,CAAA;AAAA,UACnF;AACA,UAAA,IAAA,CAAK,QAAQ,SAAA,GAAY,KAAA;AACzB,UAAA;AAAA,QACF,KAAK,UAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA,CAAsB,4BAAA,EAA8B,UAAA,EAAY,KAAK,CAAA;AAAA,UACjF;AACA,UAAA,IAAA,CAAK,QAAQ,QAAA,GAAW,KAAA;AACxB,UAAA;AAAA,QACF,KAAK,WAAA;AACH,UAAA,IAAI,KAAA,KAAU,QAAA,IAAY,KAAA,KAAU,QAAA,EAAU;AAC5C,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,mBAAA,EAAsB,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACnC,WAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,SAAA,GAAY,KAAA;AACzB,UAAA;AAAA,QACF,KAAK,aAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AACxD,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cAC9D,aAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,WAAA,GAAc,KAAA;AAC3B,UAAA;AAAA,QACF;AACE,UAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,mBAAA,EAAsB,GAAG,CAAA,CAAA,EAAI,KAAK,KAAK,CAAA;AAAA;AAC3E,IACF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AACF;;;ACllBO,IAAM,mBAAN,MAA4C;AAAA;AAAA,EAEzC,KAAA;AAAA;AAAA,EAGA,UAAA;AAAA;AAAA,EAGA,YAAA;AAAA;AAAA,EAGA,MAAA,GAAS,KAAA;AAAA,EAET,OAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAcR,WAAA,CACE,OAAA,GAII,EAAC,EACL;AACA,IAAA,IAAA,CAAK,OAAA,GAAU,EAAE,GAAG,OAAA,EAAQ;AAAA,EAC9B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAkBA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAE9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAGtB,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAA,CAAK,YAAA,GAAe,SAAA;AAIpB,IAAA,MAAM,YAAA,GAAe,IAAA,CAAK,OAAA,CAAQ,YAAA,IAAgB,IAAA;AAClD,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,KAAA,IAAS,IAAA;AACpC,IAAA,MAAM,eAAe,KAAA,KAAU,KAAA,KAAU,EAAE,KAAA,KAAU,SAAA,IAAa,EAAE,KAAA,KAAU,SAAA,CAAA;AAE9E,IAAA,IAAI,YAAA,EAAc;AAEhB,MAAA,MAAM,MAAA,GAAS,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AAExB,MAAA,MAAM,MAAA,GAAS,KAAK,CAAC,CAAA;AAGrB,MAAA,IAAI,WAAA,GAAc,YAAA,GACd,IAAA,CAAK,iBAAA,CAAkB,CAAA,EAAG,MAAM,CAAA,GAChC,IAAA,CAAK,UAAA,CAAW,CAAA,EAAG,MAAM,CAAA;AAC7B,MAAA,MAAM,UAAA,GAAa,IAAA,CAAK,UAAA,CAAW,CAAA,EAAG,MAAM,CAAA;AAG5C,MAAA,IAAI,OAAA;AACJ,MAAA,IAAI,IAAA,CAAK,QAAQ,SAAA,EAAW;AAE1B,QAAA,OAAA,GAAU,IAAA,CAAK,cAAc,WAAW,CAAA;AAExC,QAAA,WAAA,GAAc,YAAA,GACV,KAAK,gBAAA,CAAiB,WAAA,EAAa,OAAO,CAAA,GAC1C,IAAA,CAAK,SAAA,CAAU,WAAA,EAAa,OAAO,CAAA;AAAA,MACzC;AAGA,MAAA,MAAM,MAAA,GAAS,KAAA,CAAM,WAAA,EAAa,UAAU,CAAA;AAC5C,MAAA,IAAI,IAAI,MAAA,CAAO,CAAA;AAGf,MAAA,IAAI,OAAA,EAAS;AACX,QAAA,CAAA,GAAI,IAAA,CAAK,YAAA,CAAa,CAAA,EAAG,OAAO,CAAA;AAAA,MAClC;AACA,MAAA,IAAA,CAAK,KAAA,GAAQ,CAAA;AAGb,MAAA,MAAM,eAAA,GAAkB,GAAA,CAAI,MAAA,EAAQ,IAAA,CAAK,KAAK,CAAA;AAC9C,MAAA,IAAA,CAAK,UAAA,GAAa,GAAA,CAAI,MAAA,EAAQ,eAAe,CAAA;AAAA,IAC/C,CAAA,MAAO;AAEL,MAAA,MAAM,MAAA,GAAS,KAAA,CAAM,CAAA,EAAG,CAAC,CAAA;AACzB,MAAA,IAAA,CAAK,QAAQ,MAAA,CAAO,CAAA;AAAA,IAEtB;AAGA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASQ,UAAA,CAAW,MAAc,QAAA,EAA0B;AAEzD,IAAA,IAAI,IAAA,CAAK,SAAS,CAAA,EAAG;AACnB,MAAA,MAAM,IAAI,IAAA,CAAK,IAAA;AACf,MAAA,MAAM,UAAU,MAAA,CAAO,QAAA,CAAS,KAAK,QAAA,CAAS,MAAM,KAAK,CAAC,CAAA;AAC1D,MAAA,MAAM,WAAqB,EAAC;AAC5B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,QAAA,CAAS,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,IAAA,CAAK,IAAA,CAAK,SAAS,CAAC,CAAA,IAAK,CAAC,CAAA,GAAI,OAAO,CAAA;AAAA,MACjE;AACA,MAAA,OAAO,OAAO,QAAQ,CAAA;AAAA,IACxB;AAGA,IAAA,MAAM,QAAA,GAAW,IAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAClC,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEnC,IAAA,MAAM,SAAqB,EAAC;AAE5B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,GAAA,GAAM,MAAA,CAAO,IAAA,CAAK,IAAA,CAAK,IAAA,CAAK,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAA,IAAK,CAAC,CAAA;AAClE,QAAA,MAAM,OAAA,GAAU,OAAO,QAAA,CAAS,IAAA,CAAK,SAAS,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAC9D,QAAA,GAAA,CAAI,IAAA,CAAK,MAAM,OAAO,CAAA;AAAA,MACxB;AACA,MAAA,MAAA,CAAO,KAAK,GAAG,CAAA;AAAA,IACjB;AAEA,IAAA,OAAO,OAAO,MAAM,CAAA;AAAA,EACtB;AAAA,EAEQ,iBAAA,CAAkB,MAAc,QAAA,EAA0B;AAChE,IAAA,IAAI,IAAA,CAAK,SAAS,CAAA,EAAG;AACnB,MAAA,MAAM,IAAI,IAAA,CAAK,IAAA;AACf,MAAA,MAAM,UAAU,MAAA,CAAO,QAAA,CAAS,KAAK,QAAA,CAAS,MAAM,KAAK,CAAC,CAAA;AAC1D,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,GAAA,GAAM,KAAK,MAAA,GAAS,CAAA;AAC1B,QAAA,IAAA,CAAK,IAAA,CAAK,GAAG,CAAA,GAAI,MAAA,CAAO,KAAK,IAAA,CAAK,GAAG,CAAA,IAAK,CAAC,CAAA,GAAI,OAAA;AAAA,MACjD;AACA,MAAA,OAAO,IAAA;AAAA,IACT;AAEA,IAAA,MAAM,QAAA,GAAW,IAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAClC,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACnC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,OAAA,GAAU,IAAA,CAAK,MAAA,GAAS,CAAA,GAAI,SAAA;AAClC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,MAAM,OAAA,GAAU,CAAA;AACtB,QAAA,MAAM,OAAA,GAAU,OAAO,QAAA,CAAS,IAAA,CAAK,SAAS,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAC9D,QAAA,IAAA,CAAK,IAAA,CAAK,GAAG,CAAA,GAAI,MAAA,CAAO,KAAK,IAAA,CAAK,GAAG,CAAA,IAAK,CAAC,CAAA,GAAI,OAAA;AAAA,MACjD;AAAA,IACF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AAAA,EAEQ,cAAc,CAAA,EAAmB;AACvC,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,QAAkB,IAAI,KAAA,CAAM,SAAS,CAAA,CAAE,KAAK,CAAC,CAAA;AAEnD,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,GAAA,GAAM,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAA,IAAK,CAAC,CAAA;AAC5D,QAAA,KAAA,IAAS,GAAA,GAAM,GAAA;AAAA,MACjB;AACA,MAAA,KAAA,CAAM,CAAC,CAAA,GAAI,IAAA,CAAK,IAAA,CAAK,KAAK,CAAA;AAAA,IAC5B;AACA,IAAA,OAAO,OAAO,KAAK,CAAA;AAAA,EACrB;AAAA,EAEQ,SAAA,CAAU,GAAW,KAAA,EAAuB;AAClD,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,SAAqB,EAAC;AAE5B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,GAAA,GAAM,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAA,IAAK,CAAC,CAAA;AAC5D,QAAA,MAAM,CAAA,GAAI,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAClD,QAAA,GAAA,CAAI,IAAA,CAAK,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,MAAM,CAAC,CAAA;AAAA,MAChC;AACA,MAAA,MAAA,CAAO,KAAK,GAAG,CAAA;AAAA,IACjB;AACA,IAAA,OAAO,OAAO,MAAM,CAAA;AAAA,EACtB;AAAA,EAEQ,gBAAA,CAAiB,GAAW,KAAA,EAAuB;AACzD,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,SAAA;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,MAAM,OAAA,GAAU,CAAA;AACtB,QAAA,MAAM,CAAA,GAAI,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAClD,QAAA,CAAA,CAAE,IAAA,CAAK,GAAG,CAAA,GAAI,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,GAAG,CAAA,IAAK,CAAC,CAAA,GAAI,CAAA;AAAA,MACzD;AAAA,IACF;AACA,IAAA,OAAO,CAAA;AAAA,EACT;AAAA,EAEQ,YAAA,CAAa,MAAc,KAAA,EAAuB;AAGxD,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACnC,IAAA,MAAM,QAAA,GAAW,KAAK,IAAA,GAAO,CAAA,GAAK,KAAK,KAAA,CAAM,CAAC,KAAK,CAAA,GAAK,CAAA;AAExD,IAAA,IAAI,IAAA,CAAK,SAAS,CAAA,EAAG;AACnB,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,CAAA,GAAI,OAAO,IAAA,CAAK,IAAA,CAAK,KAAK,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAChD,QAAA,MAAM,CAAA,GAAI,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAClD,QAAA,GAAA,CAAI,IAAA,CAAK,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,IAAI,CAAC,CAAA;AAAA,MAC9B;AACA,MAAA,OAAO,OAAO,GAAG,CAAA;AAAA,IACnB;AAEA,IAAA,MAAM,SAAqB,EAAC;AAC5B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,MAAM,CAAA,GAAI,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAClD,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,CAAA,GAAI,MAAA,CAAO,IAAA,CAAK,IAAA,CAAK,IAAA,CAAK,SAAS,CAAA,GAAI,QAAA,GAAW,CAAC,CAAA,IAAK,CAAC,CAAA;AAC/D,QAAA,GAAA,CAAI,IAAA,CAAK,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,IAAI,CAAC,CAAA;AAAA,MAC9B;AACA,MAAA,MAAA,CAAO,KAAK,GAAG,CAAA;AAAA,IACjB;AACA,IAAA,OAAO,OAAO,MAAM,CAAA;AAAA,EACtB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,KAAA,EAAO;AAC/B,MAAA,MAAM,IAAI,eAAe,mDAAmD,CAAA;AAAA,IAC9E;AAGA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,kBAAkB,CAAA;AAGnE,IAAA,MAAM,UAAA,GAAa,GAAA,CAAI,CAAA,EAAG,IAAA,CAAK,KAAK,CAAA;AAGpC,IAAA,IAAI,IAAA,CAAK,eAAe,MAAA,EAAW;AAEjC,MAAA,MAAM,YAAA,GAAe,OAAO,IAAA,CAAK,UAAA,CAAW,KAAK,IAAA,CAAK,UAAA,CAAW,MAAM,CAAA,IAAK,CAAC,CAAA;AAC7E,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,UAAA,CAAW,MAAM,CAAA,EAAA,EAAK;AACxC,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,UAAA,CAAW,IAAA,CAAK,UAAA,CAAW,SAAS,CAAC,CAAA,IAAK,CAAC,CAAA,GAAI,YAAY,CAAA;AAAA,MAChF;AACA,MAAA,OAAO,OAAO,MAAM,CAAA;AAAA,IACtB;AAEA,IAAA,OAAO,UAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAoBA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,gDAAgD,CAAA;AAAA,IAC3E;AACA,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AAGA,IAAA,MAAM,MAAA,GAAS,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC7B,IAAA,IAAI,MAAA,CAAO,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC1B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,MAAA,CAAO,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACjF;AAAA,IACF;AAIA,IAAA,IAAI,MAAA,GAAS,CAAA;AACb,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,OAAO,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAC,CAAA,IAAK,CAAC,CAAA,GAAI,OAAO,MAAA,CAAO,IAAA,CAAK,OAAO,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAC3F,MAAA,MAAA,IAAU,IAAA,GAAO,IAAA;AAAA,IACnB;AAGA,IAAA,MAAM,aAAA,GAAgB,KAAK,CAAC,CAAA;AAC5B,IAAA,MAAM,aAAa,MAAA,CAAO,aAAA,CAAc,KAAK,aAAA,CAAc,MAAM,KAAK,CAAC,CAAA;AACvE,IAAA,IAAI,MAAA,GAAS,CAAA;AACb,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,IAAA,GAAO,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAC,CAAA,GAAI,UAAA;AACjD,MAAA,MAAA,IAAU,IAAA,GAAO,IAAA;AAAA,IACnB;AAIA,IAAA,IAAI,WAAW,CAAA,EAAG;AAChB,MAAA,OAAO,MAAA,KAAW,IAAI,CAAA,GAAM,CAAA;AAAA,IAC9B;AAEA,IAAA,OAAO,IAAI,MAAA,GAAS,MAAA;AAAA,EACtB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,IAAA,GAAe;AACjB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,KAAA,EAAO;AAC/B,MAAA,MAAM,IAAI,eAAe,wDAAwD,CAAA;AAAA,IACnF;AACA,IAAA,OAAO,IAAA,CAAK,KAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,SAAA,GAAgC;AAClC,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,qDAAqD,CAAA;AAAA,IAChF;AACA,IAAA,OAAO,IAAA,CAAK,UAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,YAAA,EAAc,IAAA,CAAK,OAAA,CAAQ,YAAA,IAAgB,IAAA;AAAA,MAC3C,SAAA,EAAW,IAAA,CAAK,OAAA,CAAQ,SAAA,IAAa,KAAA;AAAA,MACrC,KAAA,EAAO,IAAA,CAAK,OAAA,CAAQ,KAAA,IAAS;AAAA,KAC/B;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,KAAK,KAAK,MAAA,CAAO,OAAA,CAAQ,OAAO,CAAA,EAAG;AAClD,MAAA,QAAQ,GAAA;AAAK,QACX,KAAK,cAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,yCAAA,EAA4C,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACzD,cAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,YAAA,GAAe,KAAA;AAC5B,UAAA;AAAA,QACF,KAAK,WAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,sCAAA,EAAyC,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACtD,WAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,SAAA,GAAY,KAAA;AACzB,UAAA;AAAA,QACF,KAAK,OAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,kCAAA,EAAqC,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cAClD,OAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,KAAA,GAAQ,KAAA;AACrB,UAAA;AAAA,QACF;AACE,UAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,mBAAA,EAAsB,GAAG,CAAA,CAAA,EAAI,KAAK,KAAK,CAAA;AAAA;AAC3E,IACF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AACF;;;ACnbO,IAAM,qBAAN,MAA+C;AAAA,EAC5C,OAAA;AAAA,EAUA,KAAA;AAAA;AAAA,EACA,UAAA;AAAA;AAAA,EACA,YAAA;AAAA,EACA,QAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA,EACT,WAAA,GAAc,KAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EActB,WAAA,CACE,OAAA,GAQI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,OAAA,GAAU,EAAE,GAAG,OAAA,EAAQ;AAE5B,IAAA,MAAM,OAAA,GAAU,IAAA,CAAK,OAAA,CAAQ,OAAA,IAAW,IAAA;AACxC,IAAA,IAAI,OAAA,KAAY,IAAA,IAAQ,OAAA,KAAY,MAAA,EAAQ;AAC1C,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,mDAAA,EAAsD,MAAA,CAAO,OAAO,CAAC,CAAA,CAAA;AAAA,QACrE,SAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AACA,IAAA,IAAA,CAAK,QAAQ,OAAA,GAAU,OAAA;AAEvB,IAAA,MAAM,UAAA,GAAa,IAAA,CAAK,OAAA,CAAQ,UAAA,IAAc,MAAA;AAC9C,IAAA,IAAI,UAAA,KAAe,KAAA,IAAS,UAAA,KAAe,MAAA,EAAQ;AACjD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,6CAAA,EAAgD,MAAA,CAAO,UAAU,CAAC,CAAA,CAAA;AAAA,QAClE,YAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AACA,IAAA,IAAA,CAAK,QAAQ,UAAA,GAAa,UAAA;AAE1B,IAAA,IAAI,KAAK,OAAA,CAAQ,CAAA,KAAM,UAAa,IAAA,CAAK,OAAA,CAAQ,KAAK,CAAA,EAAG;AACvD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,wBAAA,EAA2B,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,CAAA;AAAA,QACzC,GAAA;AAAA,QACA,KAAK,OAAA,CAAQ;AAAA,OACf;AAAA,IACF;AACA,IAAA,IACE,IAAA,CAAK,OAAA,CAAQ,OAAA,KAAY,MAAA,KACxB,CAAC,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,OAAA,CAAQ,OAAO,CAAA,IAAK,IAAA,CAAK,OAAA,CAAQ,WAAW,CAAA,CAAA,EACnE;AACA,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,mDAAA,EAAsD,IAAA,CAAK,OAAA,CAAQ,OAAO,CAAA,CAAA;AAAA,QAC1E,SAAA;AAAA,QACA,KAAK,OAAA,CAAQ;AAAA,OACf;AAAA,IACF;AACA,IAAA,IACE,IAAA,CAAK,OAAA,CAAQ,GAAA,KAAQ,MAAA,KACpB,CAAC,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,OAAA,CAAQ,GAAG,CAAA,IAAK,IAAA,CAAK,OAAA,CAAQ,MAAM,CAAA,CAAA,EAC1D;AACA,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,2CAAA,EAA8C,IAAA,CAAK,OAAA,CAAQ,GAAG,CAAA,CAAA;AAAA,QAC9D,KAAA;AAAA,QACA,KAAK,OAAA,CAAQ;AAAA,OACf;AAAA,IACF;AACA,IAAA,IACE,IAAA,CAAK,OAAA,CAAQ,YAAA,KAAiB,MAAA,KAC7B,CAAC,MAAA,CAAO,QAAA,CAAS,IAAA,CAAK,OAAA,CAAQ,YAAY,CAAA,IAAK,IAAA,CAAK,OAAA,CAAQ,gBAAgB,CAAA,CAAA,EAC7E;AACA,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,wDAAA,EAA2D,IAAA,CAAK,OAAA,CAAQ,YAAY,CAAA,CAAA;AAAA,QACpF,cAAA;AAAA,QACA,KAAK,OAAA,CAAQ;AAAA,OACf;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaQ,QAAQ,CAAA,EAAmB;AAEjC,IAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,CAAC,CAAA,EAAG;AACvB,MAAA,OAAO,CAAA,GAAI,IAAI,CAAA,GAAI,CAAA;AAAA,IACrB;AACA,IAAA,IAAI,KAAK,CAAA,EAAG;AACV,MAAA,MAAMC,GAAAA,GAAK,IAAA,CAAK,GAAA,CAAI,CAAC,CAAC,CAAA;AACtB,MAAA,OAAO,KAAK,CAAA,GAAIA,GAAAA,CAAAA;AAAA,IAClB;AACA,IAAA,MAAM,EAAA,GAAK,IAAA,CAAK,GAAA,CAAI,CAAC,CAAA;AACrB,IAAA,OAAO,MAAM,CAAA,GAAI,EAAA,CAAA;AAAA,EACnB;AAAA,EAEQ,YAAA,GAAqB;AAC3B,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,KAAA,EAAO;AAC/B,MAAA,MAAM,IAAI,eAAe,4DAA4D,CAAA;AAAA,IACvF;AAAA,EACF;AAAA,EAEQ,UAAA,CACN,CAAA,EACA,CAAA,EACA,CAAA,EACA,GACA,MAAA,EAC4B;AAC5B,IAAA,MAAM,OAAA,GAAU,IAAA,CAAK,OAAA,CAAQ,OAAA,IAAW,GAAA;AACxC,IAAA,MAAM,GAAA,GAAM,IAAA,CAAK,OAAA,CAAQ,GAAA,IAAO,IAAA;AAChC,IAAA,MAAM,EAAA,GAAK,IAAA,CAAK,OAAA,CAAQ,YAAA,IAAgB,GAAA;AACxC,IAAA,MAAM,YAAA,GAAe,IAAA,CAAK,OAAA,CAAQ,YAAA,IAAgB,IAAA;AAGlD,IAAA,MAAM,IAAI,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACrC,IAAA,IAAI,CAAA,GAAI,CAAA;AAGR,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,OAAA,EAAS,IAAA,EAAA,EAAQ;AACzC,MAAA,MAAM,QAAQ,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACzC,MAAA,IAAI,KAAA,GAAQ,CAAA;AAGZ,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,CAAA,GAAI,eAAe,CAAA,GAAI,CAAA;AAC3B,QAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA;AAC/B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,KAAK,CAAC,CAAA,IAAK,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,QACnD;AAEA,QAAA,MAAM,EAAA,GAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAC3C,QAAA,MAAM,EAAA,GAAK,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AACzB,QAAA,MAAM,QAAQ,EAAA,GAAK,EAAA;AAEnB,QAAA,KAAA,IAAS,KAAA;AACT,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,KAAA,CAAM,CAAC,CAAA,GAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA,IAAK,KAAA,GAAQ,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,KAAK,CAAC,CAAA;AAAA,QACtE;AAAA,MACF;AAGA,MAAA,MAAM,IAAA,GAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA;AAC/B,MAAA,IAAI,SAAA,GAAY,CAAA;AAChB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,CAAA,GAAA,CAAK,MAAM,CAAC,CAAA,IAAK,KAAK,IAAA,GAAO,MAAA,IAAU,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AACrD,QAAA,MAAM,SAAS,EAAA,GAAK,CAAA;AACpB,QAAA,CAAA,CAAE,CAAC,CAAA,GAAA,CAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAK,MAAA;AACrB,QAAA,SAAA,GAAY,KAAK,GAAA,CAAI,SAAA,EAAW,IAAA,CAAK,GAAA,CAAI,MAAM,CAAC,CAAA;AAAA,MAClD;AACA,MAAA,IAAI,YAAA,EAAc;AAChB,QAAA,MAAM,KAAK,KAAA,GAAQ,IAAA;AACnB,QAAA,MAAM,UAAU,EAAA,GAAK,EAAA;AACrB,QAAA,CAAA,IAAK,OAAA;AACL,QAAA,SAAA,GAAY,KAAK,GAAA,CAAI,SAAA,EAAW,IAAA,CAAK,GAAA,CAAI,OAAO,CAAC,CAAA;AAAA,MACnD;AAGA,MAAA,IAAI,YAAY,GAAA,EAAK;AACnB,QAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,OAAO,EAAE,CAAA,EAAG,CAAA,EAAG,YAAA,GAAe,IAAI,CAAA,EAAE;AAAA,EACtC;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAiBA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAE9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAEtB,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,IAAA,CAAK,YAAA,GAAe,CAAA;AAGpB,IAAA,MAAM,KAAA,GAAQ,IAAI,YAAA,CAAa,CAAC,CAAA;AAChC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,KAAA,CAAM,CAAC,IAAI,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,IACxC;AACA,IAAA,MAAM,aAAA,GAAgB,CAAC,GAAG,IAAI,GAAA,CAAI,KAAK,CAAC,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,KAAM,IAAI,CAAC,CAAA;AAC9D,IAAA,IAAA,CAAK,QAAA,GAAW,OAAO,aAAa,CAAA;AAGpC,IAAA,MAAM,OAAA,GAAU,IAAA,CAAK,OAAA,CAAQ,OAAA,IAAW,IAAA;AACxC,IAAA,MAAM,CAAA,GAAI,IAAA,CAAK,OAAA,CAAQ,CAAA,IAAK,CAAA;AAC5B,IAAA,IAAI,EAAE,IAAI,CAAA,CAAA,EAAI;AACZ,MAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,wBAAA,EAA2B,CAAC,CAAA,CAAA,EAAI,KAAK,CAAC,CAAA;AAAA,IACxE;AACA,IAAA,MAAM,MAAA,GAAS,OAAA,KAAY,IAAA,GAAO,CAAA,GAAI,CAAA,GAAI,CAAA;AAC1C,IAAA,MAAM,UAAA,GAAa,IAAA,CAAK,OAAA,CAAQ,UAAA,IAAc,MAAA;AAE9C,IAAA,IAAI,aAAA,CAAc,UAAU,CAAA,EAAG;AAE7B,MAAA,IAAA,CAAK,WAAA,GAAc,KAAA;AAGnB,MAAA,IAAI,OAAA,GAAU,CAAA;AACd,MAAA,IAAI,aAAA,CAAc,MAAA,KAAW,CAAA,KAAM,aAAA,CAAc,CAAC,MAAM,CAAA,IAAK,aAAA,CAAc,CAAC,CAAA,KAAM,CAAA,CAAA,EAAI;AAEpF,QAAA,MAAM,UAAA,GAAa,IAAI,YAAA,CAAa,CAAC,CAAA;AACrC,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,UAAA,CAAW,CAAC,IAAI,KAAA,CAAM,CAAC,MAAM,aAAA,CAAc,CAAC,IAAI,CAAA,GAAI,CAAA;AAAA,QACtD;AACA,QAAA,OAAA,GAAU,OAAO,UAAU,CAAA;AAAA,MAC7B,CAAA,MAAA,IAAW,aAAA,CAAc,MAAA,KAAW,CAAA,EAAG;AAErC,QAAA,MAAM,UAAA,GAAa,IAAI,YAAA,CAAa,CAAC,CAAA;AACrC,QAAA,MAAM,MAAA,GAAS,aAAA,CAAc,CAAC,CAAA,KAAM,IAAI,CAAA,GAAI,CAAA;AAC5C,QAAA,UAAA,CAAW,KAAK,MAAM,CAAA;AACtB,QAAA,OAAA,GAAU,OAAO,UAAU,CAAA;AAAA,MAC7B,CAAA,MAAO;AAEL,QAAA,KAAA,MAAW,OAAO,aAAA,EAAe;AAC/B,UAAA,IAAI,GAAA,KAAQ,CAAA,IAAK,GAAA,KAAQ,CAAA,EAAG;AAE1B,YAAA,MAAM,IAAI,oBAAoB,8CAA8C,CAAA;AAAA,UAC9E;AAAA,QACF;AAAA,MACF;AAEA,MAAA,MAAM,EAAE,CAAA,EAAG,CAAA,EAAE,GAAI,IAAA,CAAK,WAAW,CAAA,EAAG,OAAA,EAAS,CAAA,EAAG,CAAA,EAAG,MAAM,CAAA;AACzD,MAAA,IAAA,CAAK,KAAA,GAAQ,OAAO,CAAC,CAAA;AACrB,MAAA,IAAA,CAAK,UAAA,GAAa,CAAA;AAAA,IACpB,CAAA,MAAO;AAEL,MAAA,IAAI,UAAA,KAAe,KAAA,IAAS,UAAA,KAAe,MAAA,EAAQ;AACjD,QAAA,MAAM,IAAI,qBAAA;AAAA,UACR,CAAA,6CAAA,EAAgD,MAAA,CAAO,UAAU,CAAC,CAAA,CAAA;AAAA,UAClE,YAAA;AAAA,UACA;AAAA,SACF;AAAA,MACF;AACA,MAAA,IAAA,CAAK,WAAA,GAAc,IAAA;AACnB,MAAA,MAAM,WAAW,aAAA,CAAc,MAAA;AAC/B,MAAA,MAAM,WAAqB,EAAC;AAC5B,MAAA,MAAM,gBAA0B,EAAC;AAEjC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,WAAA,GAAc,cAAc,CAAC,CAAA;AAEnC,QAAA,MAAM,WAAA,GAAc,IAAI,YAAA,CAAa,CAAC,CAAA;AACtC,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,WAAA,CAAY,CAAC,CAAA,GAAI,KAAA,CAAM,CAAC,CAAA,KAAM,cAAc,CAAA,GAAI,CAAA;AAAA,QAClD;AACA,QAAA,MAAM,OAAA,GAAU,OAAO,WAAW,CAAA;AAElC,QAAA,MAAM,EAAE,CAAA,EAAG,CAAA,EAAE,GAAI,IAAA,CAAK,WAAW,CAAA,EAAG,OAAA,EAAS,CAAA,EAAG,CAAA,EAAG,MAAM,CAAA;AACzD,QAAA,QAAA,CAAS,IAAA,CAAK,GAAG,CAAC,CAAA;AAClB,QAAA,aAAA,CAAc,KAAK,CAAC,CAAA;AAAA,MACtB;AAEA,MAAA,MAAM,cAAA,GAAiB,OAAO,QAAQ,CAAA;AACtC,MAAA,IAAA,CAAK,QAAQ,OAAA,CAAQ,cAAA,EAAgB,CAAC,QAAA,EAAU,CAAC,CAAC,CAAA;AAClD,MAAA,IAAA,CAAK,UAAA,GAAa,aAAA;AAAA,IACpB;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA,EAEA,IAAI,OAAA,GAA8B;AAChC,IAAA,OAAO,IAAA,CAAK,QAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,IAAA,GAAe;AACjB,IAAA,IAAA,CAAK,YAAA,EAAa;AAClB,IAAA,MAAM,OAAO,IAAA,CAAK,KAAA;AAClB,IAAA,IAAI,CAAC,IAAA,EAAM;AACT,MAAA,MAAM,IAAI,aAAa,wDAAwD,CAAA;AAAA,IACjF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,SAAA,GAA+B;AACjC,IAAA,IAAA,CAAK,YAAA,EAAa;AAClB,IAAA,IAAI,IAAA,CAAK,eAAe,MAAA,EAAW;AACjC,MAAA,OAAO,CAAA;AAAA,IACT;AACA,IAAA,OAAO,IAAA,CAAK,UAAA;AAAA,EACd;AAAA,EAEA,QAAQ,CAAA,EAAmB;AACzB,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,YAAA,CAAa,CAAC,CAAA;AACjC,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,OAAO,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAExC,IAAA,IAAI,KAAK,WAAA,EAAa;AACpB,MAAA,MAAM,UAAU,IAAA,CAAK,QAAA;AACrB,MAAA,IAAI,CAAC,OAAA,EAAS;AACZ,QAAA,MAAM,IAAI,eAAe,qCAAqC,CAAA;AAAA,MAChE;AACA,MAAA,MAAM,WAAW,OAAA,CAAQ,IAAA;AACzB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,OAAA,GAAU,EAAA;AACd,QAAA,IAAI,WAAA,GAAc,CAAA;AAClB,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,MAAM,CAAA,GAAI,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAA,GAAI,QAAA,GAAW,CAAC,CAAC,CAAA;AAC5D,UAAA,IAAI,IAAI,OAAA,EAAS;AACf,YAAA,OAAA,GAAU,CAAA;AACV,YAAA,WAAA,GAAc,CAAA;AAAA,UAChB;AAAA,QACF;AAEA,QAAA,IAAA,CAAK,CAAC,IAAI,MAAA,CAAO,OAAA,CAAQ,KAAK,OAAA,CAAQ,MAAA,GAAS,WAAW,CAAC,CAAA;AAAA,MAC7D;AAAA,IACF,CAAA,MAAO;AACL,MAAA,MAAM,UAAU,IAAA,CAAK,QAAA;AACrB,MAAA,IAAI,CAAC,OAAA,EAAS;AACZ,QAAA,MAAM,IAAI,eAAe,qCAAqC,CAAA;AAAA,MAChE;AAKA,MAAA,IAAI,OAAA,CAAQ,SAAS,CAAA,EAAG;AACtB,QAAA,MAAM,MAAM,MAAA,CAAO,OAAA,CAAQ,IAAA,CAAK,OAAA,CAAQ,MAAM,CAAC,CAAA;AAC/C,QAAA,IAAA,CAAK,KAAK,GAAG,CAAA;AAAA,MACf,CAAA,MAAO;AACL,QAAA,MAAM,OAAO,MAAA,CAAO,OAAA,CAAQ,IAAA,CAAK,OAAA,CAAQ,MAAM,CAAC,CAAA;AAChD,QAAA,MAAM,OAAO,MAAA,CAAO,OAAA,CAAQ,KAAK,OAAA,CAAQ,MAAA,GAAS,CAAC,CAAC,CAAA;AAEpD,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,MAAM,EAAA,GAAK,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAA,GAAI,CAAA,GAAI,CAAC,CAAC,CAAA;AACtD,UAAA,IAAA,CAAK,CAAC,CAAA,GAAI,EAAA,IAAM,GAAA,GAAM,IAAA,GAAO,IAAA;AAAA,QAC/B;AAAA,MACF;AAAA,IACF;AACA,IAAA,OAAO,OAAO,IAAI,CAAA;AAAA,EACpB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,aAAa,CAAA,EAAmB;AAC9B,IAAA,IAAA,CAAK,YAAA,EAAa;AAElB,IAAA,MAAM,OAAO,IAAA,CAAK,KAAA;AAClB,IAAA,IAAI,CAAC,IAAA,EAAM;AACT,MAAA,MAAM,IAAI,aAAa,uDAAuD,CAAA;AAAA,IAChF;AAGA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,oBAAoB,CAAA;AAErE,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAExB,IAAA,IAAI,KAAK,WAAA,EAAa;AACpB,MAAA,MAAM,QAAA,GAAW,IAAA,CAAK,QAAA,EAAU,IAAA,IAAQ,CAAA;AACxC,MAAA,MAAM,KAAA,GAAQ,IAAI,YAAA,CAAa,CAAA,GAAI,QAAQ,CAAA;AAC3C,MAAA,MAAM,iBAAiB,IAAA,CAAK,UAAA;AAC5B,MAAA,IAAI,CAAC,KAAA,CAAM,OAAA,CAAQ,cAAc,CAAA,EAAG;AAClC,QAAA,MAAM,IAAI,aAAa,4DAA4D,CAAA;AAAA,MACrF;AAEA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA;AAC/B,QAAA,IAAI,MAAA,GAAS,CAAA;AACb,QAAA,MAAM,SAAS,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,KAAK,CAAC,CAAA;AAGjD,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,IAAI,CAAA,GAAI,cAAA,CAAe,CAAC,CAAA,IAAK,CAAA;AAC7B,UAAA,MAAM,WAAA,GAAc,IAAA,CAAK,MAAA,GAAS,CAAA,GAAI,CAAA;AACtC,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,YAAA,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,CAAA,IAAK,CAAC,CAAA,GAAI,MAAA,CAAO,IAAA,CAAK,IAAA,CAAK,WAAA,GAAc,CAAC,KAAK,CAAC,CAAA;AAAA,UAChF;AAGA,UAAA,MAAA,CAAO,CAAC,CAAA,GAAI,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC1B,UAAA,MAAA,IAAU,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA;AAAA,QACzB;AAGA,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AAEjC,UAAA,KAAA,CAAM,CAAA,GAAI,QAAA,GAAW,CAAC,CAAA,GAAI,MAAA,GAAS,CAAA,GAAA,CAAK,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA,IAAK,MAAA,GAAS,CAAA,GAAM,QAAA;AAAA,QAC3E;AAAA,MACF;AAEA,MAAA,OAAO,MAAA,CAAO,KAAA,EAAO,EAAE,KAAA,EAAO,SAAA,EAAW,CAAA,CAAE,OAAA,CAAQ,CAAC,CAAA,EAAG,QAAQ,CAAC,CAAA;AAAA,IAClE,CAAA,MAAO;AAEL,MAAA,MAAM,QAAQ,IAAI,KAAA,CAAc,IAAI,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAC7C,MAAA,MAAM,iBAAiB,IAAA,CAAK,UAAA;AAC5B,MAAA,IAAI,MAAM,OAAA,CAAQ,cAAc,CAAA,IAAK,OAAO,mBAAmB,QAAA,EAAU;AACvE,QAAA,MAAM,IAAI,aAAa,6DAA6D,CAAA;AAAA,MACtF;AAEA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,CAAA,GAAI,cAAA;AACR,QAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA;AAC/B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,CAAA,IAAK,CAAC,CAAA,GAAI,MAAA,CAAO,KAAK,IAAA,CAAK,IAAA,CAAK,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAAA,QAChF;AAEA,QAAA,MAAM,EAAA,GAAK,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AACzB,QAAA,KAAA,CAAM,CAAA,GAAI,CAAA,GAAI,CAAC,CAAA,GAAI,CAAA,GAAI,EAAA;AACvB,QAAA,KAAA,CAAM,CAAA,GAAI,CAAA,GAAI,CAAC,CAAA,GAAI,EAAA;AAAA,MACrB;AAEA,MAAA,OAAO,OAAO,KAAK,CAAA,CAAE,QAAQ,CAAC,CAAA,EAAG,CAAC,CAAC,CAAA;AAAA,IACrC;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,IAAA,GAAO,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC3B,IAAA,IAAI,IAAA,CAAK,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACxB,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,IAAA,CAAK,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OAC/E;AAAA,IACF;AACA,IAAA,IAAI,OAAA,GAAU,CAAA;AAEd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,IAAI,OAAO,IAAA,CAAK,IAAA,CAAK,IAAA,CAAK,MAAA,GAAS,CAAC,CAAA,IAAK,CAAC,CAAA,KAAM,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAC,CAAA,IAAK,CAAC,CAAA,EAAG;AACjF,QAAA,OAAA,EAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,OAAO,UAAU,CAAA,CAAE,IAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO,EAAE,GAAG,IAAA,CAAK,OAAA,EAAQ;AAAA,EAC3B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,UAAU,MAAA,EAAuC;AAC/C,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,KAAK,KAAK,MAAA,CAAO,OAAA,CAAQ,MAAM,CAAA,EAAG;AACjD,MAAA,QAAQ,GAAA;AAAK,QACX,KAAK,SAAA;AACH,UAAA,IAAI,OAAO,UAAU,QAAA,IAAY,CAAC,OAAO,QAAA,CAAS,KAAK,CAAA,IAAK,KAAA,IAAS,CAAA,EAAG;AACtE,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,0CAAA;AAAA,cACA,SAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,OAAA,GAAU,KAAA;AACvB,UAAA;AAAA,QACF,KAAK,KAAA;AACH,UAAA,IAAI,OAAO,UAAU,QAAA,IAAY,CAAC,OAAO,QAAA,CAAS,KAAK,CAAA,IAAK,KAAA,GAAQ,CAAA,EAAG;AACrE,YAAA,MAAM,IAAI,qBAAA,CAAsB,kCAAA,EAAoC,KAAA,EAAO,KAAK,CAAA;AAAA,UAClF;AACA,UAAA,IAAA,CAAK,QAAQ,GAAA,GAAM,KAAA;AACnB,UAAA;AAAA,QACF,KAAK,GAAA;AACH,UAAA,IAAI,OAAO,UAAU,QAAA,IAAY,CAAC,OAAO,QAAA,CAAS,KAAK,CAAA,IAAK,KAAA,IAAS,CAAA,EAAG;AACtE,YAAA,MAAM,IAAI,qBAAA,CAAsB,oCAAA,EAAsC,GAAA,EAAK,KAAK,CAAA;AAAA,UAClF;AACA,UAAA,IAAA,CAAK,QAAQ,CAAA,GAAI,KAAA;AACjB,UAAA;AAAA,QACF,KAAK,cAAA;AACH,UAAA,IAAI,OAAO,UAAU,QAAA,IAAY,CAAC,OAAO,QAAA,CAAS,KAAK,CAAA,IAAK,KAAA,IAAS,CAAA,EAAG;AACtE,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,+CAAA;AAAA,cACA,cAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,YAAA,GAAe,KAAA;AAC5B,UAAA;AAAA,QACF,KAAK,cAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,gCAAA;AAAA,cACA,cAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,YAAA,GAAe,KAAA;AAC5B,UAAA;AAAA,QACF,KAAK,SAAA;AACH,UAAA,IAAI,KAAA,KAAU,MAAA,IAAU,KAAA,KAAU,IAAA,EAAM;AACtC,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,mDAAA,EAAsD,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACnE,SAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,OAAA,GAAU,KAAA;AACvB,UAAA;AAAA,QACF,KAAK,YAAA;AACH,UAAA,IAAI,KAAA,KAAU,KAAA,IAAS,KAAA,KAAU,MAAA,EAAQ;AACvC,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,6CAAA,EAAgD,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cAC7D,YAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,UAAA,GAAa,KAAA;AAC1B,UAAA;AAAA,QACF;AACE,UAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,mBAAA,EAAsB,GAAG,CAAA,CAAA,EAAI,KAAK,KAAK,CAAA;AAAA;AAC3E,IACF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AACF;;;AC5lBO,IAAM,QAAN,MAAiC;AAAA;AAAA,EAE9B,OAAA;AAAA;AAAA,EAUA,KAAA;AAAA;AAAA,EAGA,UAAA;AAAA;AAAA,EAGA,YAAA;AAAA;AAAA,EAGA,MAAA;AAAA;AAAA,EAGA,MAAA,GAAS,KAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAajB,WAAA,CACE,OAAA,GAOI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,OAAA,GAAU,EAAE,GAAG,OAAA,EAAQ;AAAA,EAC9B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAuBA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAE9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AACtB,IAAA,IAAA,CAAK,MAAA,GAAS,MAAA;AAGd,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,KAAA,IAAS,CAAA;AACpC,IAAA,IAAI,EAAE,SAAS,CAAA,CAAA,EAAI;AACjB,MAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,6BAAA,EAAgC,KAAK,CAAA,CAAA,EAAI,SAAS,KAAK,CAAA;AAAA,IACzF;AAGA,IAAA,MAAM,YAAA,GAAe,IAAA,CAAK,OAAA,CAAQ,YAAA,IAAgB,IAAA;AAGlD,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAGxB,IAAA,IAAA,CAAK,YAAA,GAAe,CAAA;AAOpB,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,MAAM,QAAQ,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAEzC,IAAA,IAAI,YAAA,EAAc;AAEhB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAAA,MAC3C;AAGA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA;AAC/B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,KAAA,CAAM,CAAC,CAAA,GAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,CAAA,IAAK,CAAC,CAAA;AAAA,QAC9D;AAAA,MACF;AAGA,MAAA,MAAM,IAAA,GAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA;AAC/B,MAAA,KAAA,IAAS,IAAA;AACT,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,KAAA,CAAM,CAAC,CAAA,GAAA,CAAK,KAAA,CAAM,CAAC,KAAK,CAAA,IAAK,IAAA;AAAA,MAC/B;AAAA,IACF;AAEA,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,OAAA,CAAQ,SAAA,IAAa,KAAA;AAC5C,IAAA,MAAM,OAAA,GAAU,IAAA,CAAK,OAAA,CAAQ,OAAA,IAAW,GAAA;AACxC,IAAA,MAAM,GAAA,GAAM,IAAA,CAAK,OAAA,CAAQ,GAAA,IAAO,IAAA;AAEhC,IAAA,IAAI,MAAA;AACJ,IAAA,IAAI,SAAA,EAAW;AACb,MAAA,MAAA,GAAS,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACpC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,OAAA,GAAU,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,CAAA;AAC/B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,MAAM,QAAA,GAAW,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,OAAA,GAAU,CAAC,CAAA,IAAK,CAAC,CAAA,IAAK,YAAA,GAAgB,KAAA,CAAM,CAAC,KAAK,CAAA,GAAK,CAAA,CAAA;AACtF,UAAA,MAAA,CAAO,CAAC,CAAA,GAAA,CAAK,MAAA,CAAO,CAAC,CAAA,IAAK,KAAK,QAAA,GAAW,QAAA;AAAA,QAC5C;AAAA,MACF;AACA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAA,CAAO,CAAC,CAAA,GAAI,IAAA,CAAK,KAAK,MAAA,CAAO,CAAC,KAAK,CAAC,CAAA;AAAA,MACtC;AAAA,IACF;AAEA,IAAA,MAAM,IAAA,GAAO,CAAC,WAAA,EAAqB,YAAA,KAAiC;AAClE,MAAA,MAAM,GAAA,GAAM,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,WAAA,GAAc,CAAA,GAAI,YAAY,CAAA,IAAK,CAAC,CAAA;AACzE,MAAA,MAAM,WAAW,GAAA,IAAO,YAAA,GAAgB,KAAA,CAAM,YAAY,KAAK,CAAA,GAAK,CAAA,CAAA;AACpE,MAAA,IAAI,aAAa,MAAA,EAAQ;AACvB,QAAA,MAAM,CAAA,GAAI,MAAA,CAAO,YAAY,CAAA,IAAK,CAAA;AAClC,QAAA,OAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,QAAA,GAAW,CAAA;AAAA,MAClC;AACA,MAAA,OAAO,QAAA;AAAA,IACT,CAAA;AAEA,IAAA,MAAM,IAAA,GAAO,CAAC,WAAA,KAAgC;AAC5C,MAAA,MAAM,GAAA,GAAM,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,WAAW,KAAK,CAAC,CAAA;AACtD,MAAA,OAAO,YAAA,GAAe,MAAM,KAAA,GAAQ,GAAA;AAAA,IACtC,CAAA;AAIA,IAAA,IAAI,UAAA;AACJ,IAAA,MAAM,MAAA,GAAS,IAAA,CAAK,OAAA,CAAQ,MAAA,IAAU,MAAA;AAEtC,IAAA,IAAI,WAAW,KAAA,EAAO;AACpB,MAAA,MAAM,GAAA,GAAM,KAAK,QAAA,CAAS,IAAA,EAAM,MAAM,CAAA,EAAG,CAAA,EAAG,KAAA,EAAO,OAAA,EAAS,GAAG,CAAA;AAC/D,MAAA,UAAA,GAAa,MAAA,CAAO,IAAI,CAAC,CAAA;AACzB,MAAA,IAAA,CAAK,SAAS,GAAA,CAAI,KAAA;AAAA,IACpB,CAAA,MAAO;AAIL,MAAA,MAAM,GAAA,GAAM,KAAA,CAAM,CAAC,CAAA,CAChB,KAAK,CAAC,CAAA,CACN,GAAA,CAAI,MAAM,KAAA,CAAM,CAAC,CAAA,CAAE,IAAA,CAAK,CAAC,CAAC,CAAA;AAE7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,IAAI,GAAA,GAAM,CAAA;AAGV,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,YAAA,MAAM,EAAA,GAAK,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACpB,YAAA,MAAM,EAAA,GAAK,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACpB,YAAA,GAAA,IAAO,EAAA,GAAK,EAAA;AAAA,UACd;AAIA,UAAA,MAAM,MAAA,GAAS,IAAI,CAAC,CAAA;AACpB,UAAA,IAAI,QAAQ,MAAA,CAAO,CAAC,IAAI,GAAA,IAAO,CAAA,KAAM,IAAI,KAAA,GAAQ,CAAA,CAAA;AAAA,QACnD;AAAA,MACF;AAIA,MAAA,MAAM,MAAM,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAEvC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,GAAA,GAAM,CAAA;AAGV,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,MAAM,IAAA,GAAO,KAAK,CAAC,CAAA;AACnB,UAAA,MAAM,IAAA,GAAO,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACtB,UAAA,GAAA,IAAO,IAAA,GAAO,IAAA;AAAA,QAChB;AACA,QAAA,GAAA,CAAI,CAAC,CAAA,GAAI,GAAA;AAAA,MACX;AAEA,MAAA,IAAI,WAAW,MAAA,EAAQ;AACrB,QAAA,MAAM,MAAM,IAAA,CAAK,sBAAA,CAAuB,GAAA,EAAK,GAAA,EAAK,SAAS,GAAG,CAAA;AAC9D,QAAA,UAAA,GAAa,MAAA,CAAO,IAAI,CAAC,CAAA;AACzB,QAAA,IAAA,CAAK,SAAS,GAAA,CAAI,KAAA;AAAA,MACpB,CAAA,MAAA,IAAW,MAAA,KAAW,UAAA,IAAc,MAAA,KAAW,MAAA,EAAQ;AACrD,QAAA,IAAI;AACF,UAAA,MAAM,SAAA,GAAY,OAAO,GAAG,CAAA;AAC5B,UAAA,MAAM,SAAA,GAAY,OAAO,GAAG,CAAA;AAC5B,UAAA,MAAM,CAAA,GAAI,SAAS,SAAS,CAAA;AAC5B,UAAA,MAAM,EAAA,GAAK,eAAA,CAAgB,CAAA,EAAG,SAAA,EAAW,IAAI,CAAA;AAC7C,UAAA,UAAA,GAAa,eAAA,CAAgB,SAAA,CAAU,CAAC,CAAA,EAAG,IAAI,KAAK,CAAA;AAAA,QACtD,SAAS,CAAA,EAAG;AACV,UAAA,IAAI,WAAW,MAAA,EAAQ;AAErB,YAAA,MAAM,GAAA,GAAM,IAAA,CAAK,iBAAA,CAAkB,GAAA,EAAK,GAAG,CAAA;AAC3C,YAAA,UAAA,GAAa,OAAO,GAAG,CAAA;AAAA,UACzB,CAAA,MAAO;AACL,YAAA,MAAM,CAAA;AAAA,UACR;AAAA,QACF;AAAA,MACF,CAAA,MAAA,IAAW,WAAW,KAAA,EAAO;AAC3B,QAAA,MAAM,SAAA,GAAY,OAAO,GAAG,CAAA;AAC5B,QAAA,MAAM,SAAA,GAAY,OAAO,GAAG,CAAA;AAC5B,QAAA,MAAM,CAAC,CAAA,EAAG,CAAA,EAAG,EAAE,CAAA,GAAI,IAAI,SAAS,CAAA;AAGhC,QAAA,MAAM,EAAA,GAAK,UAAU,CAAC,CAAA;AACtB,QAAA,MAAM,GAAA,GAAM,GAAA,CAAI,EAAA,EAAI,SAAS,CAAA;AAE7B,QAAA,MAAM,QAAQ,CAAA,CAAE,IAAA;AAChB,QAAA,IAAI,EAAE,iBAAiB,YAAA,CAAA,EAAe;AACpC,UAAA,MAAM,IAAI,oBAAoB,0CAA0C,CAAA;AAAA,QAC1E;AACA,QAAA,MAAM,UAAA,GAAa,IAAI,YAAA,CAAa,GAAA,CAAI,IAAI,CAAA;AAC5C,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,GAAA,CAAI,MAAM,CAAA,EAAA,EAAK;AACjC,UAAA,MAAM,MAAM,MAAA,CAAO,GAAA,CAAI,KAAK,GAAA,CAAI,MAAA,GAAS,CAAC,CAAC,CAAA;AAC3C,UAAA,MAAM,KAAA,GAAQ,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC1B,UAAA,UAAA,CAAW,CAAC,IAAI,IAAA,CAAK,GAAA,CAAI,KAAK,CAAA,GAAI,KAAA,GAAQ,MAAM,KAAA,GAAQ,CAAA;AAAA,QAC1D;AACA,QAAA,MAAM,MAAA,GAAS,OAAO,UAAU,CAAA;AAEhC,QAAA,MAAM,CAAA,GAAI,UAAU,EAAE,CAAA;AACtB,QAAA,UAAA,GAAa,GAAA,CAAI,GAAG,MAAM,CAAA;AAAA,MAC5B,CAAA,MAAO;AACL,QAAA,MAAM,GAAA,GAAM,IAAA,CAAK,iBAAA,CAAkB,GAAA,EAAK,GAAG,CAAA;AAC3C,QAAA,UAAA,GAAa,OAAO,GAAG,CAAA;AAAA,MACzB;AAAA,IACF;AAEA,IAAA,IAAI,aAAa,MAAA,EAAQ;AACvB,MAAA,UAAA,GAAa,IAAA,CAAK,YAAA,CAAa,UAAA,EAAY,MAAM,CAAA;AAAA,IACnD;AAEA,IAAA,IAAA,CAAK,KAAA,GAAQ,UAAA;AAKb,IAAA,IAAI,YAAA,EAAc;AAChB,MAAA,IAAI,SAAA,GAAY,CAAA;AAChB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,EAAA,GAAK,OAAO,UAAA,CAAW,IAAA,CAAK,WAAW,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAC7D,QAAA,SAAA,IAAA,CAAc,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA,IAAK,EAAA;AAAA,MACjC;AACA,MAAA,IAAA,CAAK,aAAa,KAAA,GAAQ,SAAA;AAAA,IAC5B,CAAA,MAAO;AACL,MAAA,IAAA,CAAK,UAAA,GAAa,CAAA;AAAA,IACpB;AAGA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAsBQ,iBAAA,CAAkB,GAAe,CAAA,EAAuB;AAC9D,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AAIZ,IAAA,MAAM,GAAA,GAAM,CAAA,CAAE,GAAA,CAAI,CAAC,GAAA,EAAK,CAAA,KAAM,CAAC,GAAG,GAAA,EAAK,CAAA,CAAE,CAAC,CAAA,IAAK,CAAC,CAAC,CAAA;AACjD,IAAA,IAAI,MAAA,GAAS,CAAA;AACb,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,GAAA,GAAM,IAAI,CAAC,CAAA;AACjB,MAAA,IAAI,CAAC,GAAA,EAAK;AACV,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,IAAI,IAAA,CAAK,GAAA,CAAI,GAAA,CAAI,CAAC,KAAK,CAAC,CAAA;AAC9B,QAAA,IAAI,CAAA,GAAI,QAAQ,MAAA,GAAS,CAAA;AAAA,MAC3B;AAAA,IACF;AACA,IAAA,IAAI,WAAW,CAAA,IAAK,CAAC,MAAA,CAAO,QAAA,CAAS,MAAM,CAAA,EAAG;AAC5C,MAAA,MAAM,IAAI,oBAAoB,uCAAuC,CAAA;AAAA,IACvE;AACA,IAAA,MAAM,GAAA,GAAM,MAAA,CAAO,OAAA,GAAU,CAAA,GAAI,MAAA;AAGjC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAG1B,MAAA,IAAI,MAAA,GAAS,CAAA;AACb,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,GAAG,CAAA,EAAA,EAAK;AAC9B,QAAA,IAAI,KAAK,GAAA,CAAI,GAAA,CAAI,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAC,CAAA,GAAI,IAAA,CAAK,IAAI,GAAA,CAAI,MAAM,IAAI,CAAC,CAAA,IAAK,CAAC,CAAA,EAAG;AAChE,UAAA,MAAA,GAAS,CAAA;AAAA,QACX;AAAA,MACF;AAGA,MAAA,MAAM,IAAA,GAAO,GAAA,CAAI,CAAC,CAAA,IAAK,EAAC;AACxB,MAAA,MAAM,MAAA,GAAS,GAAA,CAAI,MAAM,CAAA,IAAK,EAAC;AAC/B,MAAA,GAAA,CAAI,CAAC,CAAA,GAAI,MAAA;AACT,MAAA,GAAA,CAAI,MAAM,CAAA,GAAI,IAAA;AAEd,MAAA,MAAM,KAAA,GAAQ,GAAA,CAAI,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAC7B,MAAA,IAAI,CAAC,OAAO,QAAA,CAAS,KAAK,KAAK,IAAA,CAAK,GAAA,CAAI,KAAK,CAAA,IAAK,GAAA,EAAK;AACrD,QAAA,MAAM,IAAI,oBAAoB,uCAAuC,CAAA;AAAA,MACvE;AAGA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,GAAG,CAAA,EAAA,EAAK;AAE9B,QAAA,MAAM,KAAK,GAAA,CAAI,CAAC,CAAA,GAAI,CAAC,KAAK,CAAA,IAAK,KAAA;AAC/B,QAAA,MAAM,IAAA,GAAO,IAAI,CAAC,CAAA;AAElB,QAAA,IAAI,IAAA,EAAM;AAER,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,IAAK,CAAA,EAAG,CAAA,EAAA,EAAK;AAC3B,YAAA,IAAA,CAAK,CAAC,CAAA,GAAA,CAAK,IAAA,CAAK,CAAC,CAAA,IAAK,CAAA,IAAK,CAAA,IAAK,GAAA,CAAI,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,UACjD;AAAA,QACF;AAAA,MACF;AAAA,IACF;AAGA,IAAA,MAAM,CAAA,GAAI,KAAA,CAAM,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,GAAI,CAAA,EAAG,CAAA,IAAK,GAAG,CAAA,EAAA,EAAK;AAE/B,MAAA,CAAA,CAAE,CAAC,CAAA,GAAI,GAAA,CAAI,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAGtB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,GAAG,CAAA,EAAA,EAAK;AAC9B,QAAA,CAAA,CAAE,CAAC,CAAA,GAAA,CAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAA,CAAM,GAAA,CAAI,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA,KAAM,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,MACrD;AAGA,MAAA,MAAM,IAAA,GAAO,GAAA,CAAI,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AAC5B,MAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAI,KAAK,IAAA,CAAK,GAAA,CAAI,IAAI,CAAA,IAAK,GAAA,EAAK;AACnD,QAAA,MAAM,IAAI,oBAAoB,uCAAuC,CAAA;AAAA,MACvE;AACA,MAAA,CAAA,CAAE,CAAC,CAAA,GAAA,CAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAK,IAAA;AAAA,IACvB;AAEA,IAAA,OAAO,CAAA;AAAA,EACT;AAAA,EAEQ,sBAAA,CACN,CAAA,EACA,CAAA,EACA,OAAA,EACA,GAAA,EACgC;AAChC,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AACZ,IAAA,MAAM,IAAI,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACrC,IAAA,MAAM,IAAI,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAErC,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,EAAA,GAAK,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA;AACnB,MAAA,CAAA,CAAE,CAAC,CAAA,GAAI,EAAA;AACP,MAAA,KAAA,IAAS,EAAA,GAAK,EAAA;AAAA,IAChB;AAEA,IAAA,IAAI,UAAU,CAAA,EAAG;AACf,MAAA,OAAO,EAAE,CAAA,EAAG,KAAA,EAAO,CAAA,EAAE;AAAA,IACvB;AAEA,IAAA,MAAM,CAAA,GAAI,EAAE,KAAA,EAAM;AAClB,IAAA,MAAM,QAAQ,GAAA,GAAM,GAAA;AACpB,IAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,OAAA,EAAS,IAAA,EAAA,EAAQ;AACzC,MAAA,MAAM,KAAK,IAAI,KAAA,CAAc,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AACtC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,GAAA,GAAM,CAAA;AACV,QAAA,MAAM,GAAA,GAAM,EAAE,CAAC,CAAA;AACf,QAAA,IAAI,CAAC,GAAA,EAAK;AACV,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,GAAA,IAAA,CAAQ,IAAI,CAAC,CAAA,IAAK,CAAA,KAAM,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,QAClC;AACA,QAAA,EAAA,CAAG,CAAC,CAAA,GAAI,GAAA;AAAA,MACV;AAEA,MAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,KAAA,IAAA,CAAU,EAAE,CAAC,CAAA,IAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,MACnC;AACA,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,IAAK,UAAU,CAAA,EAAG;AAC1C,QAAA,MAAM,IAAI,mBAAA;AAAA,UACR;AAAA,SACF;AAAA,MACF;AAEA,MAAA,MAAM,QAAQ,KAAA,GAAQ,KAAA;AACtB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,CAAA,CAAE,CAAC,KAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAK,KAAA,IAAS,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AACtC,QAAA,CAAA,CAAE,CAAC,KAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAK,KAAA,IAAS,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,MACzC;AAEA,MAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,EAAA,GAAK,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA;AACnB,QAAA,KAAA,IAAS,EAAA,GAAK,EAAA;AAAA,MAChB;AACA,MAAA,KAAA,GAAQ,IAAA,GAAO,CAAA;AACf,MAAA,IAAI,QAAQ,KAAA,EAAO;AACjB,QAAA;AAAA,MACF;AAEA,MAAA,MAAM,OAAO,KAAA,GAAQ,KAAA;AACrB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,CAAA,CAAE,CAAC,KAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAK,IAAA,IAAQ,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,MACvC;AACA,MAAA,KAAA,GAAQ,KAAA;AAAA,IACV;AAEA,IAAA,OAAO,EAAE,GAAG,KAAA,EAAM;AAAA,EACpB;AAAA,EAEQ,SACN,IAAA,EACA,IAAA,EACA,UACA,SAAA,EACA,KAAA,EACA,SACA,GAAA,EACgC;AAChC,IAAA,MAAM,IAAI,IAAI,KAAA,CAAc,SAAS,CAAA,CAAE,KAAK,CAAC,CAAA;AAC7C,IAAA,MAAM,UAAU,IAAI,KAAA,CAAc,SAAS,CAAA,CAAE,KAAK,CAAC,CAAA;AACnD,IAAA,MAAM,YAAY,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,KAAK,CAAC,CAAA;AAEpD,IAAA,IAAI,SAAA,GAAY,CAAA;AAChB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,IAAI,MAAA,GAAS,CAAA;AACb,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,GAAA,GAAM,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACrB,QAAA,MAAA,IAAU,GAAA,GAAM,GAAA;AAAA,MAClB;AACA,MAAA,IAAI,SAAS,SAAA,EAAW;AACtB,QAAA,SAAA,GAAY,MAAA;AAAA,MACd;AAAA,IACF;AAEA,IAAA,MAAM,KAAA,GAAQ,QAAA,KAAa,CAAA,GAAI,CAAA,GAAI,QAAA;AACnC,IAAA,MAAM,CAAA,GAAI,YAAY,KAAA,GAAQ,KAAA;AAC9B,IAAA,MAAM,IAAA,GAAO,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA;AAE7B,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,OAAA,EAAS,IAAA,EAAA,EAAQ;AACzC,MAAA,IAAI,SAAA,GAAY,CAAA;AAEhB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,IAAI,OAAA,GAAU,CAAA;AACd,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,OAAA,IAAA,CAAY,EAAE,CAAC,CAAA,IAAK,CAAA,IAAK,IAAA,CAAK,GAAG,CAAC,CAAA;AAAA,QACpC;AAEA,QAAA,MAAM,EAAA,GAAK,KAAK,CAAC,CAAA;AACjB,QAAA,MAAM,cAAc,OAAA,GAAU,EAAA;AAC9B,QAAA,MAAM,KAAA,GAAQ,WAAA,IAAe,SAAA,CAAU,CAAC,CAAA,IAAK,CAAA,CAAA;AAC7C,QAAA,SAAA,CAAU,CAAC,CAAA,GAAI,WAAA;AAEf,QAAA,IAAI,UAAU,CAAA,EAAG;AACf,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,OAAA,CAAQ,CAAC,KAAK,OAAA,CAAQ,CAAC,KAAK,CAAA,IAAK,KAAA,GAAQ,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AAAA,UACpD;AAAA,QACF;AAEA,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,MAAM,IAAA,GAAA,CAAQ,QAAQ,CAAC,CAAA,IAAK,KAAK,KAAA,IAAS,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AAClD,UAAA,MAAM,SAAS,IAAA,GAAO,IAAA;AACtB,UAAA,CAAA,CAAE,CAAC,CAAA,GAAA,CAAK,CAAA,CAAE,CAAC,KAAK,CAAA,IAAK,MAAA;AACrB,UAAA,IAAI,IAAA,CAAK,GAAA,CAAI,MAAM,CAAA,GAAI,SAAA,EAAW;AAChC,YAAA,SAAA,GAAY,IAAA,CAAK,IAAI,MAAM,CAAA;AAAA,UAC7B;AAAA,QACF;AAAA,MACF;AAEA,MAAA,KAAA,GAAQ,IAAA,GAAO,CAAA;AACf,MAAA,IAAI,YAAY,GAAA,EAAK;AACnB,QAAA;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,EAAE,CAAA,EAAG,CAAA,EAAG,KAAA,EAAM;AAAA,EACvB;AAAA,EAEQ,YAAA,CAAa,MAAc,KAAA,EAAyB;AAC1D,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACnC,IAAA,MAAM,SAAmB,EAAC;AAC1B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,CAAA,GAAI,OAAO,IAAA,CAAK,IAAA,CAAK,KAAK,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAChD,MAAA,MAAM,CAAA,GAAI,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACtB,MAAA,MAAA,CAAO,IAAA,CAAK,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,IAAI,CAAC,CAAA;AAAA,IACjC;AACA,IAAA,OAAO,OAAO,MAAM,CAAA;AAAA,EACtB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAgBA,QAAQ,CAAA,EAAmB;AAEzB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,KAAA,EAAO;AAC/B,MAAA,MAAM,IAAI,eAAe,wCAAwC,CAAA;AAAA,IACnE;AAGA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,OAAO,CAAA;AAExD,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,CAAA,GAAI,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACxB,IAAA,MAAM,IAAA,GAAO,KAAA,CAAM,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAG5B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,IAAI,GAAA,GAAM,KAAK,UAAA,IAAc,CAAA;AAG7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,GAAA,IACE,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,CAAA,GAAI,CAAC,CAAA,IAAK,CAAC,IACxC,MAAA,CAAO,IAAA,CAAK,MAAM,IAAA,CAAK,IAAA,CAAK,MAAM,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAAA,MACtD;AACA,MAAA,IAAA,CAAK,CAAC,CAAA,GAAI,GAAA;AAAA,IACZ;AAEA,IAAA,OAAO,OAAO,IAAI,CAAA;AAAA,EACpB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAyBA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAElC,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,qCAAqC,CAAA;AAAA,IAChE;AAGA,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AAGA,IAAA,MAAM,IAAA,GAAO,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC3B,IAAA,IAAI,IAAA,CAAK,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACxB,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,IAAA,CAAK,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OAC/E;AAAA,IACF;AAEA,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AAGZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAAA,IAC3C;AACA,IAAA,KAAA,IAAS,CAAA,CAAE,IAAA;AAGX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,IAAA,GAAO,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAC7C,MAAA,MAAM,OAAA,GAAU,OAAO,IAAA,CAAK,IAAA,CAAK,KAAK,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAGtD,MAAA,KAAA,IAAA,CAAU,OAAO,OAAA,KAAY,CAAA;AAG7B,MAAA,KAAA,IAAA,CAAU,OAAO,KAAA,KAAU,CAAA;AAAA,IAC7B;AAKA,IAAA,IAAI,UAAU,CAAA,EAAG;AACf,MAAA,OAAO,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA;AAAA,IAC7B;AAGA,IAAA,OAAO,IAAI,KAAA,GAAQ,KAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,SAAA,GAAqC;AACnC,IAAA,OAAO,EAAE,GAAG,IAAA,CAAK,OAAA,EAAQ;AAAA,EAC3B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,UAAU,MAAA,EAAuC;AAC/C,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,KAAK,KAAK,MAAA,CAAO,OAAA,CAAQ,MAAM,CAAA,EAAG;AACjD,MAAA,QAAQ,GAAA;AAAK,QACX,KAAK,OAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AACxD,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,wCAAA,EAA2C,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACxD,OAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,KAAA,GAAQ,KAAA;AACrB,UAAA;AAAA,QAEF,KAAK,SAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AACxD,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,0CAAA,EAA6C,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cAC1D,SAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,OAAA,GAAU,KAAA;AACvB,UAAA;AAAA,QAEF,KAAK,KAAA;AACH,UAAA,IAAI,OAAO,KAAA,KAAU,QAAA,IAAY,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AACxD,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,sCAAA,EAAyC,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACtD,KAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,GAAA,GAAM,KAAA;AACnB,UAAA;AAAA,QAEF,KAAK,cAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,yCAAA,EAA4C,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACzD,cAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,YAAA,GAAe,KAAA;AAC5B,UAAA;AAAA,QAEF,KAAK,WAAA;AACH,UAAA,IAAI,OAAO,UAAU,SAAA,EAAW;AAC9B,YAAA,MAAM,IAAI,qBAAA;AAAA,cACR,CAAA,sCAAA,EAAyC,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA;AAAA,cACtD,WAAA;AAAA,cACA;AAAA,aACF;AAAA,UACF;AACA,UAAA,IAAA,CAAK,QAAQ,SAAA,GAAY,KAAA;AACzB,UAAA;AAAA,QAEF,KAAK,QAAA;AACH,UAAA,IACE,KAAA,KAAU,UACV,KAAA,KAAU,KAAA,IACV,UAAU,UAAA,IACV,KAAA,KAAU,MAAA,IACV,KAAA,KAAU,KAAA,EACV;AACA,YAAA,MAAM,IAAI,sBAAsB,CAAA,gBAAA,EAAmB,MAAA,CAAO,KAAK,CAAC,CAAA,CAAA,EAAI,UAAU,KAAK,CAAA;AAAA,UACrF;AACA,UAAA,IAAA,CAAK,QAAQ,MAAA,GAAS,KAAA;AACtB,UAAA;AAAA,QAEF;AACE,UAAA,MAAM,IAAI,qBAAA,CAAsB,CAAA,mBAAA,EAAsB,GAAG,CAAA,CAAA,EAAI,KAAK,KAAK,CAAA;AAAA;AAC3E,IACF;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,IAAA,GAAe;AACjB,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,KAAA,EAAO;AAC/B,MAAA,MAAM,IAAI,eAAe,6CAA6C,CAAA;AAAA,IACxE;AACA,IAAA,OAAO,IAAA,CAAK,KAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,SAAA,GAAoB;AACtB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,0CAA0C,CAAA;AAAA,IACrE;AACA,IAAA,OAAO,KAAK,UAAA,IAAc,CAAA;AAAA,EAC5B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,KAAA,GAA4B;AAC9B,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,sCAAsC,CAAA;AAAA,IACjE;AACA,IAAA,OAAO,IAAA,CAAK,MAAA;AAAA,EACd;AACF;;;ACjxBO,IAAM,OAAN,MAAW;AAAA;AAAA,EAEC,WAAA;AAAA;AAAA,EAGA,UAAA;AAAA;AAAA,EAGA,YAAA;AAAA;AAAA,EAGA,KAAA;AAAA;AAAA,EAGA,iBAAA;AAAA;AAAA,EAGA,qBAAA;AAAA;AAAA,EAGA,WAAA;AAAA;AAAA,EAGA,WAAA;AAAA;AAAA,EAGA,MAAA;AAAA;AAAA,EAGA,eAAA;AAAA;AAAA,EAGA,oBAAA;AAAA;AAAA,EAGA,eAAA;AAAA;AAAA,EAGT,YAAwB,EAAC;AAAA;AAAA,EAGzB,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAiBI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,CAAA;AAC1C,IAAA,IAAA,CAAK,UAAA,GAAa,QAAQ,UAAA,IAAc,EAAA;AACxC,IAAA,IAAA,CAAK,YAAA,GAAe,QAAQ,YAAA,IAAgB,GAAA;AAC5C,IAAA,IAAA,CAAK,KAAA,GAAQ,QAAQ,KAAA,IAAS,GAAA;AAC9B,IAAA,IAAA,CAAK,iBAAA,GAAoB,QAAQ,iBAAA,IAAqB,EAAA;AACtD,IAAA,MAAM,qBAAA,GAAwB,QAAQ,qBAAA,IAAyB,GAAA;AAC/D,IAAA,IAAA,CAAK,qBAAA,GAAwB,IAAA,CAAK,GAAA,CAAI,qBAAA,EAAuB,KAAK,KAAK,CAAA;AACvE,IAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAC3B,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,IAAA;AAC1C,IAAA,IAAA,CAAK,MAAA,GAAS,QAAQ,MAAA,IAAU,OAAA;AAChC,IAAA,IAAA,CAAK,eAAA,GAAkB,QAAQ,eAAA,IAAmB,GAAA;AAClD,IAAA,IAAA,CAAK,oBAAA,GACH,OAAA,CAAQ,oBAAA,IAAwB,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,IAAA,CAAK,KAAA,CAAM,IAAA,CAAK,UAAA,GAAa,CAAC,CAAC,CAAA;AAC7E,IAAA,IAAA,CAAK,eAAA,GAAkB,OAAA,CAAQ,eAAA,IAAmB,IAAA,CAAK,GAAA,CAAI,EAAA,EAAI,IAAA,CAAK,KAAA,CAAM,IAAA,CAAK,UAAA,GAAa,CAAC,CAAC,CAAA;AAE9F,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,eAAe,CAAA,EAAG;AAChE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,8BAAA;AAAA,QACA,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,UAAU,CAAA,IAAK,IAAA,CAAK,cAAc,CAAA,EAAG;AAC7D,MAAA,MAAM,IAAI,qBAAA,CAAsB,6BAAA,EAA+B,YAAA,EAAc,KAAK,UAAU,CAAA;AAAA,IAC9F;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,YAAY,CAAA,IAAK,IAAA,CAAK,gBAAgB,CAAA,EAAG;AACjE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,+BAAA;AAAA,QACA,cAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,KAAK,CAAA,IAAK,IAAA,CAAK,SAAS,CAAA,EAAG;AACpD,MAAA,MAAM,IAAI,qBAAA,CAAsB,kCAAA,EAAoC,OAAA,EAAS,KAAK,KAAK,CAAA;AAAA,IACzF;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,iBAAiB,CAAA,IAAK,IAAA,CAAK,qBAAqB,CAAA,EAAG;AAC3E,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,oCAAA;AAAA,QACA,mBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,MAAA,CAAO,SAAA,CAAU,qBAAqB,CAAA,IAAK,wBAAwB,CAAA,EAAG;AACzE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,+CAAA;AAAA,QACA,uBAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,eAAe,CAAA,EAAG;AAC/D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,8BAAA;AAAA,QACA,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,IAAA,CAAK,MAAA,KAAW,OAAA,IAAW,IAAA,CAAK,WAAW,aAAA,EAAe;AAC5D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,yCAAA;AAAA,QACA,QAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,mBAAmB,CAAA,EAAG;AACxE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,4CAAA;AAAA,QACA,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,oBAAoB,CAAA,IAAK,IAAA,CAAK,wBAAwB,CAAA,EAAG;AAClF,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,iDAAA;AAAA,QACA,sBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,mBAAmB,CAAA,EAAG;AACxE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,4CAAA;AAAA,QACA,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,IAAa,CAAC,OAAO,QAAA,CAAS,OAAA,CAAQ,WAAW,CAAA,EAAG;AAC9E,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,qCAAA;AAAA,QACA,aAAA;AAAA,QACA,OAAA,CAAQ;AAAA,OACV;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA,EAKQ,iBAAiB,CAAA,EAA2B;AAClD,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AACZ,IAAA,MAAM,YAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,MAAM,CAAA,EAAG;AACX,UAAA,GAAA,CAAI,KAAK,CAAC,CAAA;AAAA,QACZ,CAAA,MAAO;AACL,UAAA,IAAI,IAAA,GAAO,CAAA;AACX,UAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,UAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,UAAA,IAAI,MAAM,EAAA,EAAI;AACZ,YAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,EAAA,CAAG,QAAQ,CAAA,EAAA,EAAK;AAClC,cAAA,MAAM,QAAQ,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AACtC,cAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,YACjB;AAAA,UACF;AACA,UAAA,GAAA,CAAI,KAAK,IAAI,CAAA;AAAA,QACf;AAAA,MACF;AACA,MAAA,SAAA,CAAU,KAAK,GAAG,CAAA;AAAA,IACpB;AAEA,IAAA,OAAO,SAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKQ,sBAAA,CAAuB,GAAa,CAAA,EAAqB;AAC/D,IAAA,IAAI,IAAA,GAAO,CAAA;AACX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,QAAQ,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,QAAQ,CAAA,CAAE,CAAC,KAAK,CAAA,KAAM,CAAA,CAAE,CAAC,CAAA,IAAK,CAAA,CAAA;AACpC,MAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,IACjB;AACA,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKQ,aAAA,CAAc,CAAA,EAAW,OAAA,EAAiB,CAAA,EAAW,GAAA,EAA6B;AACxF,IAAA,MAAM,MAAA,uBAAa,GAAA,EAAY;AAC/B,IAAA,MAAM,WAAA,GAAc,IAAI,EAAA,GAAK,GAAA;AAC7B,IAAA,IAAI,QAAA,GAAW,CAAA;AAEf,IAAA,OAAO,MAAA,CAAO,IAAA,GAAO,CAAA,IAAK,QAAA,GAAW,WAAA,EAAa;AAChD,MAAA,MAAM,MAAM,IAAA,CAAK,KAAA,CAAM,GAAA,EAAI,IAAK,IAAI,CAAA,CAAE,CAAA;AACtC,MAAA,MAAM,GAAA,GAAM,GAAA,IAAO,OAAA,GAAU,GAAA,GAAM,CAAA,GAAI,GAAA;AACvC,MAAA,MAAA,CAAO,IAAI,GAAG,CAAA;AACd,MAAA,QAAA,IAAY,CAAA;AAAA,IACd;AAEA,IAAA,IAAI,MAAA,CAAO,OAAO,CAAA,EAAG;AACnB,MAAA,KAAA,IAAS,IAAI,CAAA,EAAG,CAAA,GAAI,KAAK,MAAA,CAAO,IAAA,GAAO,GAAG,CAAA,EAAA,EAAK;AAC7C,QAAA,IAAI,MAAM,OAAA,EAAS;AACjB,UAAA,MAAA,CAAO,IAAI,CAAC,CAAA;AAAA,QACd;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,KAAA,CAAM,KAAK,MAAM,CAAA;AAAA,EAC1B;AAAA;AAAA;AAAA;AAAA,EAKQ,2BAAA,CACN,CAAA,EACA,UAAA,EACA,CAAA,EACA,GAAA,EACU;AACV,IAAA,MAAM,MAAA,uBAAa,GAAA,EAAY;AAC/B,IAAA,MAAM,WAAA,GAAc,IAAI,EAAA,GAAK,GAAA;AAC7B,IAAA,IAAI,QAAA,GAAW,CAAA;AAEf,IAAA,OAAO,MAAA,CAAO,IAAA,GAAO,CAAA,IAAK,QAAA,GAAW,WAAA,EAAa;AAChD,MAAA,MAAM,GAAA,GAAM,IAAA,CAAK,KAAA,CAAM,GAAA,KAAQ,CAAC,CAAA;AAChC,MAAA,IAAI,CAAC,UAAA,CAAW,GAAA,CAAI,GAAG,CAAA,EAAG;AACxB,QAAA,MAAA,CAAO,IAAI,GAAG,CAAA;AAAA,MAChB;AACA,MAAA,QAAA,IAAY,CAAA;AAAA,IACd;AAEA,IAAA,IAAI,MAAA,CAAO,OAAO,CAAA,EAAG;AACnB,MAAA,KAAA,IAAS,IAAI,CAAA,EAAG,CAAA,GAAI,KAAK,MAAA,CAAO,IAAA,GAAO,GAAG,CAAA,EAAA,EAAK;AAC7C,QAAA,IAAI,CAAC,UAAA,CAAW,GAAA,CAAI,CAAC,CAAA,EAAG;AACtB,UAAA,MAAA,CAAO,IAAI,CAAC,CAAA;AAAA,QACd;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,KAAA,CAAM,KAAK,MAAM,CAAA;AAAA,EAC1B;AAAA;AAAA;AAAA;AAAA,EAKQ,qBAAqB,SAAA,EAAmC;AAC9D,IAAA,MAAM,IAAI,SAAA,CAAU,MAAA;AACpB,IAAA,MAAM,aAAA,GAAgB,IAAA,CAAK,GAAA,CAAI,IAAA,CAAK,UAAU,CAAA;AAC9C,IAAA,MAAM,IAAgB,EAAC;AAEvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,MAAgB,IAAI,KAAA,CAAM,CAAC,CAAA,CAAE,KAAK,CAAC,CAAA;AAGzC,MAAA,IAAI,QAAA,GAAW,KAAA;AACf,MAAA,IAAI,QAAA,GAAW,IAAA;AACf,MAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,MAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,EAAA,EAAI,IAAA,EAAA,EAAQ;AAEpC,QAAA,IAAIC,OAAAA,GAAS,CAAA;AACb,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,IAAI,MAAM,CAAA,EAAG;AACX,YAAA,MAAM,OAAA,GAAU,UAAU,CAAC,CAAA;AAC3B,YAAA,MAAM,IAAA,GAAO,OAAA,GAAW,OAAA,CAAQ,CAAC,KAAK,CAAA,GAAK,CAAA;AAC3C,YAAAA,WAAU,IAAA,CAAK,GAAA,CAAI,CAAC,IAAA,IAAQ,CAAA,GAAI,QAAQ,KAAA,CAAM,CAAA;AAAA,UAChD;AAAA,QACF;AAGA,QAAA,IAAI,OAAA,GAAU,CAAA;AACd,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,UAAA,IAAI,MAAM,CAAA,EAAG;AACX,YAAA,MAAM,OAAA,GAAU,UAAU,CAAC,CAAA;AAC3B,YAAA,MAAM,IAAA,GAAO,OAAA,GAAW,OAAA,CAAQ,CAAC,KAAK,CAAA,GAAK,CAAA;AAC3C,YAAA,MAAM,GAAA,GAAM,KAAK,GAAA,CAAI,CAAC,QAAQ,CAAA,GAAI,KAAA,GAAQ,KAAA,CAAM,CAAA,IAAKA,OAAAA,GAAS,KAAA,CAAA;AAC9D,YAAA,IAAI,MAAM,KAAA,EAAO;AACf,cAAA,OAAA,IAAW,GAAA,GAAM,IAAA,CAAK,GAAA,CAAI,GAAG,CAAA;AAAA,YAC/B;AAAA,UACF;AAAA,QACF;AAGA,QAAA,IAAI,IAAA,CAAK,GAAA,CAAI,OAAA,GAAU,aAAa,IAAI,IAAA,EAAM;AAC5C,UAAA;AAAA,QACF;AAEA,QAAA,IAAI,UAAU,aAAA,EAAe;AAC3B,UAAA,QAAA,GAAW,KAAA;AACX,UAAA,KAAA,GAAA,CAAS,QAAQ,QAAA,IAAY,CAAA;AAAA,QAC/B,CAAA,MAAO;AACL,UAAA,QAAA,GAAW,KAAA;AACX,UAAA,KAAA,GAAA,CAAS,QAAQ,QAAA,IAAY,CAAA;AAAA,QAC/B;AAAA,MACF;AAGA,MAAA,IAAI,MAAA,GAAS,CAAA;AACb,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,MAAM,CAAA,EAAG;AACX,UAAA,MAAM,OAAA,GAAU,UAAU,CAAC,CAAA;AAC3B,UAAA,MAAM,IAAA,GAAO,OAAA,GAAW,OAAA,CAAQ,CAAC,KAAK,CAAA,GAAK,CAAA;AAC3C,UAAA,MAAA,IAAU,KAAK,GAAA,CAAI,CAAC,IAAA,IAAQ,CAAA,GAAI,QAAQ,KAAA,CAAM,CAAA;AAAA,QAChD;AAAA,MACF;AAEA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,MAAM,CAAA,EAAG;AACX,UAAA,MAAM,OAAA,GAAU,UAAU,CAAC,CAAA;AAC3B,UAAA,MAAM,IAAA,GAAO,OAAA,GAAW,OAAA,CAAQ,CAAC,KAAK,CAAA,GAAK,CAAA;AAC3C,UAAA,GAAA,CAAI,CAAC,CAAA,GAAI,IAAA,CAAK,GAAA,CAAI,CAAC,QAAQ,CAAA,GAAI,KAAA,GAAQ,KAAA,CAAM,CAAA,IAAK,MAAA,GAAS,KAAA,CAAA;AAAA,QAC7D;AAAA,MACF;AAEA,MAAA,CAAA,CAAE,KAAK,GAAG,CAAA;AAAA,IACZ;AAGA,IAAA,MAAM,OAAmB,EAAC;AAC1B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,GAAA,GAAM,CAAA,CAAE,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AACzB,QAAA,MAAM,GAAA,GAAM,CAAA,CAAE,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AACzB,QAAA,GAAA,CAAI,IAAA,CAAA,CAAM,GAAA,GAAM,GAAA,KAAQ,CAAA,GAAI,CAAA,CAAE,CAAA;AAAA,MAChC;AACA,MAAA,IAAA,CAAK,KAAK,GAAG,CAAA;AAAA,IACf;AAEA,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKQ,0BAAA,CACN,CAAA,EACA,SAAA,EACA,GAAA,EACc;AACd,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AACZ,IAAA,MAAM,aAAA,GAAgB,IAAA,CAAK,GAAA,CAAI,IAAA,CAAK,UAAU,CAAA;AAC9C,IAAA,MAAM,OAAqB,EAAC;AAE5B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,kBAAkB,IAAA,CAAK,aAAA,CAAc,CAAA,EAAG,CAAA,EAAG,WAAW,GAAG,CAAA;AAC/D,MAAA,MAAM,YAAY,eAAA,CAAgB,GAAA;AAAA,QAAI,CAAC,CAAA,KACrC,IAAA,CAAK,sBAAA,CAAuB,CAAA,CAAE,CAAC,CAAA,IAAK,EAAC,EAAG,CAAA,CAAE,CAAC,CAAA,IAAK,EAAE;AAAA,OACpD;AAEA,MAAA,IAAI,QAAA,GAAW,KAAA;AACf,MAAA,IAAI,QAAA,GAAW,IAAA;AACf,MAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,MAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,EAAA,EAAI,IAAA,EAAA,EAAQ;AACpC,QAAA,IAAIA,OAAAA,GAAS,CAAA;AACb,QAAA,KAAA,MAAW,QAAQ,SAAA,EAAW;AAC5B,UAAAA,WAAU,IAAA,CAAK,GAAA,CAAI,CAAC,IAAA,IAAQ,CAAA,GAAI,QAAQ,KAAA,CAAM,CAAA;AAAA,QAChD;AAEA,QAAA,IAAI,OAAA,GAAU,CAAA;AACd,QAAA,KAAA,MAAW,QAAQ,SAAA,EAAW;AAC5B,UAAA,MAAM,GAAA,GAAM,KAAK,GAAA,CAAI,CAAC,QAAQ,CAAA,GAAI,KAAA,GAAQ,KAAA,CAAM,CAAA,IAAKA,OAAAA,GAAS,KAAA,CAAA;AAC9D,UAAA,IAAI,MAAM,KAAA,EAAO;AACf,YAAA,OAAA,IAAW,GAAA,GAAM,IAAA,CAAK,GAAA,CAAI,GAAG,CAAA;AAAA,UAC/B;AAAA,QACF;AAEA,QAAA,IAAI,IAAA,CAAK,GAAA,CAAI,OAAA,GAAU,aAAa,IAAI,IAAA,EAAM;AAC5C,UAAA;AAAA,QACF;AAEA,QAAA,IAAI,UAAU,aAAA,EAAe;AAC3B,UAAA,QAAA,GAAW,KAAA;AACX,UAAA,KAAA,GAAA,CAAS,QAAQ,QAAA,IAAY,CAAA;AAAA,QAC/B,CAAA,MAAO;AACL,UAAA,QAAA,GAAW,KAAA;AACX,UAAA,KAAA,GAAA,CAAS,QAAQ,QAAA,IAAY,CAAA;AAAA,QAC/B;AAAA,MACF;AAEA,MAAA,IAAI,MAAA,GAAS,CAAA;AACb,MAAA,KAAA,MAAW,QAAQ,SAAA,EAAW;AAC5B,QAAA,MAAA,IAAU,KAAK,GAAA,CAAI,CAAC,IAAA,IAAQ,CAAA,GAAI,QAAQ,KAAA,CAAM,CAAA;AAAA,MAChD;AAEA,MAAA,MAAM,SAAmB,SAAA,CAAU,GAAA;AAAA,QACjC,CAAC,IAAA,KAAS,IAAA,CAAK,GAAA,CAAI,CAAC,QAAQ,CAAA,GAAI,KAAA,GAAQ,KAAA,CAAM,CAAA,IAAK,MAAA,GAAS,KAAA;AAAA,OAC9D;AAEA,MAAA,IAAA,CAAK,IAAA,CAAK,EAAE,OAAA,EAAS,eAAA,EAAiB,QAAQ,CAAA;AAAA,IAChD;AAEA,IAAA,OAAO,IAAA,CAAK,gBAAA,CAAiB,IAAA,EAAM,CAAC,CAAA;AAAA,EACtC;AAAA;AAAA;AAAA;AAAA,EAKQ,gBAAA,CAAiB,MAAoB,CAAA,EAAyB;AACpE,IAAA,MAAM,IAAA,GAAmC,IAAA,CAAK,GAAA,CAAI,CAAC,GAAA,KAAQ;AACzD,MAAA,MAAM,GAAA,uBAAU,GAAA,EAAoB;AACpC,MAAA,KAAA,IAAS,IAAI,CAAA,EAAG,CAAA,GAAI,GAAA,CAAI,OAAA,CAAQ,QAAQ,CAAA,EAAA,EAAK;AAC3C,QAAA,MAAM,CAAA,GAAI,GAAA,CAAI,OAAA,CAAQ,CAAC,CAAA;AACvB,QAAA,IAAI,MAAM,MAAA,EAAW;AACrB,QAAA,MAAM,GAAA,GAAM,GAAA,CAAI,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA;AAC7B,QAAA,GAAA,CAAI,GAAA,CAAI,GAAG,GAAG,CAAA;AAAA,MAChB;AACA,MAAA,OAAO,GAAA;AAAA,IACT,CAAC,CAAA;AAED,IAAA,MAAM,MAAoB,EAAC;AAC3B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,GAAA,GAAM,KAAK,CAAC,CAAA;AAClB,MAAA,IAAI,CAAC,GAAA,EAAK;AACR,QAAA,GAAA,CAAI,IAAA,CAAK,EAAE,OAAA,EAAS,IAAI,MAAA,EAAQ,IAAI,CAAA;AACpC,QAAA;AAAA,MACF;AACA,MAAA,MAAM,OAAA,GAAU,GAAA,CAAI,OAAA,CAAQ,KAAA,EAAM;AAClC,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,OAAA,CAAQ,QAAQ,CAAA,EAAA,EAAK;AACvC,QAAA,MAAM,CAAA,GAAI,QAAQ,CAAC,CAAA;AACnB,QAAA,IAAI,MAAM,MAAA,EAAW;AACrB,QAAA,MAAM,MAAM,IAAA,CAAK,CAAC,CAAA,EAAG,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA;AAC/B,QAAA,MAAM,MAAM,IAAA,CAAK,CAAC,CAAA,EAAG,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA;AAC/B,QAAA,MAAA,CAAO,IAAA,CAAA,CAAM,GAAA,GAAM,GAAA,KAAQ,CAAA,GAAI,CAAA,CAAE,CAAA;AAAA,MACnC;AACA,MAAA,GAAA,CAAI,IAAA,CAAK,EAAE,OAAA,EAAS,MAAA,EAAQ,CAAA;AAAA,IAC9B;AAEA,IAAA,OAAO,GAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKQ,oBAAoB,CAAA,EAAuB;AACjD,IAAA,MAAM,YAAwB,EAAC;AAG/B,IAAA,IAAI,IAAA,GAAO,IAAA,CAAK,WAAA,IAAe,IAAA,CAAK,GAAA,EAAI;AACxC,IAAA,MAAM,SAAS,MAAc;AAC3B,MAAA,IAAA,GAAQ,IAAA,GAAO,aAAa,KAAA,GAAS,UAAA;AACrC,MAAA,OAAO,IAAA,GAAO,UAAA;AAAA,IAChB,CAAA;AAEA,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AAEzC,QAAA,MAAM,EAAA,GAAK,QAAO,IAAK,IAAA;AACvB,QAAA,MAAM,KAAK,MAAA,EAAO;AAClB,QAAA,MAAM,CAAA,GAAI,IAAA,CAAK,IAAA,CAAK,EAAA,GAAK,KAAK,GAAA,CAAI,EAAE,CAAC,CAAA,GAAI,IAAA,CAAK,GAAA,CAAI,CAAA,GAAI,IAAA,CAAK,KAAK,EAAE,CAAA;AAClE,QAAA,GAAA,CAAI,IAAA,CAAK,IAAI,IAAM,CAAA;AAAA,MACrB;AACA,MAAA,SAAA,CAAU,KAAK,GAAG,CAAA;AAAA,IACpB;AAEA,IAAA,OAAO,SAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKQ,SAAS,CAAA,EAAgD;AAC/D,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AACZ,IAAA,MAAM,IAAgB,EAAC;AACvB,IAAA,IAAI,IAAA,GAAO,CAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,MAAM,CAAA,EAAG;AACX,UAAA,GAAA,CAAI,KAAK,CAAC,CAAA;AAAA,QACZ,CAAA,MAAO;AACL,UAAA,IAAI,IAAA,GAAO,CAAA;AACX,UAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,UAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,UAAA,IAAI,MAAM,EAAA,EAAI;AACZ,YAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,cAAA,MAAM,QAAQ,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AACtC,cAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,YACjB;AAAA,UACF;AAEA,UAAA,MAAM,GAAA,GAAM,KAAK,CAAA,GAAI,IAAA,CAAA;AACrB,UAAA,GAAA,CAAI,KAAK,GAAG,CAAA;AACZ,UAAA,IAAA,IAAQ,GAAA;AAAA,QACV;AAAA,MACF;AACA,MAAA,CAAA,CAAE,KAAK,GAAG,CAAA;AAAA,IACZ;AAEA,IAAA,OAAO,EAAE,GAAG,IAAA,EAAK;AAAA,EACnB;AAAA;AAAA;AAAA;AAAA,EAKQ,cAAA,CACN,CAAA,EACA,SAAA,EACA,eAAA,EACA,GAAA,EACqC;AACrC,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AACZ,IAAA,MAAM,OAAoB,EAAC;AAC3B,IAAA,IAAI,IAAA,GAAO,CAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,eAAA,GAAkB,SAAA,CAAU,CAAC,CAAA,EAAG,WAAW,EAAC;AAClD,MAAA,MAAM,SAAA,GAAY,IAAI,GAAA,CAAY,eAAe,CAAA;AACjD,MAAA,SAAA,CAAU,IAAI,CAAC,CAAA;AAEf,MAAA,MAAM,YAAY,IAAA,CAAK,2BAAA,CAA4B,CAAA,EAAG,SAAA,EAAW,iBAAiB,GAAG,CAAA;AACrF,MAAA,MAAM,OAAA,GAAU,eAAA,CAAgB,MAAA,CAAO,SAAS,CAAA;AAChD,MAAA,MAAM,UAAoB,EAAC;AAE3B,MAAA,MAAM,EAAA,GAAK,CAAA,CAAE,CAAC,CAAA,IAAK,EAAC;AACpB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,OAAA,CAAQ,QAAQ,CAAA,EAAA,EAAK;AACvC,QAAA,MAAM,CAAA,GAAI,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA;AACxB,QAAA,MAAM,EAAA,GAAK,CAAA,CAAE,CAAC,CAAA,IAAK,EAAC;AACpB,QAAA,MAAM,IAAA,GAAO,IAAA,CAAK,sBAAA,CAAuB,EAAA,EAAI,EAAE,CAAA;AAC/C,QAAA,MAAM,GAAA,GAAM,KAAK,CAAA,GAAI,IAAA,CAAA;AACrB,QAAA,OAAA,CAAQ,KAAK,GAAG,CAAA;AAChB,QAAA,IAAA,IAAQ,GAAA;AAAA,MACV;AAEA,MAAA,IAAA,CAAK,IAAA,CAAK,EAAE,OAAA,EAAS,OAAA,EAAS,CAAA;AAAA,IAChC;AAEA,IAAA,OAAO,EAAE,MAAM,IAAA,EAAK;AAAA,EACtB;AAAA;AAAA;AAAA;AAAA,EAKQ,gBAAA,CAAiB,CAAA,EAAe,CAAA,EAAe,IAAA,EAAc,CAAA,EAA2B;AAC9F,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AACZ,IAAA,MAAM,YAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,OAAiB,IAAI,KAAA,CAAM,KAAK,WAAW,CAAA,CAAE,KAAK,CAAC,CAAA;AAEzD,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,IAAI,MAAM,CAAA,EAAG;AACX,UAAA,MAAM,GAAA,GAAM,CAAA,CAAE,CAAC,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA;AACzB,UAAA,MAAM,OAAO,CAAA,CAAE,CAAC,IAAI,CAAC,CAAA,IAAK,MAAM,IAAA,GAAO,KAAA,CAAA;AAEvC,UAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,UAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,UAAA,IAAI,MAAM,EAAA,EAAI;AAEZ,YAAA,IAAI,IAAA,GAAO,CAAA;AACX,YAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,cAAA,MAAM,QAAQ,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AACtC,cAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,YACjB;AACA,YAAA,MAAM,IAAA,GAAA,CAAQ,GAAA,GAAM,GAAA,KAAQ,CAAA,IAAK,CAAA,GAAI,IAAA,CAAA,CAAA;AAErC,YAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,cAAA,IAAA,CAAK,CAAC,CAAA,GAAA,CAAK,IAAA,CAAK,CAAC,KAAK,CAAA,IAAK,CAAA,GAAI,IAAA,IAAA,CAAS,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA,CAAA;AAAA,YAClE;AAAA,UACF;AAAA,QACF;AAAA,MACF;AAEA,MAAA,SAAA,CAAU,KAAK,IAAI,CAAA;AAAA,IACrB;AAEA,IAAA,OAAO,SAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKQ,sBAAA,CACN,KAAA,EACA,KAAA,EACA,IAAA,EACA,GACA,YAAA,EACY;AACZ,IAAA,MAAM,IAAI,CAAA,CAAE,MAAA;AACZ,IAAA,MAAM,YAAwB,EAAC;AAC/B,IAAA,MAAM,QAAQ,IAAA,GAAO,KAAA;AAErB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,OAAiB,IAAI,KAAA,CAAM,KAAK,WAAW,CAAA,CAAE,KAAK,CAAC,CAAA;AACzD,MAAA,MAAM,EAAA,GAAK,CAAA,CAAE,CAAC,CAAA,IAAK,EAAC;AACpB,MAAA,MAAM,IAAA,GAAO,MAAM,CAAC,CAAA;AACpB,MAAA,MAAM,IAAA,GAAO,MAAM,CAAC,CAAA;AAEpB,MAAA,IAAI,IAAA,EAAM;AACR,QAAA,KAAA,IAAS,MAAM,CAAA,EAAG,GAAA,GAAM,IAAA,CAAK,OAAA,CAAQ,QAAQ,GAAA,EAAA,EAAO;AAClD,UAAA,MAAM,CAAA,GAAI,IAAA,CAAK,OAAA,CAAQ,GAAG,CAAA,IAAK,CAAA;AAC/B,UAAA,IAAI,MAAM,CAAA,EAAG;AACb,UAAA,MAAM,EAAA,GAAK,CAAA,CAAE,CAAC,CAAA,IAAK,EAAC;AAEpB,UAAA,MAAM,OAAA,GAAU,IAAA,EAAM,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA;AAChC,UAAA,MAAM,MAAM,OAAA,GAAU,YAAA;AACtB,UAAA,MAAM,OAAA,GAAU,IAAA,CAAK,OAAA,CAAQ,GAAG,CAAA,IAAK,CAAA;AACrC,UAAA,MAAM,MAAM,OAAA,GAAU,KAAA;AAEtB,UAAA,MAAM,IAAA,GAAA,CAAQ,MAAM,GAAA,IAAO,OAAA;AAC3B,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,YAAA,IAAA,CAAK,CAAC,CAAA,GAAA,CAAK,IAAA,CAAK,CAAC,KAAK,CAAA,IAAK,CAAA,GAAI,IAAA,IAAA,CAAS,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA,CAAA;AAAA,UAClE;AAAA,QACF;AAAA,MACF;AAEA,MAAA,SAAA,CAAU,KAAK,IAAI,CAAA;AAAA,IACrB;AAEA,IAAA,OAAO,SAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,aAAa,CAAA,EAAmB;AAC9B,IAAA,6BAAA,CAA8B,CAAC,CAAA;AAE/B,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAI,WAAW,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,qBAAA,CAAsB,mCAAA,EAAqC,UAAA,EAAY,QAAQ,CAAA;AAAA,IAC3F;AACA,IAAA,IAAI,IAAA,CAAK,cAAc,QAAA,EAAU;AAC/B,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,4DAAA,EAA+D,IAAA,CAAK,UAAU,CAAA,YAAA,EAAe,QAAQ,CAAA,CAAA;AAAA,QACrG,YAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,IAAA,CAAK,MAAA,KAAW,OAAA,IAAW,QAAA,GAAW,KAAK,eAAA,EAAiB;AAC9D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,kDAAA,EAAqD,IAAA,CAAK,eAAe,CAAA,qBAAA,EAAwB,QAAQ,CAAA,uDAAA,CAAA;AAAA,QACzG,UAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AAGA,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AAAA,IAChB;AAEA,IAAA,MAAM,QAAA,GAAW,IAAA,CAAK,WAAA,IAAe,IAAA,CAAK,GAAA,EAAI;AAC9C,IAAA,MAAM,OAAO,MAAM;AACjB,MAAA,IAAI,QAAQ,QAAA,KAAa,CAAA;AACzB,MAAA,OAAO,MAAc;AACnB,QAAA,KAAA,GAAS,KAAA,GAAQ,UAAU,UAAA,KAAgB,CAAA;AAC3C,QAAA,OAAO,QAAQ,CAAA,IAAK,EAAA;AAAA,MACtB,CAAA;AAAA,IACF,CAAA,GAAG;AAEH,IAAA,MAAM,cAAA,GAAiB,KAAK,MAAA,KAAW,aAAA;AACvC,IAAA,MAAM,gBAAgB,IAAA,CAAK,GAAA,CAAI,IAAA,CAAK,oBAAA,EAAsB,WAAW,CAAC,CAAA;AACtE,IAAA,MAAM,qBAAqB,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,QAAA,GAAW,IAAI,aAAa,CAAA;AACnE,IAAA,MAAM,aAAA,GAAgB,IAAA,CAAK,GAAA,CAAI,IAAA,CAAK,iBAAiB,kBAAkB,CAAA;AACvE,IAAA,IAAI,cAAA,IAAkB,gBAAgB,CAAA,EAAG;AACvC,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,8DAAA;AAAA,QACA,sBAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AACA,IAAA,IAAI,cAAA,IAAkB,IAAA,CAAK,UAAA,IAAc,aAAA,EAAe;AACtD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,uEAAA,EAA0E,IAAA,CAAK,UAAU,CAAA,uBAAA,EAA0B,aAAa,CAAA,CAAA;AAAA,QAChI,YAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAGA,IAAA,MAAM,MAAA,GAAS,iBAAiB,IAAA,GAAO,IAAA,CAAK,qBAAqB,IAAA,CAAK,gBAAA,CAAiB,KAAK,CAAC,CAAA;AAC7F,IAAA,MAAM,UAAU,cAAA,GACZ,IAAA,CAAK,2BAA2B,KAAA,EAAO,aAAA,EAAe,GAAG,CAAA,GACzD,IAAA;AAGJ,IAAA,MAAM,CAAA,GAAI,IAAA,CAAK,mBAAA,CAAoB,QAAQ,CAAA;AAG3C,IAAA,MAAM,aAAyB,EAAC;AAChC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,UAAA,CAAW,IAAA,CAAK,IAAI,KAAA,CAAM,IAAA,CAAK,WAAW,CAAA,CAAE,IAAA,CAAK,CAAC,CAAC,CAAA;AAAA,IACrD;AAEA,IAAA,MAAM,QAAA,GAAW,GAAA;AACjB,IAAA,MAAM,aAAA,GAAgB,GAAA;AAEtB,IAAA,MAAM,QACJ,cAAA,IAAkB,OAAA,GACd,OAAA,CAAQ,GAAA,CAAI,CAAC,GAAA,KAAQ;AACnB,MAAA,MAAM,GAAA,uBAAU,GAAA,EAAoB;AACpC,MAAA,KAAA,IAAS,IAAI,CAAA,EAAG,CAAA,GAAI,GAAA,CAAI,OAAA,CAAQ,QAAQ,CAAA,EAAA,EAAK;AAC3C,QAAA,MAAM,CAAA,GAAI,GAAA,CAAI,OAAA,CAAQ,CAAC,CAAA;AACvB,QAAA,IAAI,MAAM,MAAA,EAAW;AACrB,QAAA,MAAM,GAAA,GAAM,GAAA,CAAI,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA;AAC7B,QAAA,GAAA,CAAI,GAAA,CAAI,GAAG,GAAG,CAAA;AAAA,MAChB;AACA,MAAA,OAAO,GAAA;AAAA,IACT,CAAC,IACD,EAAC;AAGP,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,IAAA,CAAK,OAAO,IAAA,EAAA,EAAQ;AAE5C,MAAA,MAAM,YAAA,GAAe,IAAA,GAAO,IAAA,CAAK,qBAAA,GAAwB,KAAK,iBAAA,GAAoB,CAAA;AAElF,MAAA,IAAI,SAAA;AACJ,MAAA,IAAI,kBAAkB,OAAA,EAAS;AAC7B,QAAA,MAAM,EAAE,IAAA,EAAM,KAAA,EAAO,IAAA,EAAK,GAAI,KAAK,cAAA,CAAe,CAAA,EAAG,OAAA,EAAS,aAAA,EAAe,GAAG,CAAA;AAChF,QAAA,SAAA,GAAY,KAAK,sBAAA,CAAuB,KAAA,EAAO,KAAA,EAAO,IAAA,EAAM,GAAG,YAAY,CAAA;AAAA,MAC7E,CAAA,MAAO;AAEL,QAAA,MAAM,QAAoB,EAAC;AAC3B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,MAAM,MAAgB,EAAC;AACvB,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,YAAA,GAAA,CAAI,MAAM,MAAA,GAAS,CAAC,IAAI,CAAC,CAAA,IAAK,KAAK,YAAY,CAAA;AAAA,UACjD;AACA,UAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AAAA,QAChB;AAGA,QAAA,MAAM,EAAE,CAAA,EAAG,IAAA,EAAK,GAAI,IAAA,CAAK,SAAS,CAAC,CAAA;AAGnC,QAAA,SAAA,GAAY,IAAA,CAAK,gBAAA,CAAiB,KAAA,EAAO,CAAA,EAAG,MAAM,CAAC,CAAA;AAAA,MACrD;AAGA,MAAA,IAAI,QAAA,GAAW,CAAA;AACf,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,IAAA,GAAO,UAAU,CAAC,CAAA;AACxB,QAAA,IAAI,IAAA,EAAM;AACR,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,YAAA,QAAA,IAAA,CAAa,IAAA,CAAK,CAAC,CAAA,IAAK,CAAA,KAAM,CAAA;AAAA,UAChC;AAAA,QACF;AAAA,MACF;AACA,MAAA,QAAA,GAAW,IAAA,CAAK,KAAK,QAAQ,CAAA;AAE7B,MAAA,IAAI,QAAA,GAAW,KAAK,WAAA,EAAa;AAC/B,QAAA;AAAA,MACF;AAGA,MAAA,MAAM,eAAA,GAAkB,IAAA,GAAO,IAAA,CAAK,qBAAA,GAAwB,QAAA,GAAW,aAAA;AAGvE,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,QAAA,MAAM,IAAA,GAAO,UAAU,CAAC,CAAA;AACxB,QAAA,MAAM,GAAA,GAAM,WAAW,CAAC,CAAA;AAExB,QAAA,IAAI,EAAA,IAAM,QAAQ,GAAA,EAAK;AACrB,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,YAAA,GAAA,CAAI,CAAC,CAAA,GAAI,eAAA,IAAmB,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA,CAAA,GAAK,IAAA,CAAK,YAAA,IAAgB,IAAA,CAAK,CAAC,CAAA,IAAK,CAAA,CAAA;AAC3E,YAAA,EAAA,CAAG,CAAC,KAAK,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,UACpC;AAAA,QACF;AAAA,MACF;AAGA,MAAA,MAAM,SAAmB,IAAI,KAAA,CAAM,KAAK,WAAW,CAAA,CAAE,KAAK,CAAC,CAAA;AAC3D,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,QAAA,IAAI,EAAA,EAAI;AACN,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,YAAA,MAAA,CAAO,CAAC,KAAK,MAAA,CAAO,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,UAC3C;AAAA,QACF;AAAA,MACF;AACA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,QAAA,MAAA,CAAO,CAAC,CAAA,GAAA,CAAK,MAAA,CAAO,CAAC,KAAK,CAAA,IAAK,QAAA;AAAA,MACjC;AACA,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,EAAA,GAAK,EAAE,CAAC,CAAA;AACd,QAAA,IAAI,EAAA,EAAI;AACN,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,YAAA,EAAA,CAAG,CAAC,KAAK,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,UACvC;AAAA,QACF;AAAA,MACF;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,SAAA,GAAY,CAAA;AACjB,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,OAAO,CAAC,CAAA;AAAA,EACjB;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,CAAA,EAAiB;AACnB,IAAA,IAAA,CAAK,aAAa,CAAC,CAAA;AACnB,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA,EAKA,IAAI,eAAA,GAA0B;AAC5B,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,gDAAgD,CAAA;AAAA,IAC3E;AACA,IAAA,OAAO,MAAA,CAAO,KAAK,SAAS,CAAA;AAAA,EAC9B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,YAAY,IAAA,CAAK,UAAA;AAAA,MACjB,cAAc,IAAA,CAAK,YAAA;AAAA,MACnB,OAAO,IAAA,CAAK,KAAA;AAAA,MACZ,mBAAmB,IAAA,CAAK,iBAAA;AAAA,MACxB,uBAAuB,IAAA,CAAK,qBAAA;AAAA,MAC5B,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,QAAQ,IAAA,CAAK,MAAA;AAAA,MACb,iBAAiB,IAAA,CAAK,eAAA;AAAA,MACtB,sBAAsB,IAAA,CAAK,oBAAA;AAAA,MAC3B,iBAAiB,IAAA,CAAK;AAAA,KACxB;AAAA,EACF;AACF;;;ACr1BO,IAAM,aAAN,MAAuC;AAAA,EAC3B,YAAA;AAAA,EAET,QAAA;AAAA,EACA,WAAA;AAAA,EACA,MAAA;AAAA;AAAA,EACA,IAAA;AAAA;AAAA,EACA,YAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQjB,WAAA,CACE,OAAA,GAEI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,YAAA,GAAe,QAAQ,YAAA,IAAgB,IAAA;AAC5C,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,YAAY,CAAA,IAAK,IAAA,CAAK,eAAe,CAAA,EAAG;AAChE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,2CAAA;AAAA,QACA,cAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAeA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAEtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAA,CAAK,YAAA,GAAe,SAAA;AAGpB,IAAA,MAAM,QAAA,uBAAe,GAAA,EAAY;AACjC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,QAAA,CAAS,GAAA,CAAI,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IAC3C;AACA,IAAA,IAAA,CAAK,QAAA,GAAW,KAAA,CAAM,IAAA,CAAK,QAAQ,CAAA,CAAE,KAAK,CAAC,CAAA,EAAG,CAAA,KAAM,CAAA,GAAI,CAAC,CAAA;AAGzD,IAAA,IAAA,CAAK,WAAA,uBAAkB,GAAA,EAAI;AAC3B,IAAA,KAAA,MAAW,GAAA,IAAO,KAAK,QAAA,EAAU;AAC/B,MAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,IAAI,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAC,CAAC,MAAM,GAAA,EAAK;AACxC,UAAA,KAAA,EAAA;AAAA,QACF;AAAA,MACF;AACA,MAAA,IAAA,CAAK,WAAA,CAAY,GAAA,CAAI,GAAA,EAAK,KAAA,GAAQ,QAAQ,CAAA;AAAA,IAC5C;AAGA,IAAA,IAAA,CAAK,MAAA,uBAAa,GAAA,EAAI;AACtB,IAAA,IAAA,CAAK,IAAA,uBAAW,GAAA,EAAI;AAEpB,IAAA,KAAA,MAAW,GAAA,IAAO,KAAK,QAAA,EAAU;AAC/B,MAAA,MAAM,eAA2B,EAAC;AAElC,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,IAAI,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAC,CAAC,MAAM,GAAA,EAAK;AACxC,UAAA,MAAM,SAAmB,EAAC;AAC1B,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,UAC1D;AACA,UAAA,YAAA,CAAa,KAAK,MAAM,CAAA;AAAA,QAC1B;AAAA,MACF;AAEA,MAAA,MAAM,QAAkB,EAAC;AACzB,MAAA,MAAM,YAAsB,EAAC;AAE7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAElC,QAAA,IAAI,GAAA,GAAM,CAAA;AACV,QAAA,KAAA,MAAW,UAAU,YAAA,EAAc;AACjC,UAAA,GAAA,IAAO,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA;AAAA,QACtB;AACA,QAAA,MAAMC,KAAAA,GAAO,MAAM,YAAA,CAAa,MAAA;AAChC,QAAA,KAAA,CAAM,KAAKA,KAAI,CAAA;AAGf,QAAA,IAAI,MAAA,GAAS,CAAA;AACb,QAAA,KAAA,MAAW,UAAU,YAAA,EAAc;AACjC,UAAA,MAAM,IAAA,GAAA,CAAQ,MAAA,CAAO,CAAC,CAAA,IAAK,CAAA,IAAKA,KAAAA;AAChC,UAAA,MAAA,IAAU,IAAA,GAAO,IAAA;AAAA,QACnB;AACA,QAAA,MAAM,QAAA,GAAW,SAAS,YAAA,CAAa,MAAA;AACvC,QAAA,IAAI,QAAA,KAAa,CAAA,IAAK,IAAA,CAAK,YAAA,KAAiB,CAAA,EAAG;AAC7C,UAAA,MAAM,IAAI,mBAAA;AAAA,YACR;AAAA,WACF;AAAA,QACF;AACA,QAAA,SAAA,CAAU,KAAK,QAAQ,CAAA;AAAA,MACzB;AAEA,MAAA,IAAA,CAAK,MAAA,CAAO,GAAA,CAAI,GAAA,EAAK,KAAK,CAAA;AAC1B,MAAA,IAAA,CAAK,IAAA,CAAK,GAAA,CAAI,GAAA,EAAK,SAAS,CAAA;AAAA,IAC9B;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,6CAA6C,CAAA;AAAA,IACxE;AAEA,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,YAAA,CAAa,CAAC,CAAA;AACjC,IAAA,MAAM,QAAA,GAAW,KAAA,CAAM,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACnC,IAAA,MAAM,QAAA,GAAW,KAAA,CAAM,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACnC,IAAA,MAAM,cAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,IAAI,OAAA,GAAU,EAAA;AACd,MAAA,IAAI,QAAA,GAAW,CAAA;AAEf,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,IAAA,GAAO,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAA,GAAI,QAAA,GAAW,CAAC,CAAC,CAAA;AAC/D,QAAA,IAAI,OAAO,OAAA,EAAS;AAClB,UAAA,OAAA,GAAU,IAAA;AACV,UAAA,QAAA,GAAW,IAAA,CAAK,QAAA,GAAW,CAAC,CAAA,IAAK,CAAA;AAAA,QACnC;AAAA,MACF;AAEA,MAAA,WAAA,CAAY,KAAK,QAAQ,CAAA;AAAA,IAC3B;AAEA,IAAA,OAAO,MAAA,CAAO,WAAA,EAAa,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EAC/C;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,aAAa,CAAA,EAAmB;AAC9B,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,QAAA,IAAY,CAAC,IAAA,CAAK,WAAA,IAAe,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,IAAA,EAAM;AACrF,MAAA,MAAM,IAAI,eAAe,6CAA6C,CAAA;AAAA,IACxE;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,YAAY,CAAA;AAE7D,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,MAAM,gBAA4B,EAAC;AAEnC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,WAAqB,EAAC;AAE5B,MAAA,KAAA,MAAW,GAAA,IAAO,KAAK,QAAA,EAAU;AAC/B,QAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,WAAA,CAAY,GAAA,CAAI,GAAG,CAAA,IAAK,CAAA;AAC3C,QAAA,MAAM,QAAQ,IAAA,CAAK,MAAA,CAAO,GAAA,CAAI,GAAG,KAAK,EAAC;AACvC,QAAA,MAAM,YAAY,IAAA,CAAK,IAAA,CAAK,GAAA,CAAI,GAAG,KAAK,EAAC;AAEzC,QAAA,IAAI,OAAA,GAAU,IAAA,CAAK,GAAA,CAAI,KAAK,CAAA;AAE5B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,MAAM,CAAA,GAAI,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AACrD,UAAA,MAAMA,KAAAA,GAAO,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACzB,UAAA,MAAM,QAAA,GAAA,CAAY,SAAA,CAAU,CAAC,CAAA,IAAK,KAAK,IAAA,CAAK,YAAA;AAG5C,UAAA,OAAA,IAAW,MAAM,IAAA,CAAK,GAAA,CAAI,CAAA,GAAI,IAAA,CAAK,KAAK,QAAQ,CAAA;AAChD,UAAA,OAAA,IAAA,CAAY,CAAA,GAAIA,KAAAA,KAAS,CAAA,IAAK,CAAA,GAAI,QAAA,CAAA;AAAA,QACpC;AAEA,QAAA,QAAA,CAAS,KAAK,OAAO,CAAA;AAAA,MACvB;AAGA,MAAA,MAAM,UAAA,GAAa,IAAA,CAAK,GAAA,CAAI,GAAG,QAAQ,CAAA;AACvC,MAAA,MAAM,QAAA,GAAW,SAAS,GAAA,CAAI,CAAC,OAAO,IAAA,CAAK,GAAA,CAAI,EAAA,GAAK,UAAU,CAAC,CAAA;AAC/D,MAAA,MAAM,WAAA,GAAc,SAAS,MAAA,CAAO,CAAC,GAAG,CAAA,KAAM,CAAA,GAAI,GAAG,CAAC,CAAA;AACtD,MAAA,MAAM,QAAQ,QAAA,CAAS,GAAA,CAAI,CAAC,EAAA,KAAO,KAAK,WAAW,CAAA;AAEnD,MAAA,aAAA,CAAc,KAAK,KAAK,CAAA;AAAA,IAC1B;AAEA,IAAA,OAAO,OAAO,aAAa,CAAA;AAAA,EAC7B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC5B,IAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACzB,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,KAAA,CAAM,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OAChF;AAAA,IACF;AACA,IAAA,IAAI,OAAA,GAAU,CAAA;AAEd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,IAAI,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAC,CAAC,CAAA,KAAM,MAAA,CAAO,MAAM,IAAA,CAAK,KAAA,CAAM,MAAA,GAAS,CAAC,CAAC,CAAA,EAAG;AACzE,QAAA,OAAA,EAAA;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,UAAU,CAAA,CAAE,IAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,IAAI,OAAA,GAA8B;AAChC,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,QAAA,EAAU;AAClC,MAAA,OAAO,MAAA;AAAA,IACT;AACA,IAAA,OAAO,OAAO,IAAA,CAAK,QAAA,EAAU,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EACjD;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,cAAc,IAAA,CAAK;AAAA,KACrB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,oBAAoB,0DAA0D,CAAA;AAAA,EAC1F;AACF;;;AC7TA,IAAe,iBAAf,MAA8B;AAAA,EACT,UAAA;AAAA,EACA,OAAA;AAAA,EACA,MAAA;AAAA,EAET,OAAA;AAAA,EACA,OAAA;AAAA,EACA,YAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA,EAEnB,WAAA,CACE,OAAA,GAII,EAAC,EACL;AACA,IAAA,IAAA,CAAK,UAAA,GAAa,QAAQ,UAAA,IAAc,CAAA;AACxC,IAAA,IAAA,CAAK,OAAA,GAAU,QAAQ,OAAA,IAAW,SAAA;AAClC,IAAA,IAAA,CAAK,MAAA,GAAS,QAAQ,MAAA,IAAU,WAAA;AAEhC,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,UAAU,CAAA,IAAK,IAAA,CAAK,aAAa,CAAA,EAAG;AAC7D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,oCAAA;AAAA,QACA,YAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,IAAA,CAAK,OAAA,KAAY,SAAA,IAAa,IAAA,CAAK,YAAY,UAAA,EAAY;AAC7D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,kDAAA,EAAqD,MAAA,CAAO,IAAA,CAAK,OAAO,CAAC,CAAA,CAAA;AAAA,QACzE,SAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,IAAA,CAAK,MAAA,KAAW,WAAA,IAAe,IAAA,CAAK,WAAW,WAAA,EAAa;AAC9D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,MAAA,CAAO,IAAA,CAAK,MAAM,CAAC,CAAA,CAAA;AAAA,QAC1E,QAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAAA,EACF;AAAA,EAEU,iBAAA,CAAkB,IAAc,EAAA,EAAsB;AAC9D,IAAA,IAAI,IAAA,GAAO,CAAA;AACX,IAAA,IAAI,IAAA,CAAK,WAAW,WAAA,EAAa;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,EAAA,CAAG,QAAQ,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,QAAQ,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AACtC,QAAA,IAAA,IAAQ,IAAA,GAAO,IAAA;AAAA,MACjB;AACA,MAAA,OAAO,IAAA,CAAK,KAAK,IAAI,CAAA;AAAA,IACvB,CAAA,MAAO;AAEL,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,EAAA,CAAG,QAAQ,CAAA,EAAA,EAAK;AAClC,QAAA,IAAA,IAAQ,IAAA,CAAK,KAAK,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAE,CAAA;AAAA,MAC9C;AACA,MAAA,OAAO,IAAA;AAAA,IACT;AAAA,EACF;AAAA,EAEU,aAAa,MAAA,EAA8D;AACnF,IAAA,IAAI,CAAC,KAAK,OAAA,EAAS;AACjB,MAAA,MAAM,IAAI,eAAe,+CAA+C,CAAA;AAAA,IAC1E;AAEA,IAAA,MAAM,QAAA,GAAW,IAAA,CAAK,OAAA,CAAQ,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC1C,IAAA,MAAM,SAAA,GAAY,IAAA,CAAK,OAAA,CAAQ,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAE3C,IAAA,MAAM,YAAwD,EAAC;AAE/D,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,cAAwB,EAAC;AAC/B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,WAAA,CAAY,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,OAAA,CAAQ,IAAA,CAAK,IAAA,CAAK,OAAA,CAAQ,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACrF;AACA,MAAA,MAAM,IAAA,GAAO,IAAA,CAAK,iBAAA,CAAkB,MAAA,EAAQ,WAAW,CAAA;AACvD,MAAA,SAAA,CAAU,KAAK,EAAE,KAAA,EAAO,CAAA,EAAG,QAAA,EAAU,MAAM,CAAA;AAAA,IAC7C;AAEA,IAAA,SAAA,CAAU,KAAK,CAAC,CAAA,EAAG,MAAM,CAAA,CAAE,QAAA,GAAW,EAAE,QAAQ,CAAA;AAChD,IAAA,OAAO,SAAA,CAAU,KAAA,CAAM,CAAA,EAAG,IAAA,CAAK,UAAU,CAAA;AAAA,EAC3C;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,YAAY,IAAA,CAAK,UAAA;AAAA,MACjB,SAAS,IAAA,CAAK,OAAA;AAAA,MACd,QAAQ,IAAA,CAAK;AAAA,KACf;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,oBAAoB,0DAA0D,CAAA;AAAA,EAC1F;AACF,CAAA;AAkCO,IAAM,oBAAA,GAAN,cAAmC,cAAA,CAAqC;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAY7E,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AACtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAI,IAAA,CAAK,aAAa,QAAA,EAAU;AAC9B,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,0CAAA,EAA6C,IAAA,CAAK,UAAU,CAAA,GAAA,EAAM,QAAQ,CAAA,CAAA;AAAA,QAC1E,YAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,OAAA,GAAU,CAAA;AACf,IAAA,IAAA,CAAK,OAAA,GAAU,CAAA;AACf,IAAA,IAAA,CAAK,YAAA,GAAe,SAAA;AACpB,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,IAAU,CAAC,KAAK,OAAA,IAAW,CAAC,KAAK,OAAA,EAAS;AAClD,MAAA,MAAM,IAAI,eAAe,uDAAuD,CAAA;AAAA,IAClF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,sBAAsB,CAAA;AAEvE,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,cAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MAC1D;AAEA,MAAA,MAAM,SAAA,GAAY,IAAA,CAAK,YAAA,CAAa,MAAM,CAAA;AAG1C,MAAA,MAAM,KAAA,uBAAY,GAAA,EAAoB;AAEtC,MAAA,KAAA,MAAW,YAAY,SAAA,EAAW;AAChC,QAAA,MAAM,KAAA,GAAQ,MAAA,CAAO,IAAA,CAAK,OAAA,CAAQ,IAAA,CAAK,KAAK,OAAA,CAAQ,MAAA,GAAS,QAAA,CAAS,KAAK,CAAC,CAAA;AAC5E,QAAA,MAAM,SAAS,IAAA,CAAK,OAAA,KAAY,YAAY,CAAA,GAAI,CAAA,IAAK,SAAS,QAAA,GAAW,KAAA,CAAA;AACzE,QAAA,KAAA,CAAM,IAAI,KAAA,EAAA,CAAQ,KAAA,CAAM,IAAI,KAAK,CAAA,IAAK,KAAK,MAAM,CAAA;AAAA,MACnD;AAGA,MAAA,IAAI,QAAA,GAAW,EAAA;AACf,MAAA,IAAI,cAAA,GAAiB,CAAA;AACrB,MAAA,KAAA,MAAW,CAAC,KAAA,EAAO,SAAS,CAAA,IAAK,KAAA,CAAM,SAAQ,EAAG;AAChD,QAAA,IAAI,YAAY,QAAA,EAAU;AACxB,UAAA,QAAA,GAAW,SAAA;AACX,UAAA,cAAA,GAAiB,KAAA;AAAA,QACnB;AAAA,MACF;AAEA,MAAA,WAAA,CAAY,KAAK,cAAc,CAAA;AAAA,IACjC;AAEA,IAAA,OAAO,MAAA,CAAO,WAAA,EAAa,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EAC/C;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,aAAa,CAAA,EAAmB;AAC9B,IAAA,IAAI,CAAC,KAAK,MAAA,IAAU,CAAC,KAAK,OAAA,IAAW,CAAC,KAAK,OAAA,EAAS;AAClD,MAAA,MAAM,IAAI,eAAe,uDAAuD,CAAA;AAAA,IAClF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,sBAAsB,CAAA;AAEvE,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAGhC,IAAA,MAAM,QAAA,uBAAe,GAAA,EAAY;AACjC,IAAA,KAAA,IAAS,IAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,OAAA,CAAQ,MAAM,CAAA,EAAA,EAAK;AAC1C,MAAA,QAAA,CAAS,GAAA,CAAI,MAAA,CAAO,IAAA,CAAK,OAAA,CAAQ,IAAA,CAAK,KAAK,OAAA,CAAQ,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACjE;AACA,IAAA,MAAM,OAAA,GAAU,KAAA,CAAM,IAAA,CAAK,QAAQ,CAAA,CAAE,KAAK,CAAC,CAAA,EAAG,CAAA,KAAM,CAAA,GAAI,CAAC,CAAA;AAEzD,IAAA,MAAM,gBAA4B,EAAC;AAEnC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MAC1D;AAEA,MAAA,MAAM,SAAA,GAAY,IAAA,CAAK,YAAA,CAAa,MAAM,CAAA;AAG1C,MAAA,MAAM,KAAA,uBAAY,GAAA,EAAoB;AACtC,MAAA,IAAI,WAAA,GAAc,CAAA;AAElB,MAAA,KAAA,MAAW,YAAY,SAAA,EAAW;AAChC,QAAA,MAAM,KAAA,GAAQ,MAAA,CAAO,IAAA,CAAK,OAAA,CAAQ,IAAA,CAAK,KAAK,OAAA,CAAQ,MAAA,GAAS,QAAA,CAAS,KAAK,CAAC,CAAA;AAC5E,QAAA,MAAM,SAAS,IAAA,CAAK,OAAA,KAAY,YAAY,CAAA,GAAI,CAAA,IAAK,SAAS,QAAA,GAAW,KAAA,CAAA;AACzE,QAAA,KAAA,CAAM,IAAI,KAAA,EAAA,CAAQ,KAAA,CAAM,IAAI,KAAK,CAAA,IAAK,KAAK,MAAM,CAAA;AACjD,QAAA,WAAA,IAAe,MAAA;AAAA,MACjB;AAGA,MAAA,MAAM,QAAkB,EAAC;AACzB,MAAA,KAAA,MAAW,OAAO,OAAA,EAAS;AACzB,QAAA,KAAA,CAAM,MAAM,KAAA,CAAM,GAAA,CAAI,GAAG,CAAA,IAAK,KAAK,WAAW,CAAA;AAAA,MAChD;AACA,MAAA,aAAA,CAAc,KAAK,KAAK,CAAA;AAAA,IAC1B;AAEA,IAAA,OAAO,OAAO,aAAa,CAAA;AAAA,EAC7B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC5B,IAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACzB,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,KAAA,CAAM,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OAChF;AAAA,IACF;AACA,IAAA,IAAI,OAAA,GAAU,CAAA;AAEd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,IAAI,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAC,CAAC,CAAA,KAAM,MAAA,CAAO,MAAM,IAAA,CAAK,KAAA,CAAM,MAAA,GAAS,CAAC,CAAC,CAAA,EAAG;AACzE,QAAA,OAAA,EAAA;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,UAAU,CAAA,CAAE,IAAA;AAAA,EACrB;AACF;AAwBO,IAAM,mBAAA,GAAN,cAAkC,cAAA,CAAoC;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAY3E,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AACtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAI,IAAA,CAAK,aAAa,QAAA,EAAU;AAC9B,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,0CAAA,EAA6C,IAAA,CAAK,UAAU,CAAA,GAAA,EAAM,QAAQ,CAAA,CAAA;AAAA,QAC1E,YAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,OAAA,GAAU,CAAA;AACf,IAAA,IAAA,CAAK,OAAA,GAAU,CAAA;AACf,IAAA,IAAA,CAAK,YAAA,GAAe,SAAA;AACpB,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AAEd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,IAAU,CAAC,KAAK,OAAA,IAAW,CAAC,KAAK,OAAA,EAAS;AAClD,MAAA,MAAM,IAAI,eAAe,sDAAsD,CAAA;AAAA,IACjF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,YAAA,IAAgB,CAAA,EAAG,qBAAqB,CAAA;AAEtE,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,MAAM,cAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,SAAmB,EAAC;AAC1B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAA,CAAO,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MAC1D;AAEA,MAAA,MAAM,SAAA,GAAY,IAAA,CAAK,YAAA,CAAa,MAAM,CAAA;AAG1C,MAAA,IAAI,SAAA,GAAY,CAAA;AAChB,MAAA,IAAI,UAAA,GAAa,CAAA;AAEjB,MAAA,KAAA,MAAW,YAAY,SAAA,EAAW;AAChC,QAAA,MAAM,KAAA,GAAQ,MAAA,CAAO,IAAA,CAAK,OAAA,CAAQ,IAAA,CAAK,KAAK,OAAA,CAAQ,MAAA,GAAS,QAAA,CAAS,KAAK,CAAC,CAAA;AAC5E,QAAA,MAAM,SAAS,IAAA,CAAK,OAAA,KAAY,YAAY,CAAA,GAAI,CAAA,IAAK,SAAS,QAAA,GAAW,KAAA,CAAA;AACzE,QAAA,SAAA,IAAa,KAAA,GAAQ,MAAA;AACrB,QAAA,UAAA,IAAc,MAAA;AAAA,MAChB;AAEA,MAAA,WAAA,CAAY,IAAA,CAAK,YAAY,UAAU,CAAA;AAAA,IACzC;AAEA,IAAA,OAAO,OAAO,WAAW,CAAA;AAAA,EAC3B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC5B,IAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACzB,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,KAAA,CAAM,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OAChF;AAAA,IACF;AAEA,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,IACtC;AACA,IAAA,KAAA,IAAS,CAAA,CAAE,IAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,QAAQ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AACzC,MAAA,MAAM,WAAW,MAAA,CAAO,KAAA,CAAM,KAAK,KAAA,CAAM,MAAA,GAAS,CAAC,CAAC,CAAA;AACpD,MAAA,KAAA,IAAA,CAAU,QAAQ,QAAA,KAAa,CAAA;AAC/B,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAAA,IAC9B;AAEA,IAAA,IAAI,UAAU,CAAA,EAAG;AACf,MAAA,OAAO,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA;AAAA,IAC7B;AAEA,IAAA,OAAO,IAAI,KAAA,GAAQ,KAAA;AAAA,EACrB;AACF;;;AC/bO,IAAM,YAAN,MAAsC;AAAA;AAAA,EAE1B,CAAA;AAAA;AAAA,EAGA,OAAA;AAAA;AAAA,EAGA,GAAA;AAAA;AAAA,EAGT,UAAoB,EAAC;AAAA;AAAA,EAGrB,IAAA,GAAO,CAAA;AAAA;AAAA,EAGP,SAAA,GAAY,CAAA;AAAA;AAAA,EAGZ,cAAwB,EAAC;AAAA;AAAA,EAGzB,MAAA,GAAS,KAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAUjB,WAAA,CACE,OAAA,GAII,EAAC,EACL;AACA,IAAA,IAAA,CAAK,CAAA,GAAI,QAAQ,CAAA,IAAK,CAAA;AACtB,IAAA,IAAA,CAAK,OAAA,GAAU,QAAQ,OAAA,IAAW,GAAA;AAClC,IAAA,IAAA,CAAK,GAAA,GAAM,QAAQ,GAAA,IAAO,IAAA;AAG1B,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,CAAC,CAAA,IAAK,IAAA,CAAK,KAAK,CAAA,EAAG;AAC3C,MAAA,MAAM,IAAI,qBAAA,CAAsB,oBAAA,EAAsB,GAAA,EAAK,KAAK,CAAC,CAAA;AAAA,IACnE;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,OAAO,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACxD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,oCAAA;AAAA,QACA,SAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,GAAG,CAAA,IAAK,IAAA,CAAK,MAAM,CAAA,EAAG;AAC9C,MAAA,MAAM,IAAI,qBAAA,CAAsB,kBAAA,EAAoB,KAAA,EAAO,KAAK,GAAG,CAAA;AAAA,IACrE;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAgBA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAEtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAGjB,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,MAAM,QAAkB,EAAC;AAEzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AACd,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAGA,IAAA,IAAA,CAAK,WAAA,GAAc,CAAC,GAAG,IAAI,GAAA,CAAI,KAAK,CAAC,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,KAAM,IAAI,CAAC,CAAA;AAC3D,IAAA,IAAI,IAAA,CAAK,WAAA,CAAY,MAAA,KAAW,CAAA,EAAG;AACjC,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,gEAAA;AAAA,QACA,GAAA;AAAA,QACA,KAAK,WAAA,CAAY;AAAA,OACnB;AAAA,IACF;AAGA,IAAA,MAAM,OAAA,GAAU,KAAA,CAAM,GAAA,CAAI,CAAC,KAAA,KAAW,KAAA,KAAU,IAAA,CAAK,WAAA,CAAY,CAAC,CAAA,GAAI,EAAA,GAAK,CAAE,CAAA;AAG7E,IAAA,IAAA,CAAK,UAAU,IAAI,KAAA,CAAM,SAAS,CAAA,CAAE,KAAK,CAAC,CAAA;AAC1C,IAAA,IAAA,CAAK,IAAA,GAAO,CAAA;AAWZ,IAAA,MAAM,YAAA,GAAe,IAAA;AAErB,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,IAAA,CAAK,SAAS,IAAA,EAAA,EAAQ;AAC9C,MAAA,IAAI,YAAA,GAAe,CAAA;AAEnB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,EAAA,GAAK,MAAM,CAAC,CAAA;AAClB,QAAA,MAAM,EAAA,GAAK,QAAQ,CAAC,CAAA;AAEpB,QAAA,IAAI,EAAA,KAAO,MAAA,IAAa,EAAA,KAAO,MAAA,EAAW;AAG1C,QAAA,IAAI,WAAW,IAAA,CAAK,IAAA;AACpB,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,QAAA,IAAA,CAAa,KAAK,OAAA,CAAQ,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,QACjD;AAGA,QAAA,MAAM,SAAS,EAAA,GAAK,QAAA;AAGpB,QAAA,IAAI,SAAS,CAAA,EAAG;AACd,UAAA,YAAA,GAAe,IAAA,CAAK,GAAA,CAAI,YAAA,EAAc,CAAA,GAAI,MAAM,CAAA;AAAA,QAClD;AAYA,QAAA,MAAM,cAAc,IAAA,CAAK,GAAA,CAAI,cAAc,CAAA,IAAO,IAAA,CAAK,IAAI,EAAA,CAAG,CAAA;AAE9D,QAAA,IAAI,SAAS,CAAA,EAAG;AAEd,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,IAAA,CAAK,QAAQ,CAAC,CAAA,GAAA,CACX,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA,KAAM,CAAA,GAAI,WAAA,CAAA,GAAe,cAAc,IAAA,CAAK,CAAA,GAAI,EAAA,IAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,UACvF;AACA,UAAA,IAAA,CAAK,IAAA,IAAQ,WAAA,GAAc,IAAA,CAAK,CAAA,GAAI,EAAA;AAAA,QACtC,CAAA,MAAO;AAEL,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,GAAA,CAAK,IAAA,CAAK,QAAQ,CAAC,CAAA,IAAK,MAAM,CAAA,GAAI,WAAA,CAAA;AAAA,UAClD;AAAA,QACF;AAAA,MACF;AAGA,MAAA,IAAI,YAAA,GAAe,KAAK,GAAA,EAAK;AAC3B,QAAA;AAAA,MACF;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,sCAAsC,CAAA;AAAA,IACjE;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,WAAW,CAAA;AAEzD,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,MAAM,cAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AAEjC,MAAA,IAAI,WAAW,IAAA,CAAK,IAAA;AACpB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,QAAA,IAAA,CAAa,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AAAA,MAClF;AAGA,MAAA,MAAM,cAAA,GAAiB,YAAY,CAAA,GAAI,IAAA,CAAK,YAAY,CAAC,CAAA,GAAI,IAAA,CAAK,WAAA,CAAY,CAAC,CAAA;AAC/E,MAAA,WAAA,CAAY,IAAA,CAAK,kBAAkB,CAAC,CAAA;AAAA,IACtC;AAEA,IAAA,OAAO,MAAA,CAAO,WAAA,EAAa,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EAC/C;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,aAAa,CAAA,EAAmB;AAC9B,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,4CAA4C,CAAA;AAAA,IACvE;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,WAAW,CAAA;AAEzD,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,MAAM,QAAoB,EAAC;AAE3B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AAEjC,MAAA,IAAI,WAAW,IAAA,CAAK,IAAA;AACpB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,QAAA,IAAA,CAAa,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AAAA,MAClF;AAGA,MAAA,MAAM,KAAK,CAAA,IAAK,CAAA,GAAI,IAAA,CAAK,GAAA,CAAI,CAAC,QAAQ,CAAA,CAAA;AACtC,MAAA,KAAA,CAAM,IAAA,CAAK,CAAC,CAAA,GAAI,EAAA,EAAI,EAAE,CAAC,CAAA;AAAA,IACzB;AAEA,IAAA,OAAO,OAAO,KAAK,CAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AACA,IAAA,IAAI,OAAA,GAAU,CAAA;AACd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,IAAI,MAAA,CAAO,WAAA,CAAY,IAAA,CAAK,WAAA,CAAY,SAAS,CAAC,CAAC,CAAA,KAAM,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA,EAAG;AACrF,QAAA,OAAA,EAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,OAAO,UAAU,CAAA,CAAE,IAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,IAAA,GAAe;AACjB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,iDAAiD,CAAA;AAAA,IAC5E;AACA,IAAA,OAAO,MAAA,CAAO,CAAC,IAAA,CAAK,OAAO,CAAC,CAAA;AAAA,EAC9B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,IAAI,SAAA,GAAoB;AACtB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,8CAA8C,CAAA;AAAA,IACzE;AACA,IAAA,OAAO,IAAA,CAAK,IAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,GAAG,IAAA,CAAK,CAAA;AAAA,MACR,SAAS,IAAA,CAAK,OAAA;AAAA,MACd,KAAK,IAAA,CAAK;AAAA,KACZ;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,oBAAoB,yDAAyD,CAAA;AAAA,EACzF;AACF;AAsBO,IAAM,YAAN,MAAqC;AAAA;AAAA,EAEzB,CAAA;AAAA;AAAA,EAGA,OAAA;AAAA;AAAA,EAGA,OAAA;AAAA;AAAA,EAGA,GAAA;AAAA;AAAA,EAGT,UAAoB,EAAC;AAAA;AAAA,EAGrB,IAAA,GAAO,CAAA;AAAA;AAAA,EAGP,SAAA,GAAY,CAAA;AAAA;AAAA,EAGZ,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAKI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,CAAA,GAAI,QAAQ,CAAA,IAAK,CAAA;AACtB,IAAA,IAAA,CAAK,OAAA,GAAU,QAAQ,OAAA,IAAW,GAAA;AAClC,IAAA,IAAA,CAAK,OAAA,GAAU,QAAQ,OAAA,IAAW,GAAA;AAClC,IAAA,IAAA,CAAK,GAAA,GAAM,QAAQ,GAAA,IAAO,IAAA;AAE1B,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,CAAC,CAAA,IAAK,IAAA,CAAK,KAAK,CAAA,EAAG;AAC3C,MAAA,MAAM,IAAI,qBAAA,CAAsB,oBAAA,EAAsB,GAAA,EAAK,KAAK,CAAC,CAAA;AAAA,IACnE;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,OAAO,CAAA,IAAK,IAAA,CAAK,UAAU,CAAA,EAAG;AACtD,MAAA,MAAM,IAAI,qBAAA,CAAsB,sBAAA,EAAwB,SAAA,EAAW,KAAK,OAAO,CAAA;AAAA,IACjF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,OAAO,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACxD,MAAA,MAAM,IAAI,qBAAA,CAAsB,0BAAA,EAA4B,SAAA,EAAW,KAAK,OAAO,CAAA;AAAA,IACrF;AACA,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,IAAA,CAAK,GAAG,CAAA,IAAK,IAAA,CAAK,MAAM,CAAA,EAAG;AAC9C,MAAA,MAAM,IAAI,qBAAA,CAAsB,kBAAA,EAAoB,KAAA,EAAO,KAAK,GAAG,CAAA;AAAA,IACrE;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAEtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAGjB,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,MAAM,QAAkB,EAAC;AAEzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AACd,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAGA,IAAA,IAAA,CAAK,UAAU,IAAI,KAAA,CAAM,SAAS,CAAA,CAAE,KAAK,CAAC,CAAA;AAC1C,IAAA,IAAA,CAAK,IAAA,GAAO,CAAA;AAEZ,IAAA,MAAM,YAAA,GAAe,IAAA;AAErB,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,IAAA,CAAK,SAAS,IAAA,EAAA,EAAQ;AAC9C,MAAA,IAAI,SAAA,GAAY,CAAA;AAEhB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,MAAM,EAAA,GAAK,MAAM,CAAC,CAAA;AAClB,QAAA,MAAM,EAAA,GAAK,MAAM,CAAC,CAAA;AAElB,QAAA,IAAI,EAAA,KAAO,MAAA,IAAa,EAAA,KAAO,MAAA,EAAW;AAG1C,QAAA,IAAI,OAAO,IAAA,CAAK,IAAA;AAChB,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,UAAA,IAAA,IAAA,CAAS,KAAK,OAAA,CAAQ,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,QAC7C;AAEA,QAAA,MAAM,QAAQ,IAAA,GAAO,EAAA;AACrB,QAAA,MAAM,QAAA,GAAW,IAAA,CAAK,GAAA,CAAI,KAAK,CAAA;AAG/B,QAAA,IAAI,QAAA,GAAW,KAAK,OAAA,EAAS;AAC3B,UAAA,SAAA,IAAa,WAAW,IAAA,CAAK,OAAA;AAG7B,UAAA,MAAM,IAAA,GAAO,KAAA,GAAQ,CAAA,GAAI,CAAA,GAAI,EAAA;AAE7B,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,IAAA,CAAK,QAAQ,CAAC,CAAA,GAAA,CACX,KAAK,OAAA,CAAQ,CAAC,KAAK,CAAA,IACpB,YAAA,IAAgB,KAAK,CAAA,GAAI,IAAA,IAAQ,GAAG,CAAC,CAAA,IAAK,MAAM,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA,CAAA,CAAA;AAAA,UACvE;AACA,UAAA,IAAA,CAAK,IAAA,IAAQ,YAAA,GAAe,IAAA,CAAK,CAAA,GAAI,IAAA;AAAA,QACvC,CAAA,MAAO;AAEL,UAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,YAAA,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,GAAA,CAAK,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA,IAAK,YAAA,IAAgB,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA,CAAA;AAAA,UAChF;AAAA,QACF;AAAA,MACF;AAEA,MAAA,IAAI,SAAA,GAAY,QAAA,GAAW,IAAA,CAAK,GAAA,EAAK;AACnC,QAAA;AAAA,MACF;AAAA,IACF;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAWA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,sCAAsC,CAAA;AAAA,IACjE;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,KAAK,CAAA;AAEnD,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAEhC,IAAA,MAAM,cAAwB,EAAC;AAE/B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,IAAI,OAAO,IAAA,CAAK,IAAA;AAChB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,IAAA,IAAA,CAAS,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA,IAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAA;AAAA,MAC9E;AACA,MAAA,WAAA,CAAY,KAAK,IAAI,CAAA;AAAA,IACvB;AAEA,IAAA,OAAO,OAAO,WAAW,CAAA;AAAA,EAC3B;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AAEA,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,IACtC;AACA,IAAA,KAAA,IAAS,CAAA,CAAE,IAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,QAAQ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AACzC,MAAA,MAAM,QAAQ,MAAA,CAAO,WAAA,CAAY,KAAK,WAAA,CAAY,MAAA,GAAS,CAAC,CAAC,CAAA;AAC7D,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAC5B,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAAA,IAC9B;AAEA,IAAA,OAAO,UAAU,CAAA,GAAK,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA,GAAO,IAAI,KAAA,GAAQ,KAAA;AAAA,EAC/D;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,GAAG,IAAA,CAAK,CAAA;AAAA,MACR,SAAS,IAAA,CAAK,OAAA;AAAA,MACd,SAAS,IAAA,CAAK,OAAA;AAAA,MACd,KAAK,IAAA,CAAK;AAAA,KACZ;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,oBAAoB,yDAAyD,CAAA;AAAA,EACzF;AACF;;;AC3lBO,IAAM,yBAAN,MAAmD;AAAA,EACvC,WAAA;AAAA,EACA,QAAA;AAAA,EACA,eAAA;AAAA,EACA,cAAA;AAAA,EACA,WAAA;AAAA,EACA,SAAA;AAAA,EACA,WAAA;AAAA,EAET,QAAkC,EAAC;AAAA,EACnC,WAAA;AAAA,EACA,SAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAQI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,GAAA;AAC1C,IAAA,IAAA,CAAK,QAAA,GAAW,QAAQ,QAAA,IAAY,EAAA;AACpC,IAAA,IAAA,CAAK,eAAA,GAAkB,QAAQ,eAAA,IAAmB,CAAA;AAClD,IAAA,IAAA,CAAK,cAAA,GAAiB,QAAQ,cAAA,IAAkB,CAAA;AAChD,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,MAAA;AAC1C,IAAA,IAAA,CAAK,SAAA,GAAY,QAAQ,SAAA,IAAa,IAAA;AACtC,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AAEA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,cAAc,CAAA,EAAG;AAC/D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,KAAK,WAAW,CAAA,CAAA;AAAA,QACjE,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,QAAQ,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACzD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,2CAAA,EAA8C,KAAK,QAAQ,CAAA,CAAA;AAAA,QAC3D,UAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,kBAAkB,CAAA,EAAG;AACvE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,kDAAA,EAAqD,KAAK,eAAe,CAAA,CAAA;AAAA,QACzE,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,cAAc,CAAA,IAAK,IAAA,CAAK,iBAAiB,CAAA,EAAG;AACrE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,iDAAA,EAAoD,KAAK,cAAc,CAAA,CAAA;AAAA,QACvE,gBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,OAAO,IAAA,CAAK,WAAA,KAAgB,QAAA,EAAU;AACxC,MAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,cAAc,CAAA,EAAG;AAC/D,QAAA,MAAM,IAAI,qBAAA;AAAA,UACR,CAAA,8CAAA,EAAiD,KAAK,WAAW,CAAA,CAAA;AAAA,UACjE,aAAA;AAAA,UACA,IAAA,CAAK;AAAA,SACP;AAAA,MACF;AAAA,IACF,WAAW,IAAA,CAAK,WAAA,KAAgB,MAAA,IAAU,IAAA,CAAK,gBAAgB,MAAA,EAAQ;AACrE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,oEAAA,EAAuE,MAAA,CAAO,IAAA,CAAK,WAAW,CAAC,CAAA,CAAA;AAAA,QAC/F,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,IAAa,CAAC,OAAO,QAAA,CAAS,OAAA,CAAQ,WAAW,CAAA,EAAG;AAC9E,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,OAAA,CAAQ,WAAW,CAAC,CAAA,CAAA;AAAA,QAC5E,aAAA;AAAA,QACA,OAAA,CAAQ;AAAA,OACV;AAAA,IACF;AAAA,EACF;AAAA,EAEQ,SAAA,GAA0B;AAChC,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,MAAA,IAAI,OAAO,IAAA,CAAK,WAAA;AAChB,MAAA,OAAO,MAAM;AACX,QAAA,IAAA,GAAA,CAAQ,IAAA,GAAO,OAAO,KAAA,IAAS,MAAA;AAC/B,QAAA,OAAO,IAAA,GAAO,MAAA;AAAA,MAChB,CAAA;AAAA,IACF;AACA,IAAA,OAAO,IAAA,CAAK,MAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAeA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AACtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAGjB,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,MAAM,QAAkB,EAAC;AACzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AACd,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAEA,IAAA,IAAA,CAAK,WAAA,GAAc,CAAC,GAAG,IAAI,GAAA,CAAI,KAAK,CAAC,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,KAAM,IAAI,CAAC,CAAA;AAG3D,IAAA,IAAI,eAAA;AACJ,IAAA,IAAI,OAAO,IAAA,CAAK,WAAA,KAAgB,QAAA,EAAU;AACxC,MAAA,eAAA,GAAkB,IAAA,CAAK,GAAA,CAAI,IAAA,CAAK,WAAA,EAAa,SAAS,CAAA;AAAA,IACxD,CAAA,MAAA,IAAW,IAAA,CAAK,WAAA,KAAgB,MAAA,EAAQ;AACtC,MAAA,eAAA,GAAkB,IAAA,CAAK,KAAA,CAAM,IAAA,CAAK,IAAA,CAAK,SAAS,CAAC,CAAA;AAAA,IACnD,CAAA,MAAO;AACL,MAAA,eAAA,GAAkB,IAAA,CAAK,KAAA,CAAM,IAAA,CAAK,IAAA,CAAK,SAAS,CAAC,CAAA;AAAA,IACnD;AACA,IAAA,eAAA,GAAkB,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,eAAe,CAAA;AAE7C,IAAA,MAAM,GAAA,GAAM,KAAK,SAAA,EAAU;AAE3B,IAAA,IAAA,CAAK,QAAQ,EAAC;AAEd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AAEzC,MAAA,MAAM,gBAA0B,EAAC;AACjC,MAAA,IAAI,KAAK,SAAA,EAAW;AAClB,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,aAAA,CAAc,KAAK,IAAA,CAAK,KAAA,CAAM,GAAA,EAAI,GAAI,QAAQ,CAAC,CAAA;AAAA,QACjD;AAAA,MACF,CAAA,MAAO;AACL,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,aAAA,CAAc,KAAK,CAAC,CAAA;AAAA,QACtB;AAAA,MACF;AAGA,MAAA,MAAM,UAAsB,EAAC;AAC7B,MAAA,MAAM,UAAoB,EAAC;AAC3B,MAAA,KAAA,MAAW,aAAa,aAAA,EAAe;AAErC,QAAA,OAAA,CAAQ,IAAA,CAAK,KAAA,CAAM,SAAS,CAAA,IAAK,EAAE,CAAA;AACnC,QAAA,OAAA,CAAQ,IAAA,CAAK,KAAA,CAAM,SAAS,CAAA,IAAK,CAAC,CAAA;AAAA,MACpC;AAGA,MAAA,MAAM,WAAA,GAMF;AAAA,QACF,UAAU,IAAA,CAAK,QAAA;AAAA,QACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,QACtB,gBAAgB,IAAA,CAAK,cAAA;AAAA,QACrB,WAAA,EAAa;AAAA,OACf;AACA,MAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,QAAA,WAAA,CAAY,WAAA,GAAc,KAAK,WAAA,GAAc,CAAA;AAAA,MAC/C;AACA,MAAA,MAAM,IAAA,GAAO,IAAI,sBAAA,CAAuB,WAAW,CAAA;AACnD,MAAA,IAAA,CAAK,GAAA,CAAI,MAAA,CAAO,OAAO,CAAA,EAAG,MAAA,CAAO,SAAS,EAAE,KAAA,EAAO,OAAA,EAAS,CAAC,CAAA;AAC7D,MAAA,IAAA,CAAK,KAAA,CAAM,KAAK,IAAI,CAAA;AAAA,IACtB;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,yDAAyD,CAAA;AAAA,IACpF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,wBAAwB,CAAA;AAEtE,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAG/B,IAAA,MAAM,iBAA6B,EAAC;AAEpC,IAAA,KAAA,MAAW,IAAA,IAAQ,KAAK,KAAA,EAAO;AAC7B,MAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC5B,MAAA,MAAM,YAAsB,EAAC;AAC7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,SAAA,CAAU,IAAA,CAAK,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,MACrD;AACA,MAAA,cAAA,CAAe,KAAK,SAAS,CAAA;AAAA,IAC/B;AAGA,IAAA,MAAM,mBAA6B,EAAC;AACpC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,KAAA,uBAAY,GAAA,EAAoB;AACtC,MAAA,KAAA,MAAW,aAAa,cAAA,EAAgB;AACtC,QAAA,MAAM,IAAA,GAAO,SAAA,CAAU,CAAC,CAAA,IAAK,CAAA;AAC7B,QAAA,KAAA,CAAM,IAAI,IAAA,EAAA,CAAO,KAAA,CAAM,IAAI,IAAI,CAAA,IAAK,KAAK,CAAC,CAAA;AAAA,MAC5C;AAEA,MAAA,IAAI,QAAA,GAAW,CAAA;AACf,MAAA,IAAI,UAAA,GAAa,CAAA;AACjB,MAAA,KAAA,MAAW,CAAC,KAAA,EAAO,KAAK,CAAA,IAAK,KAAA,EAAO;AAClC,QAAA,IAAI,QAAQ,QAAA,EAAU;AACpB,UAAA,QAAA,GAAW,KAAA;AACX,UAAA,UAAA,GAAa,KAAA;AAAA,QACf;AAAA,MACF;AACA,MAAA,gBAAA,CAAiB,KAAK,UAAU,CAAA;AAAA,IAClC;AAEA,IAAA,OAAO,MAAA,CAAO,gBAAA,EAAkB,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EACpD;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,aAAa,CAAA,EAAmB;AAC9B,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,yDAAyD,CAAA;AAAA,IACpF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,wBAAwB,CAAA;AACtE,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AAEvB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,WAAA,IAAe,EAAC;AACzC,IAAA,MAAM,WAAW,WAAA,CAAY,MAAA;AAE7B,IAAA,IAAI,aAAa,CAAA,EAAG;AAClB,MAAA,MAAM,IAAI,eAAe,yDAAyD,CAAA;AAAA,IACpF;AAEA,IAAA,MAAM,UAAA,uBAAiB,GAAA,EAAoB;AAC3C,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,CAAA,GAAI,YAAY,CAAC,CAAA;AACvB,MAAA,IAAI,CAAA,KAAM,MAAA,EAAW,UAAA,CAAW,GAAA,CAAI,GAAG,CAAC,CAAA;AAAA,IAC1C;AAEA,IAAA,MAAM,QAAoB,KAAA,CAAM,IAAA;AAAA,MAAK,EAAE,QAAQ,QAAA,EAAS;AAAA,MAAG,MACzD,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,KAAK,CAAC;AAAA,KACpC;AAEA,IAAA,KAAA,MAAW,IAAA,IAAQ,KAAK,KAAA,EAAO;AAC7B,MAAA,MAAM,SAAA,GAAY,IAAA,CAAK,YAAA,CAAa,CAAC,CAAA;AACrC,MAAA,MAAM,cAAc,IAAA,CAAK,OAAA;AACzB,MAAA,IAAI,CAAC,WAAA,EAAa;AAClB,MAAA,gBAAA,CAAiB,aAAa,SAAS,CAAA;AAEvC,MAAA,MAAM,IAAI,WAAA,CAAY,IAAA;AACtB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,QAAA,MAAM,GAAA,GAAM,OAAO,WAAA,CAAY,IAAA,CAAK,YAAY,MAAA,GAAS,CAAC,KAAK,CAAC,CAAA;AAChE,QAAA,MAAM,OAAA,GAAU,UAAA,CAAW,GAAA,CAAI,GAAG,CAAA;AAClC,QAAA,IAAI,YAAY,MAAA,EAAW;AAE3B,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,MAAM,GAAA,GAAM,MAAM,CAAC,CAAA;AACnB,UAAA,IAAI,GAAA,EAAK;AACP,YAAA,GAAA,CAAI,OAAO,CAAA,GAAA,CACR,GAAA,CAAI,OAAO,KAAK,CAAA,IAAK,MAAA,CAAO,SAAA,CAAU,IAAA,CAAK,UAAU,MAAA,GAAS,CAAA,GAAI,CAAA,GAAI,CAAC,KAAK,CAAC,CAAA;AAAA,UAClF;AAAA,QACF;AAAA,MACF;AAAA,IACF;AAEA,IAAA,MAAM,QAAA,GAAW,KAAK,KAAA,CAAM,MAAA,KAAW,IAAI,CAAA,GAAI,CAAA,GAAI,KAAK,KAAA,CAAM,MAAA;AAC9D,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,GAAA,GAAM,MAAM,CAAC,CAAA;AACnB,MAAA,IAAI,GAAA,EAAK;AACP,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,GAAA,CAAI,CAAC,CAAA,GAAA,CAAK,GAAA,CAAI,CAAC,KAAK,CAAA,IAAK,QAAA;AAAA,QAC3B;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,OAAO,KAAK,CAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,sDAAsD,CAAA;AAAA,IACjF;AACA,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AACA,IAAA,IAAI,OAAA,GAAU,CAAA;AACd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,IAAI,MAAA,CAAO,WAAA,CAAY,IAAA,CAAK,WAAA,CAAY,SAAS,CAAC,CAAC,CAAA,KAAM,MAAA,CAAO,EAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA,EAAG;AACrF,QAAA,OAAA,EAAA;AAAA,MACF;AAAA,IACF;AACA,IAAA,OAAO,UAAU,CAAA,CAAE,IAAA;AAAA,EACrB;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,IAAI,OAAA,GAA8B;AAChC,IAAA,IAAI,CAAC,IAAA,CAAK,MAAA,IAAU,CAAC,KAAK,WAAA,EAAa;AACrC,MAAA,OAAO,MAAA;AAAA,IACT;AACA,IAAA,OAAO,OAAO,IAAA,CAAK,WAAA,EAAa,EAAE,KAAA,EAAO,SAAS,CAAA;AAAA,EACpD;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,UAAU,IAAA,CAAK,QAAA;AAAA,MACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,MACtB,gBAAgB,IAAA,CAAK,cAAA;AAAA,MACrB,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,WAAW,IAAA,CAAK,SAAA;AAAA,MAChB,aAAa,IAAA,CAAK;AAAA,KACpB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AACF;AAyBO,IAAM,wBAAN,MAAiD;AAAA,EACrC,WAAA;AAAA,EACA,QAAA;AAAA,EACA,eAAA;AAAA,EACA,cAAA;AAAA,EACA,WAAA;AAAA,EACA,SAAA;AAAA,EACA,WAAA;AAAA,EAET,QAAiC,EAAC;AAAA,EAClC,SAAA;AAAA,EACA,MAAA,GAAS,KAAA;AAAA,EAEjB,WAAA,CACE,OAAA,GAQI,EAAC,EACL;AACA,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,GAAA;AAC1C,IAAA,IAAA,CAAK,QAAA,GAAW,QAAQ,QAAA,IAAY,EAAA;AACpC,IAAA,IAAA,CAAK,eAAA,GAAkB,QAAQ,eAAA,IAAmB,CAAA;AAClD,IAAA,IAAA,CAAK,cAAA,GAAiB,QAAQ,cAAA,IAAkB,CAAA;AAChD,IAAA,IAAA,CAAK,WAAA,GAAc,QAAQ,WAAA,IAAe,CAAA;AAC1C,IAAA,IAAA,CAAK,SAAA,GAAY,QAAQ,SAAA,IAAa,IAAA;AACtC,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,EAAW;AACrC,MAAA,IAAA,CAAK,cAAc,OAAA,CAAQ,WAAA;AAAA,IAC7B;AAEA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,cAAc,CAAA,EAAG;AAC/D,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,KAAK,WAAW,CAAA,CAAA;AAAA,QACjE,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,QAAQ,CAAA,IAAK,IAAA,CAAK,WAAW,CAAA,EAAG;AACzD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,2CAAA,EAA8C,KAAK,QAAQ,CAAA,CAAA;AAAA,QAC3D,UAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,eAAe,CAAA,IAAK,IAAA,CAAK,kBAAkB,CAAA,EAAG;AACvE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,kDAAA,EAAqD,KAAK,eAAe,CAAA,CAAA;AAAA,QACzE,iBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,CAAC,OAAO,SAAA,CAAU,IAAA,CAAK,cAAc,CAAA,IAAK,IAAA,CAAK,iBAAiB,CAAA,EAAG;AACrE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,iDAAA,EAAoD,KAAK,cAAc,CAAA,CAAA;AAAA,QACvE,gBAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,OAAO,IAAA,CAAK,WAAA,KAAgB,QAAA,EAAU;AACxC,MAAA,IACE,IAAA,CAAK,WAAA,KAAgB,CAAA,KACpB,CAAC,MAAA,CAAO,SAAA,CAAU,IAAA,CAAK,WAAW,CAAA,IAAK,IAAA,CAAK,WAAA,GAAc,CAAA,CAAA,EAC3D;AACA,QAAA,MAAM,IAAI,qBAAA;AAAA,UACR,CAAA,sEAAA,EAAyE,KAAK,WAAW,CAAA,CAAA;AAAA,UACzF,aAAA;AAAA,UACA,IAAA,CAAK;AAAA,SACP;AAAA,MACF;AAAA,IACF,WAAW,IAAA,CAAK,WAAA,KAAgB,MAAA,IAAU,IAAA,CAAK,gBAAgB,MAAA,EAAQ;AACrE,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,oEAAA,EAAuE,MAAA,CAAO,IAAA,CAAK,WAAW,CAAC,CAAA,CAAA;AAAA,QAC/F,aAAA;AAAA,QACA,IAAA,CAAK;AAAA,OACP;AAAA,IACF;AACA,IAAA,IAAI,OAAA,CAAQ,gBAAgB,MAAA,IAAa,CAAC,OAAO,QAAA,CAAS,OAAA,CAAQ,WAAW,CAAA,EAAG;AAC9E,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,8CAAA,EAAiD,MAAA,CAAO,OAAA,CAAQ,WAAW,CAAC,CAAA,CAAA;AAAA,QAC5E,aAAA;AAAA,QACA,OAAA,CAAQ;AAAA,OACV;AAAA,IACF;AAAA,EACF;AAAA,EAEQ,SAAA,GAA0B;AAChC,IAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,MAAA,IAAI,OAAO,IAAA,CAAK,WAAA;AAChB,MAAA,OAAO,MAAM;AACX,QAAA,IAAA,GAAA,CAAQ,IAAA,GAAO,OAAO,KAAA,IAAS,MAAA;AAC/B,QAAA,OAAO,IAAA,GAAO,MAAA;AAAA,MAChB,CAAA;AAAA,IACF;AACA,IAAA,OAAO,IAAA,CAAK,MAAA;AAAA,EACd;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAeA,GAAA,CAAI,GAAW,CAAA,EAAiB;AAC9B,IAAA,iBAAA,CAAkB,GAAG,CAAC,CAAA;AAEtB,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,IAAA,MAAM,SAAA,GAAY,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,IAAA,CAAK,SAAA,GAAY,SAAA;AAEjB,IAAA,MAAM,QAAoB,EAAC;AAC3B,IAAA,MAAM,QAAkB,EAAC;AACzB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,MAAgB,EAAC;AACvB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,GAAA,CAAI,IAAA,CAAK,MAAA,CAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,SAAS,CAAA,GAAI,SAAA,GAAY,CAAC,CAAC,CAAC,CAAA;AAAA,MACvD;AACA,MAAA,KAAA,CAAM,KAAK,GAAG,CAAA;AACd,MAAA,KAAA,CAAM,IAAA,CAAK,OAAO,CAAA,CAAE,IAAA,CAAK,EAAE,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,IACzC;AAEA,IAAA,IAAI,eAAA;AACJ,IAAA,IAAI,OAAO,IAAA,CAAK,WAAA,KAAgB,QAAA,EAAU;AACxC,MAAA,IAAI,IAAA,CAAK,gBAAgB,CAAA,EAAK;AAC5B,QAAA,eAAA,GAAkB,SAAA;AAAA,MACpB,CAAA,MAAO;AACL,QAAA,eAAA,GAAkB,IAAA,CAAK,GAAA,CAAI,IAAA,CAAK,WAAA,EAAa,SAAS,CAAA;AAAA,MACxD;AAAA,IACF,CAAA,MAAA,IAAW,IAAA,CAAK,WAAA,KAAgB,MAAA,EAAQ;AACtC,MAAA,eAAA,GAAkB,IAAA,CAAK,KAAA,CAAM,IAAA,CAAK,IAAA,CAAK,SAAS,CAAC,CAAA;AAAA,IACnD,CAAA,MAAO;AACL,MAAA,eAAA,GAAkB,IAAA,CAAK,KAAA,CAAM,IAAA,CAAK,IAAA,CAAK,SAAS,CAAC,CAAA;AAAA,IACnD;AACA,IAAA,eAAA,GAAkB,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,eAAe,CAAA;AAE7C,IAAA,MAAM,GAAA,GAAM,KAAK,SAAA,EAAU;AAE3B,IAAA,IAAA,CAAK,QAAQ,EAAC;AAEd,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,aAAa,CAAA,EAAA,EAAK;AACzC,MAAA,MAAM,gBAA0B,EAAC;AACjC,MAAA,IAAI,KAAK,SAAA,EAAW;AAClB,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,aAAA,CAAc,KAAK,IAAA,CAAK,KAAA,CAAM,GAAA,EAAI,GAAI,QAAQ,CAAC,CAAA;AAAA,QACjD;AAAA,MACF,CAAA,MAAO;AACL,QAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,UAAA,aAAA,CAAc,KAAK,CAAC,CAAA;AAAA,QACtB;AAAA,MACF;AAGA,MAAA,MAAM,UAAsB,EAAC;AAC7B,MAAA,MAAM,UAAoB,EAAC;AAC3B,MAAA,KAAA,MAAW,aAAa,aAAA,EAAe;AACrC,QAAA,OAAA,CAAQ,IAAA,CAAK,KAAA,CAAM,SAAS,CAAA,IAAK,EAAE,CAAA;AACnC,QAAA,OAAA,CAAQ,IAAA,CAAK,KAAA,CAAM,SAAS,CAAA,IAAK,CAAC,CAAA;AAAA,MACpC;AAEA,MAAA,MAAM,WAAA,GAMF;AAAA,QACF,UAAU,IAAA,CAAK,QAAA;AAAA,QACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,QACtB,gBAAgB,IAAA,CAAK,cAAA;AAAA,QACrB,WAAA,EAAa;AAAA,OACf;AACA,MAAA,IAAI,IAAA,CAAK,gBAAgB,MAAA,EAAW;AAClC,QAAA,WAAA,CAAY,WAAA,GAAc,KAAK,WAAA,GAAc,CAAA;AAAA,MAC/C;AACA,MAAA,MAAM,IAAA,GAAO,IAAI,qBAAA,CAAsB,WAAW,CAAA;AAClD,MAAA,IAAA,CAAK,IAAI,MAAA,CAAO,OAAO,CAAA,EAAG,MAAA,CAAO,OAAO,CAAC,CAAA;AACzC,MAAA,IAAA,CAAK,KAAA,CAAM,KAAK,IAAI,CAAA;AAAA,IACtB;AAEA,IAAA,IAAA,CAAK,MAAA,GAAS,IAAA;AACd,IAAA,OAAO,IAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAaA,QAAQ,CAAA,EAAmB;AACzB,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,wDAAwD,CAAA;AAAA,IACnF;AAEA,IAAA,qBAAA,CAAsB,CAAA,EAAG,IAAA,CAAK,SAAA,IAAa,CAAA,EAAG,uBAAuB,CAAA;AAErE,IAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAE/B,IAAA,MAAM,iBAA6B,EAAC;AAEpC,IAAA,KAAA,MAAW,IAAA,IAAQ,KAAK,KAAA,EAAO;AAC7B,MAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAC5B,MAAA,MAAM,YAAsB,EAAC;AAC7B,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,SAAA,CAAU,IAAA,CAAK,OAAO,KAAA,CAAM,IAAA,CAAK,MAAM,MAAA,GAAS,CAAC,CAAC,CAAC,CAAA;AAAA,MACrD;AACA,MAAA,cAAA,CAAe,KAAK,SAAS,CAAA;AAAA,IAC/B;AAGA,IAAA,MAAM,mBAA6B,EAAC;AACpC,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,IAAI,GAAA,GAAM,CAAA;AACV,MAAA,KAAA,MAAW,aAAa,cAAA,EAAgB;AACtC,QAAA,GAAA,IAAO,SAAA,CAAU,CAAC,CAAA,IAAK,CAAA;AAAA,MACzB;AACA,MAAA,gBAAA,CAAiB,IAAA,CAAK,GAAA,GAAM,IAAA,CAAK,KAAA,CAAM,MAAM,CAAA;AAAA,IAC/C;AAEA,IAAA,OAAO,OAAO,gBAAgB,CAAA;AAAA,EAChC;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAcA,KAAA,CAAM,GAAW,CAAA,EAAmB;AAClC,IAAA,IAAI,CAAC,KAAK,MAAA,EAAQ;AAChB,MAAA,MAAM,IAAI,eAAe,qDAAqD,CAAA;AAAA,IAChF;AACA,IAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,MAAA,MAAM,IAAI,UAAA,CAAW,CAAA,kCAAA,EAAqC,CAAA,CAAE,IAAI,CAAA,CAAE,CAAA;AAAA,IACpE;AACA,IAAA,gBAAA,CAAiB,GAAG,GAAG,CAAA;AACvB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,MAAM,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAA,IAAK,CAAA;AACpC,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,CAAA,EAAG;AACzB,QAAA,MAAM,IAAI,oBAAoB,2CAA2C,CAAA;AAAA,MAC3E;AAAA,IACF;AACA,IAAA,MAAM,WAAA,GAAc,IAAA,CAAK,OAAA,CAAQ,CAAC,CAAA;AAClC,IAAA,IAAI,WAAA,CAAY,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AAC/B,MAAA,MAAM,IAAI,UAAA;AAAA,QACR,CAAA,oDAAA,EAAuD,WAAA,CAAY,IAAI,CAAA,IAAA,EAAO,EAAE,IAAI,CAAA;AAAA,OACtF;AAAA,IACF;AAEA,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,IAAA,IAAI,KAAA,GAAQ,CAAA;AAEZ,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,KAAA,IAAS,OAAO,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AAAA,IACtC;AACA,IAAA,KAAA,IAAS,CAAA,CAAE,IAAA;AAEX,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,CAAE,MAAM,CAAA,EAAA,EAAK;AAC/B,MAAA,MAAM,QAAQ,MAAA,CAAO,CAAA,CAAE,KAAK,CAAA,CAAE,MAAA,GAAS,CAAC,CAAC,CAAA;AACzC,MAAA,MAAM,QAAQ,MAAA,CAAO,WAAA,CAAY,KAAK,WAAA,CAAY,MAAA,GAAS,CAAC,CAAC,CAAA;AAC7D,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAC5B,MAAA,KAAA,IAAA,CAAU,QAAQ,KAAA,KAAU,CAAA;AAAA,IAC9B;AAEA,IAAA,OAAO,UAAU,CAAA,GAAK,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA,GAAO,IAAI,KAAA,GAAQ,KAAA;AAAA,EAC/D;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAOA,SAAA,GAAqC;AACnC,IAAA,OAAO;AAAA,MACL,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,UAAU,IAAA,CAAK,QAAA;AAAA,MACf,iBAAiB,IAAA,CAAK,eAAA;AAAA,MACtB,gBAAgB,IAAA,CAAK,cAAA;AAAA,MACrB,aAAa,IAAA,CAAK,WAAA;AAAA,MAClB,WAAW,IAAA,CAAK,SAAA;AAAA,MAChB,aAAa,IAAA,CAAK;AAAA,KACpB;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,UAAU,OAAA,EAAwC;AAChD,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AACF","file":"chunk-MLBMYKCG.js","sourcesContent":["// Base types for ML estimators\nexport type {\n  Classifier,\n  Clusterer,\n  Estimator,\n  OutlierDetector,\n  Regressor,\n  Transformer,\n} from \"./base\";\n// Clustering models\nexport { DBSCAN } from \"./clustering/DBSCAN\";\nexport { KMeans } from \"./clustering/KMeans\";\n// Dimensionality reduction\nexport { PCA } from \"./decomposition\";\n// Ensemble methods\nexport {\n  GradientBoostingClassifier,\n  GradientBoostingRegressor,\n} from \"./ensemble\";\n// Linear models - regression and classification\nexport { Lasso } from \"./linear/Lasso\";\nexport { LinearRegression } from \"./linear/LinearRegression\";\nexport { LogisticRegression } from \"./linear/LogisticRegression\";\nexport { Ridge } from \"./linear/Ridge\";\n// Manifold learning\nexport { TSNE } from \"./manifold\";\n// Naive Bayes\nexport { GaussianNB } from \"./naive_bayes\";\n// Neighbors\nexport { KNeighborsClassifier, KNeighborsRegressor } from \"./neighbors\";\n// Support Vector Machines\nexport { LinearSVC, LinearSVR } from \"./svm\";\n// Tree-based models\nexport {\n  DecisionTreeClassifier,\n  DecisionTreeRegressor,\n  RandomForestClassifier,\n  RandomForestRegressor,\n} from \"./tree\";\n","/**\n * Internal validation utilities for ML models.\n * This file is not exported from the public API.\n *\n * @internal\n */\n\nimport { DataValidationError, ShapeError } from \"../core\";\nimport type { Tensor } from \"../ndarray\";\nimport { isContiguous } from \"../ndarray/tensor/strides\";\n\nexport function assertContiguous(t: Tensor, name: string): void {\n  if (!isContiguous(t.shape, t.strides)) {\n    throw new DataValidationError(\n      `${name} must be contiguous in row-major order; materialize a contiguous tensor before passing to ML routines`\n    );\n  }\n}\n\n/**\n * Validate inputs for supervised learning fit methods.\n *\n * Checks:\n * - X is 2D, y is 1D\n * - X and y have matching number of samples\n * - No empty data (at least 1 sample and 1 feature)\n * - No NaN or Inf values\n *\n * @param X - Feature matrix of shape (n_samples, n_features)\n * @param y - Target vector of shape (n_samples,)\n * @throws {ShapeError} If dimensions are invalid\n * @throws {DataValidationError} If data contains invalid values\n *\n * @internal\n */\nexport function validateFitInputs(X: Tensor, y: Tensor): void {\n  // Check dimensions\n  if (X.ndim !== 2) {\n    throw new ShapeError(`X must be 2-dimensional; got ndim=${X.ndim}`);\n  }\n  if (y.ndim !== 1) {\n    throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n  }\n  assertContiguous(X, \"X\");\n  assertContiguous(y, \"y\");\n\n  // Check for empty data\n  const nSamples = X.shape[0] ?? 0;\n  const nFeatures = X.shape[1] ?? 0;\n\n  if (nSamples === 0) {\n    throw new DataValidationError(\"X must have at least one sample\");\n  }\n  if (nFeatures === 0) {\n    throw new DataValidationError(\"X must have at least one feature\");\n  }\n\n  // Check shape match\n  if (nSamples !== y.shape[0]) {\n    throw new ShapeError(\n      `X and y must have the same number of samples; got X.shape[0]=${nSamples}, y.shape[0]=${y.shape[0]}`\n    );\n  }\n\n  // Check for NaN/Inf in X\n  for (let i = 0; i < X.size; i++) {\n    const val = X.data[X.offset + i] ?? 0;\n    if (!Number.isFinite(val)) {\n      throw new DataValidationError(\"X contains non-finite values (NaN or Inf)\");\n    }\n  }\n\n  // Check for NaN/Inf in y\n  for (let i = 0; i < y.size; i++) {\n    const val = y.data[y.offset + i] ?? 0;\n    if (!Number.isFinite(val)) {\n      throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n    }\n  }\n}\n\n/**\n * Validate inputs for unsupervised learning fit methods.\n *\n * Checks:\n * - X is 2D\n * - No empty data (at least 1 sample and 1 feature)\n * - No NaN or Inf values\n *\n * @param X - Feature matrix of shape (n_samples, n_features)\n * @throws {ShapeError} If dimensions are invalid\n * @throws {DataValidationError} If data contains invalid values\n *\n * @internal\n */\nexport function validateUnsupervisedFitInputs(X: Tensor): void {\n  // Check dimensions\n  if (X.ndim !== 2) {\n    throw new ShapeError(`X must be 2-dimensional; got ndim=${X.ndim}`);\n  }\n  assertContiguous(X, \"X\");\n\n  // Check for empty data\n  const nSamples = X.shape[0] ?? 0;\n  const nFeatures = X.shape[1] ?? 0;\n\n  if (nSamples === 0) {\n    throw new DataValidationError(\"X must have at least one sample\");\n  }\n  if (nFeatures === 0) {\n    throw new DataValidationError(\"X must have at least one feature\");\n  }\n\n  // Check for NaN/Inf in X\n  for (let i = 0; i < X.size; i++) {\n    const val = X.data[X.offset + i] ?? 0;\n    if (!Number.isFinite(val)) {\n      throw new DataValidationError(\"X contains non-finite values (NaN or Inf)\");\n    }\n  }\n}\n\n/**\n * Validate inputs for prediction methods.\n *\n * Checks:\n * - X is 2D\n * - X has correct number of features\n * - No NaN or Inf values\n *\n * @param X - Feature matrix of shape (n_samples, n_features)\n * @param nFeaturesExpected - Expected number of features from training\n * @param modelName - Name of the model (for error messages)\n * @throws {ShapeError} If dimensions are invalid\n * @throws {DataValidationError} If data contains invalid values\n *\n * @internal\n */\nexport function validatePredictInputs(\n  X: Tensor,\n  nFeaturesExpected: number,\n  modelName: string\n): void {\n  // Check dimensions\n  if (X.ndim !== 2) {\n    throw new ShapeError(`X must be 2-dimensional; got ndim=${X.ndim}`);\n  }\n  assertContiguous(X, \"X\");\n\n  // Check feature count\n  const nFeatures = X.shape[1] ?? 0;\n  if (nFeatures !== nFeaturesExpected) {\n    throw new ShapeError(\n      `X has ${nFeatures} features but ${modelName} was fitted with ${nFeaturesExpected} features`\n    );\n  }\n\n  // Check for NaN/Inf\n  for (let i = 0; i < X.size; i++) {\n    const val = X.data[X.offset + i] ?? 0;\n    if (!Number.isFinite(val)) {\n      throw new DataValidationError(\"X contains non-finite values (NaN or Inf)\");\n    }\n  }\n}\n","import { InvalidParameterError, NotFittedError, NotImplementedError } from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { validateUnsupervisedFitInputs } from \"../_validation\";\nimport type { Clusterer } from \"../base\";\n\n/**\n * DBSCAN (Density-Based Spatial Clustering of Applications with Noise).\n *\n * Clusters points based on density. Points in high-density regions are\n * grouped together, while points in low-density regions are marked as noise.\n *\n * **Algorithm**:\n * 1. For each point, find all neighbors within eps distance\n * 2. If a point has at least minSamples neighbors, it's a core point\n * 3. Core points and their neighbors form clusters\n * 4. Points not reachable from any core point are noise (label = -1)\n *\n * **Advantages**:\n * - No need to specify number of clusters\n * - Can find arbitrarily shaped clusters\n * - Robust to outliers\n *\n * @example\n * ```ts\n * import { DBSCAN } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1, 2], [2, 2], [2, 3], [8, 7], [8, 8], [25, 80]]);\n * const dbscan = new DBSCAN({ eps: 3, minSamples: 2 });\n * const labels = dbscan.fitPredict(X);\n * // labels: [0, 0, 0, 1, 1, -1]  (-1 = noise)\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.cluster.DBSCAN.html | scikit-learn DBSCAN}\n */\nexport class DBSCAN implements Clusterer {\n  private eps: number;\n  private minSamples: number;\n  private metric: \"euclidean\" | \"manhattan\";\n\n  private labels_?: Tensor;\n  private coreIndices_?: number[];\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly eps?: number;\n      readonly minSamples?: number;\n      readonly metric?: \"euclidean\" | \"manhattan\";\n    } = {}\n  ) {\n    this.eps = options.eps ?? 0.5;\n    this.minSamples = options.minSamples ?? 5;\n    this.metric = options.metric ?? \"euclidean\";\n\n    if (!Number.isFinite(this.eps) || this.eps <= 0) {\n      throw new InvalidParameterError(\"eps must be a finite number > 0\", \"eps\", this.eps);\n    }\n    if (!Number.isInteger(this.minSamples) || this.minSamples < 1) {\n      throw new InvalidParameterError(\n        \"minSamples must be an integer >= 1\",\n        \"minSamples\",\n        this.minSamples\n      );\n    }\n    if (this.metric !== \"euclidean\" && this.metric !== \"manhattan\") {\n      throw new InvalidParameterError(\n        `metric must be \"euclidean\" or \"manhattan\"`,\n        \"metric\",\n        this.metric\n      );\n    }\n  }\n\n  /**\n   * Check if two points are neighbors (distance <= eps).\n   */\n  private isNeighbor(a: number[], b: number[]): boolean {\n    if (this.metric === \"manhattan\") {\n      let sum = 0;\n      for (let i = 0; i < a.length; i++) {\n        sum += Math.abs((a[i] ?? 0) - (b[i] ?? 0));\n        if (sum > this.eps) return false;\n      }\n      return sum <= this.eps;\n    }\n\n    // Euclidean distance\n    let sumSq = 0;\n    const epsSq = this.eps * this.eps;\n    for (let i = 0; i < a.length; i++) {\n      const diff = (a[i] ?? 0) - (b[i] ?? 0);\n      sumSq += diff * diff;\n      if (sumSq > epsSq) return false;\n    }\n    return sumSq <= epsSq;\n  }\n\n  /**\n   * Find all neighbors within eps distance.\n   */\n  private getNeighbors(data: number[][], pointIdx: number): number[] {\n    const neighbors: number[] = [];\n    const point = data[pointIdx];\n    if (!point) return neighbors;\n\n    for (let i = 0; i < data.length; i++) {\n      const other = data[i];\n      if (other && this.isNeighbor(point, other)) {\n        neighbors.push(i);\n      }\n    }\n\n    return neighbors;\n  }\n\n  /**\n   * Perform DBSCAN clustering on data X.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param _y - Ignored (exists for API compatibility)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  fit(X: Tensor, _y?: Tensor): this {\n    validateUnsupervisedFitInputs(X);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    // Extract data\n    const data: number[][] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      data.push(row);\n    }\n\n    // Initialize labels to undefined (-2 means unvisited)\n    const labels: number[] = new Array(nSamples).fill(-2);\n    const coreIndices: number[] = [];\n\n    let clusterId = 0;\n\n    for (let i = 0; i < nSamples; i++) {\n      // Skip if already processed\n      if (labels[i] !== -2) continue;\n\n      const neighbors = this.getNeighbors(data, i);\n\n      if (neighbors.length < this.minSamples) {\n        // Mark as noise (for now, might be claimed by another cluster later)\n        labels[i] = -1;\n        continue;\n      }\n\n      // Start a new cluster\n      coreIndices.push(i);\n      labels[i] = clusterId;\n\n      // Process neighbors\n      const seedSet = new Set(neighbors);\n\n      for (const q of seedSet) {\n        // If noise, claim it for this cluster\n        if (labels[q] === -1) {\n          labels[q] = clusterId;\n        }\n\n        // If unvisited\n        if (labels[q] === -2) {\n          labels[q] = clusterId;\n\n          const qNeighbors = this.getNeighbors(data, q);\n\n          if (qNeighbors.length >= this.minSamples) {\n            coreIndices.push(q);\n            // Add new neighbors to seed set\n            for (const n of qNeighbors) {\n              if (labels[n] === -2 || labels[n] === -1) {\n                seedSet.add(n);\n              }\n            }\n          }\n        }\n      }\n\n      clusterId++;\n    }\n\n    this.labels_ = tensor(labels, { dtype: \"int32\" });\n    this.coreIndices_ = coreIndices;\n    this.fitted = true;\n\n    return this;\n  }\n\n  /**\n   * Predict cluster labels for samples in X.\n   *\n   * @param _X - Samples (unused)\n   * @throws {NotImplementedError} Always — DBSCAN is transductive and does not support prediction on new data\n   */\n  predict(_X: Tensor): Tensor {\n    throw new NotImplementedError(\n      \"DBSCAN is a transductive clustering algorithm and does not support prediction on new data. Use fitPredict() instead.\"\n    );\n  }\n\n  /**\n   * Fit DBSCAN and return cluster labels.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param _y - Ignored (exists for API compatibility)\n   * @returns Cluster labels of shape (n_samples,). Noise points are labeled -1.\n   * @throws {ShapeError} If X is not 2D\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   * @throws {NotFittedError} If fit did not produce labels (internal error)\n   */\n  fitPredict(X: Tensor, _y?: Tensor): Tensor {\n    this.fit(X);\n    if (!this.labels_) {\n      throw new NotFittedError(\"DBSCAN fit did not produce labels\");\n    }\n    return this.labels_;\n  }\n\n  /**\n   * Get cluster labels assigned during fitting.\n   *\n   * @returns Tensor of cluster labels. Noise points are labeled -1.\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get labels(): Tensor {\n    if (!this.fitted || !this.labels_) {\n      throw new NotFittedError(\"DBSCAN must be fitted to access labels\");\n    }\n    return this.labels_;\n  }\n\n  /**\n   * Get indices of core samples discovered during fitting.\n   *\n   * Core samples are points with at least `minSamples` neighbors within `eps`.\n   *\n   * @returns Array of core sample indices\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get coreIndices(): number[] {\n    if (!this.fitted || !this.coreIndices_) {\n      throw new NotFittedError(\"DBSCAN must be fitted to access core indices\");\n    }\n    return [...this.coreIndices_];\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      eps: this.eps,\n      minSamples: this.minSamples,\n      metric: this.metric,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param params - Parameters to set (eps, minSamples, metric)\n   * @returns this\n   * @throws {InvalidParameterError} If any parameter value is invalid\n   */\n  setParams(params: Record<string, unknown>): this {\n    for (const [key, value] of Object.entries(params)) {\n      switch (key) {\n        case \"eps\":\n          if (typeof value !== \"number\" || value <= 0) {\n            throw new InvalidParameterError(\"eps must be > 0\", \"eps\", value);\n          }\n          this.eps = value;\n          break;\n        case \"minSamples\":\n          if (typeof value !== \"number\" || !Number.isInteger(value) || value < 1) {\n            throw new InvalidParameterError(\n              \"minSamples must be an integer >= 1\",\n              \"minSamples\",\n              value\n            );\n          }\n          this.minSamples = value;\n          break;\n        case \"metric\":\n          if (value !== \"euclidean\" && value !== \"manhattan\") {\n            throw new InvalidParameterError(\n              `metric must be \"euclidean\" or \"manhattan\"`,\n              \"metric\",\n              value\n            );\n          }\n          this.metric = value;\n          break;\n        default:\n          throw new InvalidParameterError(`Unknown parameter: ${key}`, key, value);\n      }\n    }\n    return this;\n  }\n}\n","import { InvalidParameterError, NotFittedError } from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { validatePredictInputs, validateUnsupervisedFitInputs } from \"../_validation\";\nimport type { Clusterer } from \"../base\";\n\n/**\n * K-Means clustering algorithm.\n *\n * Partitions n samples into k clusters by minimizing the within-cluster\n * sum of squared distances to cluster centroids.\n *\n * **Algorithm**: Lloyd's algorithm (iterative refinement)\n * 1. Initialize k centroids (random or k-means++)\n * 2. Assign each point to nearest centroid\n * 3. Update centroids as mean of assigned points\n * 4. Repeat until convergence or max iterations\n *\n * **Time Complexity**: O(n * k * i * d) where:\n * - n = number of samples\n * - k = number of clusters\n * - i = number of iterations\n * - d = number of features\n *\n * @example\n * ```ts\n * import { KMeans } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1, 2], [1.5, 1.8], [5, 8], [8, 8], [1, 0.6], [9, 11]]);\n * const kmeans = new KMeans({ nClusters: 2, randomState: 42 });\n * kmeans.fit(X);\n *\n * const labels = kmeans.predict(X);\n * console.log('Cluster labels:', labels);\n * console.log('Centroids:', kmeans.clusterCenters);\n * ```\n *\n * @see {@link https://en.wikipedia.org/wiki/K-means_clustering | Wikipedia: K-means}\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html | scikit-learn KMeans}\n */\nexport class KMeans implements Clusterer {\n  private nClusters: number;\n  private maxIter: number;\n  private tol: number;\n  private init: \"random\" | \"kmeans++\";\n  private randomState: number | undefined;\n\n  private clusterCenters_?: Tensor;\n  private labels_?: Tensor;\n  private inertia_?: number;\n  private nIter_?: number;\n  private nFeaturesIn_?: number;\n  private fitted = false;\n\n  /**\n   * Create a new K-Means clustering model.\n   *\n   * @param options - Configuration options\n   * @param options.nClusters - Number of clusters (default: 8)\n   * @param options.maxIter - Maximum number of iterations (default: 300)\n   * @param options.tol - Tolerance for convergence (default: 1e-4)\n   * @param options.init - Initialization method: 'random' or 'kmeans++' (default: 'kmeans++')\n   * @param options.randomState - Random seed for reproducibility\n   */\n  constructor(\n    options: {\n      readonly nClusters?: number;\n      readonly maxIter?: number;\n      readonly tol?: number;\n      readonly init?: \"random\" | \"kmeans++\";\n      readonly randomState?: number;\n    } = {}\n  ) {\n    this.nClusters = options.nClusters ?? 8;\n    this.maxIter = options.maxIter ?? 300;\n    this.tol = options.tol ?? 1e-4;\n    this.init = options.init ?? \"kmeans++\";\n    if (options.randomState !== undefined) {\n      this.randomState = options.randomState;\n    }\n\n    if (!Number.isInteger(this.nClusters) || this.nClusters < 1) {\n      throw new InvalidParameterError(\n        \"nClusters must be an integer >= 1\",\n        \"nClusters\",\n        this.nClusters\n      );\n    }\n    if (!Number.isInteger(this.maxIter) || this.maxIter < 1) {\n      throw new InvalidParameterError(\"maxIter must be an integer >= 1\", \"maxIter\", this.maxIter);\n    }\n    if (!Number.isFinite(this.tol) || this.tol < 0) {\n      throw new InvalidParameterError(\"tol must be a finite number >= 0\", \"tol\", this.tol);\n    }\n    if (this.init !== \"random\" && this.init !== \"kmeans++\") {\n      throw new InvalidParameterError(\n        `init must be \"random\" or \"kmeans++\"; received ${String(this.init)}`,\n        \"init\",\n        this.init\n      );\n    }\n    if (options.randomState !== undefined && !Number.isFinite(options.randomState)) {\n      throw new InvalidParameterError(\n        `randomState must be a finite number; received ${String(options.randomState)}`,\n        \"randomState\",\n        options.randomState\n      );\n    }\n  }\n\n  /**\n   * Fit K-Means clustering on training data.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Ignored (exists for compatibility)\n   * @returns this - The fitted estimator\n   */\n  fit(X: Tensor, _y?: Tensor): this {\n    validateUnsupervisedFitInputs(X);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    this.nFeaturesIn_ = nFeatures;\n\n    if (nSamples < this.nClusters) {\n      throw new InvalidParameterError(\n        `n_samples=${nSamples} should be >= n_clusters=${this.nClusters}`,\n        \"nClusters\",\n        this.nClusters\n      );\n    }\n\n    // Initialize centroids\n    let centroids = this.initializeCentroids(X);\n\n    let prevInertia = Number.POSITIVE_INFINITY;\n\n    for (let iter = 0; iter < this.maxIter; iter++) {\n      // Assign points to nearest centroid\n      const labels = this.assignClusters(X, centroids);\n\n      // Update centroids\n      const newCentroids: number[][] = [];\n      for (let k = 0; k < this.nClusters; k++) {\n        const clusterPoints: number[][] = [];\n        for (let i = 0; i < nSamples; i++) {\n          if (Number(labels.data[labels.offset + i]) === k) {\n            const point: number[] = [];\n            for (let j = 0; j < nFeatures; j++) {\n              point.push(Number(X.data[X.offset + i * nFeatures + j]));\n            }\n            clusterPoints.push(point);\n          }\n        }\n\n        if (clusterPoints.length > 0) {\n          const centroid: number[] = [];\n          for (let j = 0; j < nFeatures; j++) {\n            let sum = 0;\n            for (const point of clusterPoints) {\n              sum += point[j] ?? 0;\n            }\n            centroid.push(sum / clusterPoints.length);\n          }\n          newCentroids.push(centroid);\n        } else {\n          // Keep old centroid if no points assigned\n          const oldCentroid: number[] = [];\n          for (let j = 0; j < nFeatures; j++) {\n            oldCentroid.push(Number(centroids.data[centroids.offset + k * nFeatures + j]));\n          }\n          newCentroids.push(oldCentroid);\n        }\n      }\n\n      centroids = tensor(newCentroids);\n\n      // Calculate inertia (sum of squared distances to centroids)\n      const inertia = this.calculateInertia(X, centroids, labels);\n\n      // Check convergence\n      if (Math.abs(prevInertia - inertia) < this.tol) {\n        this.nIter_ = iter + 1;\n        break;\n      }\n\n      prevInertia = inertia;\n      this.nIter_ = iter + 1;\n    }\n\n    this.clusterCenters_ = centroids;\n    this.labels_ = this.assignClusters(X, centroids);\n    this.inertia_ = this.calculateInertia(X, centroids, this.labels_);\n    this.fitted = true;\n\n    return this;\n  }\n\n  /**\n   * Predict cluster labels for samples.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Cluster labels of shape (n_samples,)\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted || !this.clusterCenters_) {\n      throw new NotFittedError(\"KMeans must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"KMeans\");\n\n    return this.assignClusters(X, this.clusterCenters_);\n  }\n\n  /**\n   * Fit and predict in one step.\n   *\n   * @param X - Training data\n   * @param y - Ignored (exists for compatibility)\n   * @returns Cluster labels\n   */\n  fitPredict(X: Tensor, _y?: Tensor): Tensor {\n    this.fit(X);\n    if (!this.labels_) {\n      throw new NotFittedError(\"KMeans fit did not produce labels\");\n    }\n    return this.labels_;\n  }\n\n  /**\n   * Initialize centroids using specified method.\n   */\n  private initializeCentroids(X: Tensor): Tensor {\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    if (this.init === \"random\") {\n      // Random initialization\n      const indices = new Set<number>();\n      const rng = this.createRNG();\n\n      while (indices.size < this.nClusters) {\n        indices.add(Math.floor(rng() * nSamples));\n      }\n\n      const centroids: number[][] = [];\n      for (const idx of indices) {\n        const centroid: number[] = [];\n        for (let j = 0; j < nFeatures; j++) {\n          centroid.push(Number(X.data[X.offset + idx * nFeatures + j]));\n        }\n        centroids.push(centroid);\n      }\n\n      return tensor(centroids);\n    } else {\n      // K-means++ initialization\n      const rng = this.createRNG();\n      const centroids: number[][] = [];\n      const minDistSq = new Float64Array(nSamples).fill(Infinity);\n\n      // Choose first centroid randomly\n      const firstIdx = Math.floor(rng() * nSamples);\n      const firstCentroid: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        firstCentroid.push(Number(X.data[X.offset + firstIdx * nFeatures + j]));\n      }\n      centroids.push(firstCentroid);\n\n      // Update distances for first centroid\n      for (let i = 0; i < nSamples; i++) {\n        let dist = 0;\n        for (let j = 0; j < nFeatures; j++) {\n          const diff = Number(X.data[X.offset + i * nFeatures + j]) - (firstCentroid[j] ?? 0);\n          dist += diff * diff;\n        }\n        minDistSq[i] = dist;\n      }\n\n      // Choose remaining centroids\n      for (let k = 1; k < this.nClusters; k++) {\n        // Choose next centroid with probability proportional to distance squared\n        const totalDist = minDistSq.reduce((a, b) => a + b, 0);\n        let r = rng() * totalDist;\n        let nextIdx = 0;\n\n        for (let i = 0; i < nSamples; i++) {\n          r -= minDistSq[i] ?? 0;\n          if (r <= 0) {\n            nextIdx = i;\n            break;\n          }\n        }\n\n        const nextCentroid: number[] = [];\n        for (let j = 0; j < nFeatures; j++) {\n          nextCentroid.push(Number(X.data[X.offset + nextIdx * nFeatures + j]));\n        }\n        centroids.push(nextCentroid);\n\n        // Update minimum squared distances\n        for (let i = 0; i < nSamples; i++) {\n          let dist = 0;\n          for (let j = 0; j < nFeatures; j++) {\n            const diff = Number(X.data[X.offset + i * nFeatures + j]) - (nextCentroid[j] ?? 0);\n            dist += diff * diff;\n          }\n          if (dist < (minDistSq[i] ?? Infinity)) {\n            minDistSq[i] = dist;\n          }\n        }\n      }\n\n      return tensor(centroids);\n    }\n  }\n\n  /**\n   * Assign each sample to nearest centroid.\n   */\n  private assignClusters(X: Tensor, centroids: Tensor): Tensor {\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const labels: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      let minDist = Number.POSITIVE_INFINITY;\n      let minLabel = 0;\n\n      for (let k = 0; k < this.nClusters; k++) {\n        let dist = 0;\n        for (let j = 0; j < nFeatures; j++) {\n          const diff =\n            Number(X.data[X.offset + i * nFeatures + j]) -\n            Number(centroids.data[centroids.offset + k * nFeatures + j]);\n          dist += diff * diff;\n        }\n\n        if (dist < minDist) {\n          minDist = dist;\n          minLabel = k;\n        }\n      }\n\n      labels.push(minLabel);\n    }\n\n    return tensor(labels, { dtype: \"int32\" });\n  }\n\n  /**\n   * Calculate inertia (sum of squared distances to centroids).\n   */\n  private calculateInertia(X: Tensor, centroids: Tensor, labels: Tensor): number {\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    let inertia = 0;\n\n    for (let i = 0; i < nSamples; i++) {\n      const label = Number(labels.data[labels.offset + i]);\n      for (let j = 0; j < nFeatures; j++) {\n        const diff =\n          Number(X.data[X.offset + i * nFeatures + j]) -\n          Number(centroids.data[centroids.offset + label * nFeatures + j]);\n        inertia += diff * diff;\n      }\n    }\n\n    return inertia;\n  }\n\n  /**\n   * Create a simple RNG for reproducibility.\n   */\n  private createRNG(): () => number {\n    if (this.randomState !== undefined) {\n      let seed = this.randomState;\n      return () => {\n        seed = (seed * 9301 + 49297) % 233280;\n        return seed / 233280;\n      };\n    }\n    return Math.random;\n  }\n\n  /**\n   * Get cluster centers.\n   */\n  get clusterCenters(): Tensor {\n    if (!this.fitted || !this.clusterCenters_) {\n      throw new NotFittedError(\"KMeans must be fitted to access cluster centers\");\n    }\n    return this.clusterCenters_;\n  }\n\n  /**\n   * Get training labels.\n   */\n  get labels(): Tensor {\n    if (!this.fitted || !this.labels_) {\n      throw new NotFittedError(\"KMeans must be fitted to access labels\");\n    }\n    return this.labels_;\n  }\n\n  /**\n   * Get inertia (sum of squared distances to centroids).\n   */\n  get inertia(): number {\n    if (!this.fitted || this.inertia_ === undefined) {\n      throw new NotFittedError(\"KMeans must be fitted to access inertia\");\n    }\n    return this.inertia_;\n  }\n\n  /**\n   * Get number of iterations run.\n   */\n  get nIter(): number {\n    if (!this.fitted || this.nIter_ === undefined) {\n      throw new NotFittedError(\"KMeans must be fitted to access n_iter\");\n    }\n    return this.nIter_;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nClusters: this.nClusters,\n      maxIter: this.maxIter,\n      tol: this.tol,\n      init: this.init,\n      randomState: this.randomState,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param params - Parameters to set (nClusters, maxIter, tol, init, randomState)\n   * @returns this\n   * @throws {InvalidParameterError} If any parameter value is invalid\n   */\n  setParams(params: Record<string, unknown>): this {\n    for (const [key, value] of Object.entries(params)) {\n      switch (key) {\n        case \"nClusters\":\n          if (typeof value !== \"number\" || !Number.isInteger(value) || value < 1) {\n            throw new InvalidParameterError(\n              \"nClusters must be an integer >= 1\",\n              \"nClusters\",\n              value\n            );\n          }\n          this.nClusters = value;\n          break;\n        case \"maxIter\":\n          if (typeof value !== \"number\" || !Number.isInteger(value) || value < 1) {\n            throw new InvalidParameterError(\"maxIter must be an integer >= 1\", \"maxIter\", value);\n          }\n          this.maxIter = value;\n          break;\n        case \"tol\":\n          if (typeof value !== \"number\" || value < 0) {\n            throw new InvalidParameterError(\"tol must be >= 0\", \"tol\", value);\n          }\n          this.tol = value;\n          break;\n        case \"init\":\n          if (value !== \"random\" && value !== \"kmeans++\") {\n            throw new InvalidParameterError(`init must be \"random\" or \"kmeans++\"`, \"init\", value);\n          }\n          this.init = value;\n          break;\n        case \"randomState\":\n          if (value !== undefined && (typeof value !== \"number\" || !Number.isFinite(value))) {\n            throw new InvalidParameterError(\n              \"randomState must be a finite number\",\n              \"randomState\",\n              value\n            );\n          }\n          this.randomState = value === undefined ? undefined : value;\n          break;\n        default:\n          throw new InvalidParameterError(`Unknown parameter: ${key}`, key, value);\n      }\n    }\n    return this;\n  }\n}\n","import {\n  DataValidationError,\n  InvalidParameterError,\n  NotFittedError,\n  NotImplementedError,\n  ShapeError,\n} from \"../../core\";\nimport { svd } from \"../../linalg\";\nimport { mean, type Tensor, tensor } from \"../../ndarray\";\nimport {\n  assertContiguous,\n  validatePredictInputs,\n  validateUnsupervisedFitInputs,\n} from \"../_validation\";\nimport type { Transformer } from \"../base\";\n\n/**\n * Principal Component Analysis (PCA).\n *\n * Linear dimensionality reduction using Singular Value Decomposition (SVD)\n * to project data to a lower dimensional space.\n *\n * **Algorithm**:\n * 1. Center the data by subtracting the mean\n * 2. Compute SVD: X = U * Σ * V^T\n * 3. Principal components are columns of V\n * 4. Transform data by projecting onto principal components\n *\n * **Time Complexity**: O(min(n*d^2, d*n^2)) where n=samples, d=features\n *\n * @example\n * ```ts\n * import { PCA } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[2.5, 2.4], [0.5, 0.7], [2.2, 2.9], [1.9, 2.2], [3.1, 3.0]]);\n * const pca = new PCA({ nComponents: 1 });\n * pca.fit(X);\n *\n * const XTransformed = pca.transform(X);\n * console.log('Explained variance ratio:', pca.explainedVarianceRatio);\n * ```\n *\n * @see {@link https://en.wikipedia.org/wiki/Principal_component_analysis | Wikipedia: PCA}\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.decomposition.PCA.html | scikit-learn PCA}\n */\nexport class PCA implements Transformer {\n  private readonly nComponents?: number;\n  private readonly whiten: boolean;\n\n  private components_?: Tensor;\n  private explainedVariance_?: Tensor;\n  private explainedVarianceRatio_?: Tensor;\n  private mean_?: Tensor;\n  private nComponentsActual_?: number;\n  private nFeaturesIn_?: number;\n  private fitted = false;\n\n  /**\n   * Create a new PCA model.\n   *\n   * @param options - Configuration options\n   * @param options.nComponents - Number of components to keep (default: min(n_samples, n_features))\n   * @param options.whiten - Whether to whiten the data (default: false)\n   */\n  constructor(\n    options: {\n      readonly nComponents?: number;\n      readonly whiten?: boolean;\n    } = {}\n  ) {\n    if (options.nComponents !== undefined) {\n      this.nComponents = options.nComponents;\n    }\n    this.whiten = options.whiten ?? false;\n\n    if (this.nComponents !== undefined) {\n      if (!Number.isInteger(this.nComponents) || this.nComponents < 1) {\n        throw new InvalidParameterError(\n          \"nComponents must be an integer >= 1\",\n          \"nComponents\",\n          this.nComponents\n        );\n      }\n    }\n  }\n\n  /**\n   * Fit PCA on training data.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Ignored (exists for compatibility)\n   * @returns this\n   */\n  fit(X: Tensor, _y?: Tensor): this {\n    validateUnsupervisedFitInputs(X);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    this.nFeaturesIn_ = nFeatures;\n\n    if (nSamples < 2) {\n      throw new DataValidationError(\"X must have at least 2 samples for PCA\");\n    }\n\n    // Determine number of components\n    const nComponentsActual = this.nComponents ?? Math.min(nSamples, nFeatures);\n    if (nComponentsActual > Math.min(nSamples, nFeatures)) {\n      throw new InvalidParameterError(\n        `nComponents=${nComponentsActual} must be <= min(n_samples, n_features)=${Math.min(nSamples, nFeatures)}`,\n        \"nComponents\",\n        nComponentsActual\n      );\n    }\n\n    // Center the data\n    const meanVec = mean(X, 0);\n    this.mean_ = meanVec;\n\n    const XCentered = this.centerData(X, meanVec);\n\n    // Compute SVD\n    const [_U, s, Vt] = svd(XCentered, false);\n\n    // Extract components (rows of Vt are principal components)\n    const components: number[][] = [];\n    for (let i = 0; i < nComponentsActual; i++) {\n      const component: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        component.push(Number(Vt.data[Vt.offset + i * nFeatures + j]));\n      }\n      components.push(component);\n    }\n    this.components_ = tensor(components);\n\n    // Compute explained variance\n    const explainedVariance: number[] = [];\n    for (let i = 0; i < nComponentsActual; i++) {\n      const sv = Number(s.data[s.offset + i]);\n      explainedVariance.push((sv * sv) / (nSamples - 1));\n    }\n    this.explainedVariance_ = tensor(explainedVariance);\n\n    // Compute explained variance ratio\n    let totalVariance = 0;\n    for (let i = 0; i < s.size; i++) {\n      const sv = Number(s.data[s.offset + i]);\n      totalVariance += (sv * sv) / (nSamples - 1);\n    }\n    const explainedVarianceRatio =\n      totalVariance === 0\n        ? explainedVariance.map(() => 0)\n        : explainedVariance.map((v) => v / totalVariance);\n    this.explainedVarianceRatio_ = tensor(explainedVarianceRatio);\n\n    this.nComponentsActual_ = nComponentsActual;\n    this.fitted = true;\n\n    return this;\n  }\n\n  /**\n   * Transform data to principal component space.\n   *\n   * @param X - Data of shape (n_samples, n_features)\n   * @returns Transformed data of shape (n_samples, n_components)\n   */\n  transform(X: Tensor): Tensor {\n    if (!this.fitted || !this.components_ || !this.mean_) {\n      throw new NotFittedError(\"PCA must be fitted before transform\");\n    }\n\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"PCA\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const nComponents = this.nComponentsActual_ ?? 0;\n\n    // Center the data\n    const XCentered = this.centerData(X, this.mean_);\n\n    // Project onto principal components: X_transformed = X_centered @ components.T\n    const transformed: number[][] = [];\n    const varianceEps = 1e-12;\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let k = 0; k < nComponents; k++) {\n        let sum = 0;\n        for (let j = 0; j < nFeatures; j++) {\n          sum +=\n            Number(XCentered.data[XCentered.offset + i * nFeatures + j]) *\n            Number(this.components_.data[this.components_.offset + k * nFeatures + j]);\n        }\n        // If whitening is enabled, scale each component to unit variance.\n        if (this.whiten) {\n          const variance = Number(\n            this.explainedVariance_?.data[this.explainedVariance_.offset + k] ?? 0\n          );\n          row.push(sum / Math.sqrt(variance + varianceEps));\n        } else {\n          row.push(sum);\n        }\n      }\n      transformed.push(row);\n    }\n\n    return tensor(transformed);\n  }\n\n  /**\n   * Fit and transform in one step.\n   *\n   * @param X - Training data\n   * @param y - Ignored (exists for compatibility)\n   * @returns Transformed data\n   */\n  fitTransform(X: Tensor, _y?: Tensor): Tensor {\n    this.fit(X);\n    return this.transform(X);\n  }\n\n  /**\n   * Transform data back to original space.\n   *\n   * @param X - Transformed data of shape (n_samples, n_components)\n   * @returns Reconstructed data of shape (n_samples, n_features)\n   */\n  inverseTransform(X: Tensor): Tensor {\n    if (!this.fitted || !this.components_ || !this.mean_) {\n      throw new NotFittedError(\"PCA must be fitted before inverse transform\");\n    }\n\n    if (X.ndim !== 2) {\n      throw new ShapeError(`X must be 2-dimensional; got ndim=${X.ndim}`);\n    }\n    assertContiguous(X, \"X\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nComponents = this.nComponentsActual_ ?? 0;\n    const nFeatures = this.components_.shape[1] ?? 0;\n    if ((X.shape[1] ?? 0) !== nComponents) {\n      throw new ShapeError(\n        `X must have ${nComponents} components; got ${(X.shape[1] ?? 0).toString()}`\n      );\n    }\n\n    for (let i = 0; i < X.size; i++) {\n      const val = X.data[X.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"X contains non-finite values (NaN or Inf)\");\n      }\n    }\n\n    // Reconstruct: X_reconstructed = X_transformed @ components\n    const reconstructed: number[][] = [];\n    const varianceEps = 1e-12;\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        let sum = 0;\n        for (let k = 0; k < nComponents; k++) {\n          const xVal = Number(X.data[X.offset + i * nComponents + k]);\n          const variance = Number(\n            this.explainedVariance_?.data[this.explainedVariance_.offset + k] ?? 0\n          );\n          // Undo whitening by restoring the original component scale.\n          const scaled = this.whiten ? xVal * Math.sqrt(variance + varianceEps) : xVal;\n          sum +=\n            scaled * Number(this.components_.data[this.components_.offset + k * nFeatures + j]);\n        }\n        // Add back the mean\n        sum += Number(this.mean_.data[this.mean_.offset + j]);\n        row.push(sum);\n      }\n      reconstructed.push(row);\n    }\n\n    return tensor(reconstructed);\n  }\n\n  /**\n   * Center data by subtracting mean.\n   */\n  private centerData(X: Tensor, meanVec: Tensor): Tensor {\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const centered: number[][] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        const val = Number(X.data[X.offset + i * nFeatures + j]);\n        const meanVal = Number(meanVec.data[meanVec.offset + j]);\n        row.push(val - meanVal);\n      }\n      centered.push(row);\n    }\n\n    return tensor(centered);\n  }\n\n  /**\n   * Get principal components.\n   */\n  get components(): Tensor {\n    if (!this.fitted || !this.components_) {\n      throw new NotFittedError(\"PCA must be fitted to access components\");\n    }\n    return this.components_;\n  }\n\n  /**\n   * Get explained variance.\n   */\n  get explainedVariance(): Tensor {\n    if (!this.fitted || !this.explainedVariance_) {\n      throw new NotFittedError(\"PCA must be fitted to access explained variance\");\n    }\n    return this.explainedVariance_;\n  }\n\n  /**\n   * Get explained variance ratio.\n   */\n  get explainedVarianceRatio(): Tensor {\n    if (!this.fitted || !this.explainedVarianceRatio_) {\n      throw new NotFittedError(\"PCA must be fitted to access explained variance ratio\");\n    }\n    return this.explainedVarianceRatio_;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nComponents: this.nComponents,\n      whiten: this.whiten,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\"PCA does not support setParams after construction\");\n  }\n}\n","import {\n  DataValidationError,\n  DeepboxError,\n  InvalidParameterError,\n  NotFittedError,\n  NotImplementedError,\n  ShapeError,\n} from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Classifier, Regressor } from \"../base\";\n\ntype TreeNode = {\n  readonly isLeaf: boolean;\n  readonly prediction?: number | undefined;\n  readonly classProbabilities?: number[] | undefined;\n  readonly featureIndex?: number;\n  readonly threshold?: number;\n  readonly left?: TreeNode;\n  readonly right?: TreeNode;\n};\n\n/**\n * Decision Tree Classifier.\n *\n * A non-parametric supervised learning method that learns simple decision rules\n * inferred from the data features.\n *\n * **Algorithm**: CART (Classification and Regression Trees)\n * - Uses Gini impurity for classification\n * - Recursively splits data based on feature thresholds\n * - Supports max_depth and min_samples_split for regularization\n *\n * @example\n * ```ts\n * import { DecisionTreeClassifier } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1, 2], [3, 4], [5, 6], [7, 8]]);\n * const y = tensor([0, 0, 1, 1]);\n *\n * const clf = new DecisionTreeClassifier({ maxDepth: 3 });\n * clf.fit(X, y);\n * const predictions = clf.predict(X);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html | scikit-learn DecisionTreeClassifier}\n */\nexport class DecisionTreeClassifier implements Classifier {\n  private maxDepth: number;\n  private minSamplesSplit: number;\n  private minSamplesLeaf: number;\n  private maxFeatures: number | undefined;\n  private randomState: number | undefined;\n\n  private tree?: TreeNode;\n  private nFeatures?: number;\n  private classLabels?: number[];\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly maxDepth?: number;\n      readonly minSamplesSplit?: number;\n      readonly minSamplesLeaf?: number;\n      readonly maxFeatures?: number;\n      readonly randomState?: number;\n    } = {}\n  ) {\n    this.maxDepth = options.maxDepth ?? 10;\n    this.minSamplesSplit = options.minSamplesSplit ?? 2;\n    this.minSamplesLeaf = options.minSamplesLeaf ?? 1;\n    if (options.maxFeatures !== undefined) {\n      this.maxFeatures = options.maxFeatures;\n    }\n    if (options.randomState !== undefined) {\n      this.randomState = options.randomState;\n    }\n\n    if (this.randomState !== undefined && !Number.isFinite(this.randomState)) {\n      throw new InvalidParameterError(\n        `randomState must be a finite number; received ${String(this.randomState)}`,\n        \"randomState\",\n        this.randomState\n      );\n    }\n\n    if (!Number.isInteger(this.maxDepth) || this.maxDepth < 1) {\n      throw new InvalidParameterError(\n        `maxDepth must be an integer >= 1; received ${this.maxDepth}`,\n        \"maxDepth\",\n        this.maxDepth\n      );\n    }\n    if (!Number.isInteger(this.minSamplesSplit) || this.minSamplesSplit < 2) {\n      throw new InvalidParameterError(\n        `minSamplesSplit must be an integer >= 2; received ${this.minSamplesSplit}`,\n        \"minSamplesSplit\",\n        this.minSamplesSplit\n      );\n    }\n    if (!Number.isInteger(this.minSamplesLeaf) || this.minSamplesLeaf < 1) {\n      throw new InvalidParameterError(\n        `minSamplesLeaf must be an integer >= 1; received ${this.minSamplesLeaf}`,\n        \"minSamplesLeaf\",\n        this.minSamplesLeaf\n      );\n    }\n    if (\n      this.maxFeatures !== undefined &&\n      (!Number.isInteger(this.maxFeatures) || this.maxFeatures < 1)\n    ) {\n      throw new InvalidParameterError(\n        `maxFeatures must be an integer >= 1; received ${this.maxFeatures}`,\n        \"maxFeatures\",\n        this.maxFeatures\n      );\n    }\n  }\n\n  private getRng(): () => number {\n    if (this.randomState === undefined) {\n      return Math.random;\n    }\n    let seed = this.randomState;\n    return () => {\n      seed = (seed * 1664525 + 1013904223) >>> 0;\n      return seed / 4294967296;\n    };\n  }\n\n  /**\n   * Build a decision tree classifier from the training set (X, y).\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target class labels of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    this.nFeatures = nFeatures;\n\n    // Extract data as arrays\n    const XData: number[][] = [];\n    const yData: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      XData.push(row);\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    // Get unique classes\n    this.classLabels = [...new Set(yData)].sort((a, b) => a - b);\n\n    // Build tree\n    const indices = Array.from({ length: nSamples }, (_, i) => i);\n    this.tree = this.buildTree(XData, yData, indices, 0);\n    this.fitted = true;\n\n    return this;\n  }\n\n  private buildTree(\n    XData: number[][],\n    yData: number[],\n    indices: number[],\n    depth: number\n  ): TreeNode {\n    const n = indices.length;\n\n    // Check stopping conditions\n    if (depth >= this.maxDepth || n < this.minSamplesSplit || n < this.minSamplesLeaf) {\n      return {\n        isLeaf: true,\n        prediction: this.getMajorityClass(yData, indices),\n        classProbabilities: this.getClassProbabilities(yData, indices),\n      };\n    }\n\n    // Check if all samples have same class\n    const classes = new Set(indices.map((i) => yData[i]));\n    if (classes.size === 1) {\n      const firstIdx = indices[0] ?? 0;\n      return {\n        isLeaf: true,\n        prediction: yData[firstIdx] ?? 0,\n        classProbabilities: this.getClassProbabilities(yData, indices),\n      };\n    }\n\n    // Find best split\n    const { featureIndex, threshold, leftIndices, rightIndices } = this.findBestSplit(\n      XData,\n      yData,\n      indices\n    );\n\n    if (leftIndices.length === 0 || rightIndices.length === 0) {\n      return {\n        isLeaf: true,\n        prediction: this.getMajorityClass(yData, indices),\n        classProbabilities: this.getClassProbabilities(yData, indices),\n      };\n    }\n\n    // Recursively build subtrees\n    const left = this.buildTree(XData, yData, leftIndices, depth + 1);\n    const right = this.buildTree(XData, yData, rightIndices, depth + 1);\n\n    return {\n      isLeaf: false,\n      featureIndex,\n      threshold,\n      left,\n      right,\n    };\n  }\n\n  private getMajorityClass(yData: number[], indices: number[]): number {\n    const counts = new Map<number, number>();\n    for (const i of indices) {\n      const label = yData[i] ?? 0;\n      counts.set(label, (counts.get(label) ?? 0) + 1);\n    }\n\n    let maxCount = 0;\n    let maxLabel = 0;\n    for (const [label, count] of counts) {\n      if (count > maxCount) {\n        maxCount = count;\n        maxLabel = label;\n      }\n    }\n    return maxLabel;\n  }\n\n  private getClassProbabilities(yData: number[], indices: number[]): number[] {\n    const labels = this.classLabels ?? [];\n    if (labels.length === 0 || indices.length === 0) {\n      return [];\n    }\n\n    const labelIndex = new Map<number, number>();\n    for (let i = 0; i < labels.length; i++) {\n      const v = labels[i];\n      if (v !== undefined) labelIndex.set(v, i);\n    }\n\n    const counts = new Array<number>(labels.length).fill(0);\n    for (const index of indices) {\n      const label = yData[index] ?? 0;\n      const idx = labelIndex.get(label);\n      if (idx !== undefined) counts[idx] = (counts[idx] ?? 0) + 1;\n    }\n\n    const invN = 1 / indices.length;\n    return counts.map((c) => c * invN);\n  }\n\n  private findBestSplit(\n    XData: number[][],\n    yData: number[],\n    indices: number[]\n  ): {\n    featureIndex: number;\n    threshold: number;\n    leftIndices: number[];\n    rightIndices: number[];\n  } {\n    let bestGini = Infinity;\n    let bestFeature = 0;\n    let bestThreshold = 0;\n    let bestLeft: number[] = [];\n    let bestRight: number[] = [];\n\n    const nFeatures = XData[0]?.length ?? 0;\n    let featureIndices = Array.from({ length: nFeatures }, (_, i) => i);\n\n    if (this.maxFeatures !== undefined && this.maxFeatures < nFeatures) {\n      const rng = this.getRng();\n      // Fisher-Yates shuffle partial\n      for (let i = 0; i < this.maxFeatures; i++) {\n        const j = i + Math.floor(rng() * (nFeatures - i));\n        const temp = featureIndices[i];\n        if (temp !== undefined) {\n          const swapVal = featureIndices[j];\n          if (swapVal !== undefined) {\n            featureIndices[i] = swapVal;\n            featureIndices[j] = temp;\n          }\n        }\n      }\n      featureIndices = featureIndices.slice(0, this.maxFeatures);\n    }\n\n    const n = indices.length;\n    // Pre-calculate total class counts\n    const totalCounts = new Map<number, number>();\n    for (const i of indices) {\n      const label = yData[i] ?? 0;\n      totalCounts.set(label, (totalCounts.get(label) ?? 0) + 1);\n    }\n\n    for (const f of featureIndices) {\n      // Sort indices by feature value\n      // Create a copy to sort\n      const sortedIndices = [...indices].sort(\n        (a, b) => (XData[a]?.[f] ?? 0) - (XData[b]?.[f] ?? 0)\n      );\n\n      const leftCounts = new Map<number, number>();\n      const rightCounts = new Map<number, number>(totalCounts);\n      let leftSize = 0;\n      let rightSize = n;\n\n      for (let i = 0; i < n - 1; i++) {\n        const idx = sortedIndices[i];\n        if (idx === undefined) continue;\n        const label = yData[idx] ?? 0;\n        const val = XData[idx]?.[f] ?? 0;\n        const nextIdx = sortedIndices[i + 1];\n        if (nextIdx === undefined) continue;\n        const nextVal = XData[nextIdx]?.[f] ?? 0;\n\n        // Move from Right to Left\n        const currentRight = rightCounts.get(label) ?? 0;\n        if (currentRight <= 1) rightCounts.delete(label);\n        else rightCounts.set(label, currentRight - 1);\n        rightSize--;\n\n        leftCounts.set(label, (leftCounts.get(label) ?? 0) + 1);\n        leftSize++;\n\n        if (val === nextVal) continue; // Cannot split between same values\n\n        if (leftSize < this.minSamplesLeaf || rightSize < this.minSamplesLeaf) continue;\n\n        // Calculate weighted Gini\n        const leftGini = this.giniFromCounts(leftCounts, leftSize);\n        const rightGini = this.giniFromCounts(rightCounts, rightSize);\n        const weightedGini = (leftSize * leftGini + rightSize * rightGini) / n;\n\n        if (weightedGini < bestGini) {\n          bestGini = weightedGini;\n          bestFeature = f;\n          bestThreshold = (val + nextVal) / 2;\n          bestLeft = sortedIndices.slice(0, i + 1);\n          bestRight = sortedIndices.slice(i + 1);\n        }\n      }\n    }\n\n    return {\n      featureIndex: bestFeature,\n      threshold: bestThreshold,\n      leftIndices: bestLeft,\n      rightIndices: bestRight,\n    };\n  }\n\n  private giniFromCounts(counts: Map<number, number>, n: number): number {\n    let impurity = 1.0;\n    for (const count of counts.values()) {\n      const p = count / n;\n      impurity -= p * p;\n    }\n    return impurity;\n  }\n\n  /**\n   * Predict class labels for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted class labels of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted || !this.tree) {\n      throw new NotFittedError(\"DecisionTreeClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"DecisionTreeClassifier\");\n\n    const predictions: number[] = [];\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    for (let i = 0; i < nSamples; i++) {\n      const sample: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        sample.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      predictions.push(this.predictSample(sample, this.tree));\n    }\n\n    return tensor(predictions, { dtype: \"int32\" });\n  }\n\n  private predictSample(sample: number[], node: TreeNode): number {\n    let current = node;\n    while (!current.isLeaf) {\n      const featureValue = sample[current.featureIndex ?? 0] ?? 0;\n      if (featureValue <= (current.threshold ?? 0)) {\n        if (!current.left)\n          throw new DeepboxError(\"Corrupted tree: Internal node missing left child\");\n        current = current.left;\n      } else {\n        if (!current.right)\n          throw new DeepboxError(\"Corrupted tree: Internal node missing right child\");\n        current = current.right;\n      }\n    }\n    return current.prediction ?? 0;\n  }\n\n  /**\n   * Predict class probabilities for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Class probability matrix of shape (n_samples, n_classes)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predictProba(X: Tensor): Tensor {\n    if (!this.fitted || !this.tree || !this.classLabels) {\n      throw new NotFittedError(\"DecisionTreeClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"DecisionTreeClassifier\");\n    assertContiguous(X, \"X\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const nClasses = this.classLabels.length;\n\n    const proba: number[][] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const sample: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        sample.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n\n      const leaf = this.predictLeaf(sample, this.tree);\n      const row = leaf.classProbabilities\n        ? [...leaf.classProbabilities]\n        : new Array(nClasses).fill(0);\n      proba.push(row);\n    }\n\n    return tensor(proba);\n  }\n\n  private predictLeaf(sample: number[], node: TreeNode): TreeNode {\n    if (node.isLeaf) {\n      return node;\n    }\n\n    const featureValue = sample[node.featureIndex ?? 0] ?? 0;\n    if (featureValue <= (node.threshold ?? 0)) {\n      if (!node.left) {\n        return node;\n      }\n      return this.predictLeaf(sample, node.left);\n    }\n    if (!node.right) {\n      return node;\n    }\n    return this.predictLeaf(sample, node.right);\n  }\n\n  /**\n   * Return the mean accuracy on the given test data and labels.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True labels of shape (n_samples,)\n   * @returns Accuracy score in range [0, 1]\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n    let correct = 0;\n    for (let i = 0; i < y.size; i++) {\n      if (Number(predictions.data[predictions.offset + i]) === Number(y.data[y.offset + i])) {\n        correct++;\n      }\n    }\n    return correct / y.size;\n  }\n\n  /**\n   * Get the unique class labels discovered during fitting.\n   *\n   * @returns Tensor of class labels or undefined if not fitted\n   */\n  get classes(): Tensor | undefined {\n    if (!this.fitted || !this.classLabels) {\n      return undefined;\n    }\n    return tensor(this.classLabels, { dtype: \"int32\" });\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      maxDepth: this.maxDepth,\n      minSamplesSplit: this.minSamplesSplit,\n      minSamplesLeaf: this.minSamplesLeaf,\n      maxFeatures: this.maxFeatures,\n      randomState: this.randomState,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\n      \"DecisionTreeClassifier does not support setParams after construction\"\n    );\n  }\n}\n\n/**\n * Decision Tree Regressor.\n *\n * Uses MSE reduction to find optimal splits for regression tasks.\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeRegressor.html | scikit-learn DecisionTreeRegressor}\n */\nexport class DecisionTreeRegressor implements Regressor {\n  private maxDepth: number;\n  private minSamplesSplit: number;\n  private minSamplesLeaf: number;\n  private maxFeatures: number | undefined;\n  private randomState: number | undefined;\n\n  private tree?: TreeNode;\n  private nFeatures?: number;\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly maxDepth?: number;\n      readonly minSamplesSplit?: number;\n      readonly minSamplesLeaf?: number;\n      readonly maxFeatures?: number;\n      readonly randomState?: number;\n    } = {}\n  ) {\n    this.maxDepth = options.maxDepth ?? 10;\n    this.minSamplesSplit = options.minSamplesSplit ?? 2;\n    this.minSamplesLeaf = options.minSamplesLeaf ?? 1;\n    if (options.maxFeatures !== undefined) {\n      this.maxFeatures = options.maxFeatures;\n    }\n    if (options.randomState !== undefined) {\n      this.randomState = options.randomState;\n    }\n\n    if (this.randomState !== undefined && !Number.isFinite(this.randomState)) {\n      throw new InvalidParameterError(\n        `randomState must be a finite number; received ${String(this.randomState)}`,\n        \"randomState\",\n        this.randomState\n      );\n    }\n\n    if (!Number.isInteger(this.maxDepth) || this.maxDepth < 1) {\n      throw new InvalidParameterError(\n        `maxDepth must be an integer >= 1; received ${this.maxDepth}`,\n        \"maxDepth\",\n        this.maxDepth\n      );\n    }\n    if (!Number.isInteger(this.minSamplesSplit) || this.minSamplesSplit < 2) {\n      throw new InvalidParameterError(\n        `minSamplesSplit must be an integer >= 2; received ${this.minSamplesSplit}`,\n        \"minSamplesSplit\",\n        this.minSamplesSplit\n      );\n    }\n    if (!Number.isInteger(this.minSamplesLeaf) || this.minSamplesLeaf < 1) {\n      throw new InvalidParameterError(\n        `minSamplesLeaf must be an integer >= 1; received ${this.minSamplesLeaf}`,\n        \"minSamplesLeaf\",\n        this.minSamplesLeaf\n      );\n    }\n    if (\n      this.maxFeatures !== undefined &&\n      (!Number.isInteger(this.maxFeatures) || this.maxFeatures < 1)\n    ) {\n      throw new InvalidParameterError(\n        `maxFeatures must be an integer >= 1; received ${this.maxFeatures}`,\n        \"maxFeatures\",\n        this.maxFeatures\n      );\n    }\n  }\n\n  private getRng(): () => number {\n    if (this.randomState === undefined) {\n      return Math.random;\n    }\n    let seed = this.randomState;\n    return () => {\n      seed = (seed * 1664525 + 1013904223) >>> 0;\n      return seed / 4294967296;\n    };\n  }\n\n  /**\n   * Build a decision tree regressor from the training set (X, y).\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n    assertContiguous(X, \"X\");\n    assertContiguous(y, \"y\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    this.nFeatures = nFeatures;\n\n    const XData: number[][] = [];\n    const yData: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      XData.push(row);\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    const indices = Array.from({ length: nSamples }, (_, i) => i);\n    this.tree = this.buildTree(XData, yData, indices, 0);\n    this.fitted = true;\n\n    return this;\n  }\n\n  private buildTree(\n    XData: number[][],\n    yData: number[],\n    indices: number[],\n    depth: number\n  ): TreeNode {\n    const n = indices.length;\n    if (n === 0) {\n      throw new DataValidationError(\"Cannot build a decision tree from an empty dataset\");\n    }\n\n    if (depth >= this.maxDepth || n < this.minSamplesSplit || n < this.minSamplesLeaf) {\n      return { isLeaf: true, prediction: this.getMean(yData, indices) };\n    }\n\n    const { featureIndex, threshold, leftIndices, rightIndices } = this.findBestSplit(\n      XData,\n      yData,\n      indices\n    );\n\n    if (leftIndices.length === 0 || rightIndices.length === 0) {\n      return { isLeaf: true, prediction: this.getMean(yData, indices) };\n    }\n\n    const left = this.buildTree(XData, yData, leftIndices, depth + 1);\n    const right = this.buildTree(XData, yData, rightIndices, depth + 1);\n\n    return {\n      isLeaf: false,\n      featureIndex,\n      threshold,\n      left,\n      right,\n    };\n  }\n\n  private getMean(yData: number[], indices: number[]): number {\n    let sum = 0;\n    for (const i of indices) {\n      sum += yData[i] ?? 0;\n    }\n    return sum / indices.length;\n  }\n\n  private findBestSplit(\n    XData: number[][],\n    yData: number[],\n    indices: number[]\n  ): {\n    featureIndex: number;\n    threshold: number;\n    leftIndices: number[];\n    rightIndices: number[];\n  } {\n    let bestScore = -Infinity; // We maximize the proxy score\n    let bestFeature = 0;\n    let bestThreshold = 0;\n    let bestLeft: number[] = [];\n    let bestRight: number[] = [];\n\n    const nFeatures = XData[0]?.length ?? 0;\n    let featureIndices = Array.from({ length: nFeatures }, (_, i) => i);\n\n    if (this.maxFeatures !== undefined && this.maxFeatures < nFeatures) {\n      const rng = this.getRng();\n      for (let i = 0; i < this.maxFeatures; i++) {\n        const j = i + Math.floor(rng() * (nFeatures - i));\n        const a = featureIndices[i];\n        const b = featureIndices[j];\n        if (a === undefined || b === undefined) {\n          throw new DeepboxError(`Internal error: featureIndices out of bounds: i=${i}, j=${j}`);\n        }\n        featureIndices[i] = b;\n        featureIndices[j] = a;\n      }\n      featureIndices = featureIndices.slice(0, this.maxFeatures);\n    }\n\n    const n = indices.length;\n    let totalSum = 0;\n\n    for (const i of indices) {\n      const yVal = yData[i] ?? 0;\n      totalSum += yVal;\n    }\n\n    for (const f of featureIndices) {\n      // Sort indices by feature value\n      const sortedIndices = [...indices].sort(\n        (a, b) => (XData[a]?.[f] ?? 0) - (XData[b]?.[f] ?? 0)\n      );\n\n      let leftSum = 0;\n      let leftCnt = 0;\n      let rightSum = totalSum;\n      let rightCnt = n;\n\n      for (let i = 0; i < n - 1; i++) {\n        const idx = sortedIndices[i];\n        if (idx === undefined) continue;\n        const val = XData[idx]?.[f] ?? 0;\n        const nextIdx = sortedIndices[i + 1];\n        if (nextIdx === undefined) continue;\n        const nextVal = XData[nextIdx]?.[f] ?? 0;\n        const yVal = yData[idx] ?? 0;\n\n        // Move from Right to Left\n        leftSum += yVal;\n        leftCnt++;\n        rightSum -= yVal;\n        rightCnt--;\n\n        if (val === nextVal) continue; // Cannot split between same values\n\n        if (leftCnt < this.minSamplesLeaf || rightCnt < this.minSamplesLeaf) continue;\n\n        // Proxy score to maximize: (SumL^2 / nL) + (SumR^2 / nR)\n        // This is equivalent to minimizing weighted MSE\n        const score = (leftSum * leftSum) / leftCnt + (rightSum * rightSum) / rightCnt;\n\n        if (score > bestScore) {\n          bestScore = score;\n          bestFeature = f;\n          bestThreshold = (val + nextVal) / 2;\n          bestLeft = sortedIndices.slice(0, i + 1);\n          bestRight = sortedIndices.slice(i + 1);\n        }\n      }\n    }\n\n    // If no split found (e.g. pure node or all features constant), return empty\n    if (bestScore === -Infinity) {\n      return {\n        featureIndex: 0,\n        threshold: 0,\n        leftIndices: [],\n        rightIndices: [],\n      };\n    }\n\n    return {\n      featureIndex: bestFeature,\n      threshold: bestThreshold,\n      leftIndices: bestLeft,\n      rightIndices: bestRight,\n    };\n  }\n\n  /**\n   * Predict target values for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted || !this.tree) {\n      throw new NotFittedError(\"DecisionTreeRegressor must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"DecisionTreeRegressor\");\n    assertContiguous(X, \"X\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    const predictions: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const sample: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        sample.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      predictions.push(this.predictSample(sample, this.tree));\n    }\n\n    return tensor(predictions);\n  }\n\n  private predictSample(sample: number[], node: TreeNode): number {\n    if (node.isLeaf) {\n      return node.prediction ?? 0;\n    }\n\n    const featureValue = sample[node.featureIndex ?? 0] ?? 0;\n    if (featureValue <= (node.threshold ?? 0)) {\n      if (!node.left) throw new DeepboxError(\"Corrupted tree: Internal node missing left child\");\n      return this.predictSample(sample, node.left);\n    } else {\n      if (!node.right) throw new DeepboxError(\"Corrupted tree: Internal node missing right child\");\n      return this.predictSample(sample, node.right);\n    }\n  }\n\n  /**\n   * Return the R² score on the given test data and target values.\n   *\n   * R² = 1 - SS_res / SS_tot.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True target values of shape (n_samples,)\n   * @returns R² score (best possible is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n\n    // Calculate R² score\n    let ssRes = 0;\n    let ssTot = 0;\n    let yMean = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      yMean += Number(y.data[y.offset + i]);\n    }\n    yMean /= y.size;\n\n    for (let i = 0; i < y.size; i++) {\n      const yTrue = Number(y.data[y.offset + i]);\n      const yPred = Number(predictions.data[predictions.offset + i]);\n      ssRes += (yTrue - yPred) ** 2;\n      ssTot += (yTrue - yMean) ** 2;\n    }\n\n    return ssTot === 0 ? (ssRes === 0 ? 1.0 : 0.0) : 1 - ssRes / ssTot;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      maxDepth: this.maxDepth,\n      minSamplesSplit: this.minSamplesSplit,\n      minSamplesLeaf: this.minSamplesLeaf,\n      maxFeatures: this.maxFeatures,\n      randomState: this.randomState,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\n      \"DecisionTreeRegressor does not support setParams after construction\"\n    );\n  }\n}\n","import {\n  DataValidationError,\n  InvalidParameterError,\n  NotFittedError,\n  NotImplementedError,\n  ShapeError,\n} from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Classifier, Regressor } from \"../base\";\nimport { DecisionTreeRegressor } from \"../tree/DecisionTree\";\n\n/**\n * Gradient Boosting Regressor.\n *\n * Builds an additive model in a forward stage-wise fashion using\n * regression trees as weak learners. Optimizes squared error loss.\n *\n * **Algorithm**: Gradient Boosting with regression trees\n * - Stage-wise additive modeling\n * - Uses gradient of squared loss (residuals)\n *\n * @example\n * ```ts\n * import { GradientBoostingRegressor } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1], [2], [3], [4], [5]]);\n * const y = tensor([1.2, 2.1, 2.9, 4.0, 5.1]);\n *\n * const gbr = new GradientBoostingRegressor({ nEstimators: 100 });\n * gbr.fit(X, y);\n * const predictions = gbr.predict(X);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingRegressor.html | scikit-learn GradientBoostingRegressor}\n */\nexport class GradientBoostingRegressor implements Regressor {\n  /** Number of boosting stages (trees) */\n  private nEstimators: number;\n\n  /** Learning rate shrinks the contribution of each tree */\n  private learningRate: number;\n\n  /** Maximum depth of individual regression trees */\n  private maxDepth: number;\n\n  /** Minimum samples required to split */\n  private minSamplesSplit: number;\n\n  /** Array of weak learners (regression trees) */\n  private estimators: DecisionTreeRegressor[] = [];\n\n  /** Initial prediction (mean of targets) */\n  private initPrediction = 0;\n\n  /** Number of features */\n  private nFeatures = 0;\n\n  /** Whether the model has been fitted */\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly nEstimators?: number;\n      readonly learningRate?: number;\n      readonly maxDepth?: number;\n      readonly minSamplesSplit?: number;\n    } = {}\n  ) {\n    this.nEstimators = options.nEstimators ?? 100;\n    this.learningRate = options.learningRate ?? 0.1;\n    this.maxDepth = options.maxDepth ?? 3;\n    this.minSamplesSplit = options.minSamplesSplit ?? 2;\n\n    if (!Number.isInteger(this.nEstimators) || this.nEstimators <= 0) {\n      throw new InvalidParameterError(\n        \"nEstimators must be a positive integer\",\n        \"nEstimators\",\n        this.nEstimators\n      );\n    }\n    if (!Number.isFinite(this.learningRate) || this.learningRate <= 0) {\n      throw new InvalidParameterError(\n        \"learningRate must be positive\",\n        \"learningRate\",\n        this.learningRate\n      );\n    }\n    if (!Number.isInteger(this.maxDepth) || this.maxDepth < 1) {\n      throw new InvalidParameterError(\n        \"maxDepth must be an integer >= 1\",\n        \"maxDepth\",\n        this.maxDepth\n      );\n    }\n    if (!Number.isInteger(this.minSamplesSplit) || this.minSamplesSplit < 2) {\n      throw new InvalidParameterError(\n        \"minSamplesSplit must be an integer >= 2\",\n        \"minSamplesSplit\",\n        this.minSamplesSplit\n      );\n    }\n  }\n\n  /**\n   * Fit the gradient boosting regressor on training data.\n   *\n   * Builds an additive model by sequentially fitting regression trees\n   * to the negative gradient (residuals) of the loss function.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    this.nFeatures = nFeatures;\n\n    const yData: number[] = [];\n    for (let i = 0; i < nSamples; i++) {\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    // Initialize with mean prediction (F0)\n    this.initPrediction = yData.reduce((sum, val) => sum + val, 0) / nSamples;\n\n    // Current predictions\n    const predictions = new Array<number>(nSamples).fill(this.initPrediction);\n\n    // Build ensemble\n    this.estimators = [];\n\n    for (let m = 0; m < this.nEstimators; m++) {\n      // Compute residuals (negative gradient of squared loss)\n      const residuals: number[] = [];\n      for (let i = 0; i < nSamples; i++) {\n        residuals.push((yData[i] ?? 0) - (predictions[i] ?? 0));\n      }\n\n      // Fit a regression tree to residuals\n      const tree = new DecisionTreeRegressor({\n        maxDepth: this.maxDepth,\n        minSamplesSplit: this.minSamplesSplit,\n        minSamplesLeaf: 1,\n      });\n      tree.fit(X, tensor(residuals));\n      this.estimators.push(tree);\n\n      // Update predictions\n      const treePred = tree.predict(X);\n      for (let i = 0; i < nSamples; i++) {\n        predictions[i] =\n          (predictions[i] ?? 0) + this.learningRate * Number(treePred.data[treePred.offset + i]);\n      }\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Predict target values for samples in X.\n   *\n   * Aggregates the initial prediction and the scaled contributions of all trees.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"GradientBoostingRegressor must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"GradientBoostingRegressor\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const predictions = new Array<number>(nSamples).fill(this.initPrediction);\n\n    for (const tree of this.estimators) {\n      const treePred = tree.predict(X);\n      for (let i = 0; i < nSamples; i++) {\n        predictions[i] =\n          (predictions[i] ?? 0) + this.learningRate * Number(treePred.data[treePred.offset + i]);\n      }\n    }\n\n    return tensor(predictions);\n  }\n\n  /**\n   * Return the R² score on the given test data and target values.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True target values of shape (n_samples,)\n   * @returns R² score (best possible is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n\n    let ssRes = 0;\n    let ssTot = 0;\n    let yMean = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      yMean += Number(y.data[y.offset + i]);\n    }\n    yMean /= y.size;\n\n    for (let i = 0; i < y.size; i++) {\n      const yTrue = Number(y.data[y.offset + i]);\n      const yPred = Number(predictions.data[predictions.offset + i]);\n      ssRes += (yTrue - yPred) ** 2;\n      ssTot += (yTrue - yMean) ** 2;\n    }\n\n    return ssTot === 0 ? (ssRes === 0 ? 1.0 : 0.0) : 1 - ssRes / ssTot;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nEstimators: this.nEstimators,\n      learningRate: this.learningRate,\n      maxDepth: this.maxDepth,\n      minSamplesSplit: this.minSamplesSplit,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\n      \"GradientBoostingRegressor does not support setParams after construction\"\n    );\n  }\n}\n\n/**\n * Gradient Boosting Classifier.\n *\n * Uses gradient boosting with shallow regression trees for binary classification.\n * Optimizes log loss (cross-entropy) using sigmoid function.\n *\n * @example\n * ```ts\n * import { GradientBoostingClassifier } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1, 2], [2, 3], [3, 1], [4, 2]]);\n * const y = tensor([0, 0, 1, 1]);\n *\n * const gbc = new GradientBoostingClassifier({ nEstimators: 100 });\n * gbc.fit(X, y);\n * const predictions = gbc.predict(X);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html | scikit-learn GradientBoostingClassifier}\n */\nexport class GradientBoostingClassifier implements Classifier {\n  /** Number of boosting stages */\n  private nEstimators: number;\n\n  /** Learning rate */\n  private learningRate: number;\n\n  /** Maximum depth */\n  private maxDepth: number;\n\n  /** Minimum samples to split */\n  private minSamplesSplit: number;\n\n  /** Array of weak learners */\n  private estimators: DecisionTreeRegressor[] = [];\n\n  /** Initial log-odds prediction */\n  private initPrediction = 0;\n\n  /** Number of features */\n  private nFeatures = 0;\n\n  /** Unique class labels */\n  private classLabels: number[] = [];\n\n  /** Whether fitted */\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly nEstimators?: number;\n      readonly learningRate?: number;\n      readonly maxDepth?: number;\n      readonly minSamplesSplit?: number;\n    } = {}\n  ) {\n    this.nEstimators = options.nEstimators ?? 100;\n    this.learningRate = options.learningRate ?? 0.1;\n    this.maxDepth = options.maxDepth ?? 3;\n    this.minSamplesSplit = options.minSamplesSplit ?? 2;\n\n    if (!Number.isInteger(this.nEstimators) || this.nEstimators <= 0) {\n      throw new InvalidParameterError(\n        \"nEstimators must be a positive integer\",\n        \"nEstimators\",\n        this.nEstimators\n      );\n    }\n    if (!Number.isFinite(this.learningRate) || this.learningRate <= 0) {\n      throw new InvalidParameterError(\n        \"learningRate must be positive\",\n        \"learningRate\",\n        this.learningRate\n      );\n    }\n    if (!Number.isInteger(this.maxDepth) || this.maxDepth < 1) {\n      throw new InvalidParameterError(\n        \"maxDepth must be an integer >= 1\",\n        \"maxDepth\",\n        this.maxDepth\n      );\n    }\n    if (!Number.isInteger(this.minSamplesSplit) || this.minSamplesSplit < 2) {\n      throw new InvalidParameterError(\n        \"minSamplesSplit must be an integer >= 2\",\n        \"minSamplesSplit\",\n        this.minSamplesSplit\n      );\n    }\n  }\n\n  /**\n   * Fit the gradient boosting classifier on training data.\n   *\n   * Builds an additive model by sequentially fitting regression trees\n   * to the pseudo-residuals (gradient of log loss).\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target class labels of shape (n_samples,). Must contain exactly 2 classes.\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {InvalidParameterError} If y does not contain exactly 2 classes\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    this.nFeatures = nFeatures;\n\n    const yData: number[] = [];\n    for (let i = 0; i < nSamples; i++) {\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    // Get unique classes\n    this.classLabels = [...new Set(yData)].sort((a, b) => a - b);\n    if (this.classLabels.length !== 2) {\n      throw new InvalidParameterError(\n        \"GradientBoostingClassifier requires exactly 2 classes\",\n        \"y\",\n        this.classLabels.length\n      );\n    }\n\n    // Map to {0, 1}\n    const yBinary = yData.map((label) => (label === this.classLabels[0] ? 0 : 1));\n\n    // Initialize with log-odds\n    const posCount = yBinary.filter((v) => v === 1).length;\n    const negCount = nSamples - posCount;\n    this.initPrediction = Math.log((posCount + 1) / (negCount + 1)); // Add smoothing\n\n    // Current raw scores (log-odds)\n    const rawScores = new Array<number>(nSamples).fill(this.initPrediction);\n\n    // Build ensemble\n    this.estimators = [];\n\n    for (let m = 0; m < this.nEstimators; m++) {\n      // Compute pseudo-residuals (gradient of log loss)\n      const residuals: number[] = [];\n      for (let i = 0; i < nSamples; i++) {\n        const prob = 1 / (1 + Math.exp(-(rawScores[i] ?? 0))); // Sigmoid\n        const y_i = yBinary[i] ?? 0;\n        residuals.push(y_i - prob);\n      }\n\n      // Fit regression tree to residuals\n      const tree = new DecisionTreeRegressor({\n        maxDepth: this.maxDepth,\n        minSamplesSplit: this.minSamplesSplit,\n        minSamplesLeaf: 1,\n      });\n      tree.fit(X, tensor(residuals));\n      this.estimators.push(tree);\n\n      // Update raw scores\n      const treePred = tree.predict(X);\n      for (let i = 0; i < nSamples; i++) {\n        rawScores[i] =\n          (rawScores[i] ?? 0) + this.learningRate * Number(treePred.data[treePred.offset + i]);\n      }\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Predict class labels for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted class labels of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"GradientBoostingClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"GradientBoostingClassifier\");\n\n    const nSamples = X.shape[0] ?? 0;\n\n    const rawScores = new Array<number>(nSamples).fill(this.initPrediction);\n    for (const tree of this.estimators) {\n      const treePred = tree.predict(X);\n      for (let i = 0; i < nSamples; i++) {\n        rawScores[i] =\n          (rawScores[i] ?? 0) + this.learningRate * Number(treePred.data[treePred.offset + i]);\n      }\n    }\n\n    const predictions: number[] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const prob = 1 / (1 + Math.exp(-(rawScores[i] ?? 0)));\n      const predictedClass = prob >= 0.5 ? this.classLabels[1] : this.classLabels[0];\n      predictions.push(predictedClass ?? 0);\n    }\n\n    return tensor(predictions, { dtype: \"int32\" });\n  }\n\n  /**\n   * Predict class probabilities for samples in X.\n   *\n   * Returns a matrix of shape (n_samples, 2) where columns are\n   * [P(class_0), P(class_1)].\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Class probability matrix of shape (n_samples, 2)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predictProba(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"GradientBoostingClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"GradientBoostingClassifier\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const rawScores = new Array<number>(nSamples).fill(this.initPrediction);\n    for (const tree of this.estimators) {\n      const treePred = tree.predict(X);\n      for (let i = 0; i < nSamples; i++) {\n        rawScores[i] =\n          (rawScores[i] ?? 0) + this.learningRate * Number(treePred.data[treePred.offset + i]);\n      }\n    }\n\n    const proba: number[][] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const prob = 1 / (1 + Math.exp(-(rawScores[i] ?? 0)));\n      proba.push([1 - prob, prob]);\n    }\n\n    return tensor(proba);\n  }\n\n  /**\n   * Return the mean accuracy on the given test data and labels.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True labels of shape (n_samples,)\n   * @returns Accuracy score in range [0, 1]\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n    let correct = 0;\n    for (let i = 0; i < y.size; i++) {\n      if (Number(predictions.data[predictions.offset + i]) === Number(y.data[y.offset + i])) {\n        correct++;\n      }\n    }\n    return correct / y.size;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nEstimators: this.nEstimators,\n      learningRate: this.learningRate,\n      maxDepth: this.maxDepth,\n      minSamplesSplit: this.minSamplesSplit,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\n      \"GradientBoostingClassifier does not support setParams after construction\"\n    );\n  }\n}\n","import { DataValidationError, InvalidParameterError, NotFittedError, ShapeError } from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Regressor } from \"../base\";\n\n/**\n * Lasso Regression (L1 Regularized Linear Regression).\n *\n * Lasso performs both regularization and feature selection by adding\n * an L1 penalty that can drive coefficients exactly to zero.\n *\n * @example\n * ```ts\n * import { Lasso } from 'deepbox/ml';\n *\n * const model = new Lasso({ alpha: 0.1, maxIter: 1000 });\n * model.fit(X_train, y_train);\n *\n * // Many coefficients will be exactly 0\n * console.log(model.coef);\n *\n * const predictions = model.predict(X_test);\n * ```\n *\n * @category Linear Models\n * @implements {Regressor}\n */\nexport class Lasso implements Regressor {\n  /** Configuration options for the Lasso regression model */\n  private options: {\n    alpha?: number;\n    fitIntercept?: boolean;\n    normalize?: boolean;\n    maxIter?: number;\n    tol?: number;\n    warmStart?: boolean;\n    positive?: boolean;\n    selection?: \"cyclic\" | \"random\";\n    randomState?: number;\n  };\n\n  /** Model coefficients (weights) after fitting - shape (n_features,) */\n  private coef_?: Tensor;\n\n  /** Intercept (bias) term after fitting */\n  private intercept_ = 0;\n\n  /** Number of features seen during fit - used for validation */\n  private nFeaturesIn_?: number;\n\n  /** Number of iterations run by coordinate descent */\n  private nIter_: number | undefined;\n\n  /** Whether the model has been fitted to data */\n  private fitted = false;\n\n  /**\n   * Create a new Lasso Regression model.\n   *\n   * @param options - Configuration options\n   * @param options.alpha - Regularization strength (default: 1.0). Must be >= 0. Controls sparsity of solution.\n   * @param options.fitIntercept - Whether to calculate the intercept (default: true)\n   * @param options.normalize - Whether to normalize features before regression (default: false)\n   * @param options.maxIter - Maximum iterations for coordinate descent (default: 1000)\n   * @param options.tol - Tolerance for convergence (default: 1e-4). Smaller = more precise but slower.\n   * @param options.warmStart - Whether to reuse previous solution as initialization (default: false)\n   * @param options.positive - Whether to force coefficients to be positive (default: false)\n   * @param options.selection - Coordinate selection: 'cyclic' (default) or 'random'\n   */\n  constructor(\n    options: {\n      readonly alpha?: number;\n      readonly fitIntercept?: boolean;\n      readonly normalize?: boolean;\n      readonly maxIter?: number;\n      readonly tol?: number;\n      readonly warmStart?: boolean;\n      readonly positive?: boolean;\n      readonly selection?: \"cyclic\" | \"random\";\n      readonly randomState?: number;\n    } = {}\n  ) {\n    this.options = { ...options };\n    if (this.options.randomState !== undefined && !Number.isFinite(this.options.randomState)) {\n      throw new InvalidParameterError(\n        `randomState must be a finite number; received ${String(this.options.randomState)}`,\n        \"randomState\",\n        this.options.randomState\n      );\n    }\n  }\n\n  private createRNG(): () => number {\n    if (this.options.randomState !== undefined) {\n      let seed = this.options.randomState;\n      return () => {\n        seed = (seed * 9301 + 49297) % 233280;\n        return seed / 233280;\n      };\n    }\n    return Math.random;\n  }\n\n  /**\n   * Fit Lasso regression model using Coordinate Descent.\n   *\n   * Solves the L1-regularized least squares problem:\n   * minimize (1/(2*n)) ||y - Xw||² + α||w||₁\n   *\n   * **Algorithm**: Coordinate Descent with Soft Thresholding\n   * 1. Initialize coefficients (warm start if enabled)\n   * 2. For each iteration:\n   *    - For each feature (cyclic or random order):\n   *      - Compute residual correlation\n   *      - Apply soft thresholding operator\n   *      - Update predictions incrementally\n   * 3. Check convergence based on coefficient changes\n   *\n   * **Time Complexity**: O(k * n * p) where k = iterations, n = samples, p = features\n   * **Space Complexity**: O(n + p)\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,)\n   * @returns this - The fitted estimator for method chaining\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {DataValidationError} If X or y are empty\n   * @throws {InvalidParameterError} If alpha < 0\n   */\n  fit(X: Tensor, y: Tensor): this {\n    // Validate inputs (dimensions, empty data, NaN/Inf)\n    validateFitInputs(X, y);\n    this.nIter_ = undefined;\n\n    // Extract and validate regularization parameter\n    const alpha = this.options.alpha ?? 1.0;\n    if (!(alpha >= 0)) {\n      throw new InvalidParameterError(`alpha must be >= 0; received ${alpha}`, \"alpha\", alpha);\n    }\n\n    // Extract optimization parameters\n    const maxIter = this.options.maxIter ?? 1000;\n    const tol = this.options.tol ?? 1e-4;\n    const fitIntercept = this.options.fitIntercept ?? true;\n    const normalize = this.options.normalize ?? false;\n    const positive = this.options.positive ?? false;\n    const selection = this.options.selection ?? \"cyclic\";\n    const rng = this.createRNG();\n\n    // Extract dimensions: m = number of samples, n = number of features\n    const m = X.shape[0] ?? 0;\n    const n = X.shape[1] ?? 0;\n\n    // Store number of features for prediction validation\n    this.nFeaturesIn_ = n;\n\n    // Compute means for centering (if fitIntercept is true)\n    // Centering is crucial for proper intercept calculation and numerical stability\n    let yMean = 0;\n    const xMean = new Array<number>(n).fill(0);\n\n    if (fitIntercept) {\n      // Compute sum of y values\n      for (let i = 0; i < m; i++) {\n        yMean += Number(y.data[y.offset + i] ?? 0);\n      }\n\n      // Compute sum of each feature column\n      for (let i = 0; i < m; i++) {\n        const rowBase = X.offset + i * n;\n        for (let j = 0; j < n; j++) {\n          xMean[j] = (xMean[j] ?? 0) + Number(X.data[rowBase + j] ?? 0);\n        }\n      }\n\n      // Convert sums to means\n      const invM = m === 0 ? 0 : 1 / m;\n      yMean *= invM;\n      for (let j = 0; j < n; j++) {\n        xMean[j] = (xMean[j] ?? 0) * invM;\n      }\n    }\n\n    let xScale: number[] | undefined;\n    if (normalize) {\n      xScale = new Array<number>(n).fill(0);\n      for (let i = 0; i < m; i++) {\n        const rowBase = X.offset + i * n;\n        for (let j = 0; j < n; j++) {\n          const centered = Number(X.data[rowBase + j] ?? 0) - (fitIntercept ? (xMean[j] ?? 0) : 0);\n          xScale[j] = (xScale[j] ?? 0) + centered * centered;\n        }\n      }\n      for (let j = 0; j < n; j++) {\n        xScale[j] = Math.sqrt(xScale[j] ?? 0);\n      }\n    }\n\n    const getX = (sampleIndex: number, featureIndex: number): number => {\n      const raw = Number(X.data[X.offset + sampleIndex * n + featureIndex] ?? 0);\n      const centered = raw - (fitIntercept ? (xMean[featureIndex] ?? 0) : 0);\n      if (normalize && xScale) {\n        const s = xScale[featureIndex] ?? 0;\n        return s === 0 ? 0 : centered / s;\n      }\n      return centered;\n    };\n\n    // Precompute column squared norms of centered X: (1/m) * Σᵢ xᵢⱼ²\n    // This is used in the coordinate descent update formula\n    // Caching these norms significantly improves performance (O(n) vs O(nm) per iteration)\n    const colNorm2 = new Array<number>(n).fill(0);\n    for (let j = 0; j < n; j++) {\n      let s = 0;\n      for (let i = 0; i < m; i++) {\n        const xij = getX(i, j);\n        s += xij * xij;\n      }\n      // Normalize by number of samples\n      colNorm2[j] = m === 0 ? 0 : s / m;\n    }\n\n    // Initialize coefficients\n    // If warm_start is enabled and we have previous coefficients, reuse them\n    // Otherwise initialize to zeros\n    const w = new Array<number>(n).fill(0);\n    if (this.options.warmStart && this.coef_ && this.coef_.ndim === 1 && this.coef_.size === n) {\n      // Copy previous coefficients for warm start\n      for (let j = 0; j < n; j++) {\n        w[j] = Number(this.coef_.data[this.coef_.offset + j] ?? 0);\n      }\n    }\n\n    // Maintain current predictions in centered space: ŷ = X_centered @ w\n    // We update this incrementally during coordinate descent for efficiency\n    // This avoids recomputing all predictions from scratch each iteration\n    const yHat = new Array<number>(m).fill(0);\n    for (let i = 0; i < m; i++) {\n      let pred = 0;\n      for (let j = 0; j < n; j++) {\n        pred += getX(i, j) * (w[j] ?? 0);\n      }\n      yHat[i] = pred;\n    }\n\n    // Precompute 1/m for efficiency\n    const invM = m === 0 ? 0 : 1 / m;\n\n    // Coordinate descent main loop\n    // Each iteration updates all coefficients once\n    for (let iter = 0; iter < maxIter; iter++) {\n      // Track maximum coefficient change for convergence check\n      let maxChange = 0;\n\n      // Determine order of coordinate updates\n      // Cyclic: iterate through features in order (more cache-friendly)\n      // Random: shuffle features each iteration (can help convergence)\n      let indices: number[] | null = null;\n      if (selection === \"random\") {\n        indices = Array.from({ length: n }, (_, j) => j);\n        // Fisher-Yates shuffle for random selection\n        for (let k = n - 1; k > 0; k--) {\n          const r = Math.floor(rng() * (k + 1));\n          const tmp = indices[k];\n          indices[k] = indices[r] ?? 0;\n          indices[r] = tmp ?? 0;\n        }\n      }\n\n      // Update each coefficient using coordinate descent\n      const iterOrder = indices ?? Array.from({ length: n }, (_, j) => j);\n      for (const j of iterOrder) {\n        const denom = colNorm2[j] ?? 0;\n\n        // Skip features with zero variance (constant columns)\n        if (denom === 0) {\n          const prevW = w[j] ?? 0;\n          if (prevW !== 0) {\n            const delta = -prevW;\n            for (let i = 0; i < m; i++) {\n              yHat[i] = (yHat[i] ?? 0) + delta * getX(i, j);\n            }\n            maxChange = Math.max(maxChange, Math.abs(delta));\n          }\n          w[j] = 0;\n          continue;\n        }\n\n        // Compute correlation between feature j and current residual\n        // rho = (1/m) * Σᵢ xᵢⱼ * (yᵢ - ŷᵢ + wⱼ * xᵢⱼ)\n        // The term (wⱼ * xᵢⱼ) adds back the contribution of feature j to the residual\n        let rho = 0;\n        for (let i = 0; i < m; i++) {\n          const xij = getX(i, j);\n          const yi = Number(y.data[y.offset + i] ?? 0) - (fitIntercept ? yMean : 0);\n          // Current residual plus contribution from feature j\n          const r = yi - (yHat[i] ?? 0) + (w[j] ?? 0) * xij;\n          rho += xij * r;\n        }\n        rho *= invM;\n\n        // Apply soft thresholding operator (proximal operator for L1 norm)\n        // This is the key step that induces sparsity in Lasso\n        // newW = soft_threshold(rho, α) / ||xⱼ||²\n        let newW = this.softThreshold(rho, alpha) / denom;\n\n        // Enforce non-negativity constraint if requested\n        if (positive && newW < 0) {\n          newW = 0;\n        }\n\n        // Compute change in coefficient\n        const delta = newW - (w[j] ?? 0);\n\n        // Update predictions incrementally if coefficient changed\n        // This is much more efficient than recomputing all predictions\n        // ŷ_new = ŷ_old + Δwⱼ * xⱼ\n        if (delta !== 0) {\n          for (let i = 0; i < m; i++) {\n            yHat[i] = (yHat[i] ?? 0) + delta * getX(i, j);\n          }\n        }\n\n        // Update coefficient\n        w[j] = newW;\n\n        // Track maximum change for convergence check\n        maxChange = Math.max(maxChange, Math.abs(delta));\n      }\n\n      // Check convergence: if no coefficient changed by more than tol, we're done\n      if (maxChange < tol) {\n        this.nIter_ = iter + 1;\n        break;\n      }\n    }\n\n    // Store final iteration count if we didn't converge early\n    if (this.nIter_ === undefined) {\n      this.nIter_ = maxIter;\n    }\n\n    // Rescale coefficients back to original feature space if normalized\n    if (normalize && xScale) {\n      for (let j = 0; j < n; j++) {\n        const s = xScale[j] ?? 1;\n        w[j] = s === 0 ? 0 : (w[j] ?? 0) / s;\n      }\n    }\n\n    // Store final coefficients\n    this.coef_ = tensor(w);\n\n    // Compute intercept if needed\n    // intercept = mean(y) - mean(X) @ coef\n    // This accounts for the centering we did during optimization\n    if (fitIntercept) {\n      let xMeanDotW = 0;\n      for (let j = 0; j < n; j++) {\n        xMeanDotW += (xMean[j] ?? 0) * (w[j] ?? 0);\n      }\n      this.intercept_ = yMean - xMeanDotW;\n    } else {\n      this.intercept_ = 0;\n    }\n\n    // Mark model as fitted\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Soft thresholding operator (proximal operator for L1 norm).\n   *\n   * This is the key operation in Lasso that induces sparsity.\n   *\n   * Formula:\n   * - If x > λ: return x - λ\n   * - If x < -λ: return x + λ\n   * - Otherwise: return 0\n   *\n   * Geometrically, this \"shrinks\" x towards zero by λ,\n   * and sets it exactly to zero if |x| ≤ λ.\n   *\n   * **Time Complexity**: O(1)\n   *\n   * @param x - Input value\n   * @param lambda - Threshold parameter (regularization strength)\n   * @returns Soft-thresholded value\n   */\n  private softThreshold(x: number, lambda: number): number {\n    // Guard against non-finite inputs\n    if (!Number.isFinite(x) || !Number.isFinite(lambda)) {\n      throw new DataValidationError(\"Non-finite value encountered during soft-thresholding\");\n    }\n    if (x > lambda) return x - lambda;\n    if (x < -lambda) return x + lambda;\n    return 0;\n  }\n\n  /**\n   * Get the model coefficients (weights).\n   *\n   * Many coefficients will be exactly zero due to L1 regularization (sparsity).\n   *\n   * @returns Coefficient tensor of shape (n_features,)\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get coef(): Tensor {\n    if (!this.fitted || !this.coef_) {\n      throw new NotFittedError(\"Lasso must be fitted to access coefficients\");\n    }\n    return this.coef_;\n  }\n\n  /**\n   * Get the intercept (bias term).\n   *\n   * @returns Intercept value\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get intercept(): number {\n    if (!this.fitted) {\n      throw new NotFittedError(\"Lasso must be fitted to access intercept\");\n    }\n    return this.intercept_;\n  }\n\n  /**\n   * Get the number of iterations run by coordinate descent.\n   *\n   * @returns Number of iterations until convergence\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get nIter(): number | undefined {\n    if (!this.fitted) {\n      throw new NotFittedError(\"Lasso must be fitted to access nIter\");\n    }\n    return this.nIter_;\n  }\n\n  /**\n   * Predict using the Lasso regression model.\n   *\n   * Computes predictions as: ŷ = X @ coef + intercept\n   *\n   * **Time Complexity**: O(nm) where n = samples, m = features\n   * **Space Complexity**: O(n)\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    // Check if model has been fitted\n    if (!this.fitted || !this.coef_) {\n      throw new NotFittedError(\"Lasso must be fitted before prediction\");\n    }\n\n    // Validate input\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"Lasso\");\n\n    const m = X.shape[0] ?? 0; // Number of samples to predict\n    const n = X.shape[1] ?? 0; // Number of features\n    const pred = Array(m).fill(0);\n\n    // Compute predictions: ŷ[i] = Σⱼ X[i,j] * coef[j] + intercept\n    for (let i = 0; i < m; i++) {\n      // Compute weighted sum of features\n      for (let j = 0; j < n; j++) {\n        pred[i] +=\n          Number(X.data[X.offset + i * n + j] ?? 0) *\n          Number(this.coef_.data[this.coef_.offset + j] ?? 0);\n      }\n      // Add intercept\n      pred[i] += this.intercept_;\n    }\n\n    return tensor(pred);\n  }\n\n  /**\n   * Return the R² score on the given test data and target values.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True target values of shape (n_samples,)\n   * @returns R² score (best possible is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (!this.fitted) {\n      throw new NotFittedError(\"Lasso must be fitted before scoring\");\n    }\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const pred = this.predict(X);\n    if (pred.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${pred.size}, y=${y.size}`\n      );\n    }\n    let ssRes = 0,\n      ssTot = 0;\n    let yMean = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      yMean += Number(y.data[y.offset + i] ?? 0);\n    }\n    yMean /= y.size;\n\n    for (let i = 0; i < y.size; i++) {\n      const yVal = Number(y.data[y.offset + i] ?? 0);\n      const predVal = Number(pred.data[pred.offset + i] ?? 0);\n      ssRes += (yVal - predVal) ** 2;\n      ssTot += (yVal - yMean) ** 2;\n    }\n\n    if (ssTot === 0) {\n      return ssRes === 0 ? 1.0 : 0.0;\n    }\n\n    return 1 - ssRes / ssTot;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return { ...this.options };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param params - Parameters to set (alpha, maxIter, tol, fitIntercept, normalize)\n   * @returns this\n   * @throws {InvalidParameterError} If any parameter value is invalid\n   */\n  setParams(params: Record<string, unknown>): this {\n    for (const [key, value] of Object.entries(params)) {\n      switch (key) {\n        case \"alpha\":\n          if (typeof value !== \"number\" || !Number.isFinite(value)) {\n            throw new InvalidParameterError(\"alpha must be a finite number\", \"alpha\", value);\n          }\n          this.options.alpha = value;\n          break;\n        case \"maxIter\":\n          if (typeof value !== \"number\" || !Number.isFinite(value)) {\n            throw new InvalidParameterError(\"maxIter must be a finite number\", \"maxIter\", value);\n          }\n          this.options.maxIter = value;\n          break;\n        case \"tol\":\n          if (typeof value !== \"number\" || !Number.isFinite(value)) {\n            throw new InvalidParameterError(\"tol must be a finite number\", \"tol\", value);\n          }\n          this.options.tol = value;\n          break;\n        case \"fitIntercept\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\n              \"fitIntercept must be a boolean\",\n              \"fitIntercept\",\n              value\n            );\n          }\n          this.options.fitIntercept = value;\n          break;\n        case \"normalize\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\"normalize must be a boolean\", \"normalize\", value);\n          }\n          this.options.normalize = value;\n          break;\n        case \"warmStart\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\"warmStart must be a boolean\", \"warmStart\", value);\n          }\n          this.options.warmStart = value;\n          break;\n        case \"positive\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\"positive must be a boolean\", \"positive\", value);\n          }\n          this.options.positive = value;\n          break;\n        case \"selection\":\n          if (value !== \"cyclic\" && value !== \"random\") {\n            throw new InvalidParameterError(\n              `Invalid selection: ${String(value)}`,\n              \"selection\",\n              value\n            );\n          }\n          this.options.selection = value;\n          break;\n        case \"randomState\":\n          if (typeof value !== \"number\" || !Number.isFinite(value)) {\n            throw new InvalidParameterError(\n              `randomState must be a finite number; received ${String(value)}`,\n              \"randomState\",\n              value\n            );\n          }\n          this.options.randomState = value;\n          break;\n        default:\n          throw new InvalidParameterError(`Unknown parameter: ${key}`, key, value);\n      }\n    }\n    return this;\n  }\n}\n","import { DataValidationError, InvalidParameterError, NotFittedError, ShapeError } from \"../../core\";\nimport { lstsq } from \"../../linalg\";\nimport { dot, mean, sub, type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Regressor } from \"../base\";\n\n/**\n * Ordinary Least Squares Linear Regression.\n *\n * Fits a linear model with coefficients w = (w1, ..., wp) to minimize\n * the residual sum of squares between the observed targets and the\n * targets predicted by the linear approximation.\n *\n * @example\n * ```ts\n * import { LinearRegression } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * // Create training data\n * const X = tensor([[1, 1], [1, 2], [2, 2], [2, 3]]);\n * const y = tensor([1, 2, 2, 3]);\n *\n * // Fit model\n * const model = new LinearRegression({ fitIntercept: true });\n * model.fit(X, y);\n *\n * // Make predictions\n * const X_test = tensor([[3, 5]]);\n * const predictions = model.predict(X_test);\n *\n * // Get R^2 score\n * const score = model.score(X, y);\n * ```\n */\nexport class LinearRegression implements Regressor {\n  /** Model coefficients (weights) of shape (n_features,) or (n_features, n_targets) */\n  private coef_?: Tensor;\n\n  /** Independent term (bias/intercept) in the linear model */\n  private intercept_?: Tensor;\n\n  /** Number of features seen during fit */\n  private nFeaturesIn_?: number;\n\n  /** Whether the model has been fitted */\n  private fitted = false;\n\n  private options: {\n    fitIntercept?: boolean;\n    normalize?: boolean;\n    copyX?: boolean;\n  };\n\n  /**\n   * Create a new Linear Regression model.\n   *\n   * @param options - Configuration options\n   * @param options.fitIntercept - Whether to calculate the intercept (default: true)\n   * @param options.normalize - Whether to normalize features before regression (default: false)\n   * @param options.copyX - Whether to copy X or overwrite it (default: true)\n   */\n  constructor(\n    options: {\n      readonly fitIntercept?: boolean;\n      readonly normalize?: boolean;\n      readonly copyX?: boolean;\n    } = {}\n  ) {\n    this.options = { ...options };\n  }\n\n  /**\n   * Fit linear model using Ordinary Least Squares.\n   *\n   * Uses SVD-based least squares solver for numerical stability.\n   * When fitIntercept is true, centers the data before fitting.\n   *\n   * **Algorithm Complexity**: O(n * p^2) where n = samples, p = features\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,). Multi-output regression is not currently supported.\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {DataValidationError} If X or y are empty\n   */\n  fit(X: Tensor, y: Tensor): this {\n    // Validate inputs (dimensions, empty data, NaN/Inf)\n    validateFitInputs(X, y);\n\n    // Store number of features for validation during predict\n    const nFeatures = X.shape[1] ?? 0;\n    this.nFeaturesIn_ = nFeatures;\n\n    // Handle intercept by centering the data\n    // This is more numerically stable than adding a column of ones\n    const fitIntercept = this.options.fitIntercept ?? true;\n    const copyX = this.options.copyX ?? true;\n    const allowInPlace = copyX === false && (X.dtype === \"float32\" || X.dtype === \"float64\");\n\n    if (fitIntercept) {\n      // Compute mean of X along axis 0 (column means)\n      const X_mean = mean(X, 0);\n      // Compute mean of y\n      const y_mean = mean(y);\n\n      // Center X and y by subtracting means\n      let X_processed = allowInPlace\n        ? this.centerDataInPlace(X, X_mean)\n        : this.centerData(X, X_mean);\n      const y_centered = this.centerData(y, y_mean);\n\n      // Handle normalization if requested\n      let X_scale: Tensor | undefined;\n      if (this.options.normalize) {\n        // Compute L2 norm of centered X\n        X_scale = this.computeL2Norm(X_processed);\n        // Divide by norm\n        X_processed = allowInPlace\n          ? this.scaleDataInPlace(X_processed, X_scale)\n          : this.scaleData(X_processed, X_scale);\n      }\n\n      // Solve least squares on processed data: X_processed * w = y_centered\n      const result = lstsq(X_processed, y_centered);\n      let w = result.x;\n\n      // If normalized, rescale coefficients: w_original = w_normalized / scale\n      if (X_scale) {\n        w = this.rescaleCoefs(w, X_scale);\n      }\n      this.coef_ = w;\n\n      // Compute intercept: b = y_mean - X_mean @ coef\n      const X_mean_dot_coef = dot(X_mean, this.coef_);\n      this.intercept_ = sub(y_mean, X_mean_dot_coef);\n    } else {\n      // No intercept - solve directly using least squares\n      const result = lstsq(X, y);\n      this.coef_ = result.x;\n      // intercept_ remains undefined when fitIntercept is false\n    }\n\n    // Mark model as fitted\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Center data by subtracting the mean.\n   *\n   * @param data - Input tensor to center\n   * @param dataMean - Mean tensor to subtract\n   * @returns Centered tensor\n   */\n  private centerData(data: Tensor, dataMean: Tensor): Tensor {\n    // For 1D data, manually subtract mean to avoid dtype mismatch\n    if (data.ndim === 1) {\n      const n = data.size;\n      const meanVal = Number(dataMean.data[dataMean.offset] ?? 0);\n      const centered: number[] = [];\n      for (let i = 0; i < n; i++) {\n        centered.push(Number(data.data[data.offset + i] ?? 0) - meanVal);\n      }\n      return tensor(centered);\n    }\n\n    // For 2D data, broadcast subtract row mean from each row\n    const nSamples = data.shape[0] ?? 0;\n    const nFeatures = data.shape[1] ?? 0;\n\n    const result: number[][] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        const val = Number(data.data[data.offset + i * nFeatures + j] ?? 0);\n        const meanVal = Number(dataMean.data[dataMean.offset + j] ?? 0);\n        row.push(val - meanVal);\n      }\n      result.push(row);\n    }\n\n    return tensor(result);\n  }\n\n  private centerDataInPlace(data: Tensor, dataMean: Tensor): Tensor {\n    if (data.ndim === 1) {\n      const n = data.size;\n      const meanVal = Number(dataMean.data[dataMean.offset] ?? 0);\n      for (let i = 0; i < n; i++) {\n        const idx = data.offset + i;\n        data.data[idx] = Number(data.data[idx] ?? 0) - meanVal;\n      }\n      return data;\n    }\n\n    const nSamples = data.shape[0] ?? 0;\n    const nFeatures = data.shape[1] ?? 0;\n    for (let i = 0; i < nSamples; i++) {\n      const rowBase = data.offset + i * nFeatures;\n      for (let j = 0; j < nFeatures; j++) {\n        const idx = rowBase + j;\n        const meanVal = Number(dataMean.data[dataMean.offset + j] ?? 0);\n        data.data[idx] = Number(data.data[idx] ?? 0) - meanVal;\n      }\n    }\n    return data;\n  }\n\n  private computeL2Norm(X: Tensor): Tensor {\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const norms: number[] = new Array(nFeatures).fill(0);\n\n    for (let j = 0; j < nFeatures; j++) {\n      let sumSq = 0;\n      for (let i = 0; i < nSamples; i++) {\n        const val = Number(X.data[X.offset + i * nFeatures + j] ?? 0);\n        sumSq += val * val;\n      }\n      norms[j] = Math.sqrt(sumSq);\n    }\n    return tensor(norms);\n  }\n\n  private scaleData(X: Tensor, scale: Tensor): Tensor {\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const result: number[][] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        const val = Number(X.data[X.offset + i * nFeatures + j] ?? 0);\n        const s = Number(scale.data[scale.offset + j] ?? 1);\n        row.push(s === 0 ? 0 : val / s); // Handle zero norm\n      }\n      result.push(row);\n    }\n    return tensor(result);\n  }\n\n  private scaleDataInPlace(X: Tensor, scale: Tensor): Tensor {\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    for (let i = 0; i < nSamples; i++) {\n      const rowBase = X.offset + i * nFeatures;\n      for (let j = 0; j < nFeatures; j++) {\n        const idx = rowBase + j;\n        const s = Number(scale.data[scale.offset + j] ?? 1);\n        X.data[idx] = s === 0 ? 0 : Number(X.data[idx] ?? 0) / s;\n      }\n    }\n    return X;\n  }\n\n  private rescaleCoefs(coef: Tensor, scale: Tensor): Tensor {\n    // coef is (n_features,) or (n_features, n_targets)\n    // We need to divide each row j by scale[j]\n    const nFeatures = coef.shape[0] ?? 0;\n    const nTargets = coef.ndim > 1 ? (coef.shape[1] ?? 1) : 1;\n\n    if (coef.ndim === 1) {\n      const res: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        const c = Number(coef.data[coef.offset + j] ?? 0);\n        const s = Number(scale.data[scale.offset + j] ?? 1);\n        res.push(s === 0 ? 0 : c / s);\n      }\n      return tensor(res);\n    }\n\n    const result: number[][] = [];\n    for (let j = 0; j < nFeatures; j++) {\n      const row: number[] = [];\n      const s = Number(scale.data[scale.offset + j] ?? 1);\n      for (let k = 0; k < nTargets; k++) {\n        const c = Number(coef.data[coef.offset + j * nTargets + k] ?? 0);\n        row.push(s === 0 ? 0 : c / s);\n      }\n      result.push(row);\n    }\n    return tensor(result);\n  }\n\n  /**\n   * Predict using the linear model.\n   *\n   * Computes y_pred = X * coef_ + intercept_\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted || !this.coef_) {\n      throw new NotFittedError(\"LinearRegression must be fitted before prediction\");\n    }\n\n    // Validate input\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"LinearRegression\");\n\n    // Compute predictions: y_pred = X * w + b\n    const y_pred_raw = dot(X, this.coef_);\n\n    // Add intercept if model was fitted with fitIntercept=true\n    if (this.intercept_ !== undefined) {\n      // Manually add intercept to avoid dtype mismatch\n      const interceptVal = Number(this.intercept_.data[this.intercept_.offset] ?? 0);\n      const result: number[] = [];\n      for (let i = 0; i < y_pred_raw.size; i++) {\n        result.push(Number(y_pred_raw.data[y_pred_raw.offset + i] ?? 0) + interceptVal);\n      }\n      return tensor(result);\n    }\n\n    return y_pred_raw;\n  }\n\n  /**\n   * Return the coefficient of determination R^2 of the prediction.\n   *\n   * R^2 = 1 - (SS_res / SS_tot)\n   *\n   * Where:\n   * - SS_res = Σ(y_true - y_pred)^2 (residual sum of squares)\n   * - SS_tot = Σ(y_true - y_mean)^2 (total sum of squares)\n   *\n   * Best possible score is 1.0, and it can be negative (worse than random).\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True target values of shape (n_samples,)\n   * @returns R² score (best possible is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (!this.fitted) {\n      throw new NotFittedError(\"LinearRegression must be fitted before scoring\");\n    }\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n\n    // Get predictions\n    const y_pred = this.predict(X);\n    if (y_pred.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${y_pred.size}, y=${y.size}`\n      );\n    }\n\n    // Compute residual sum of squares: SS_res = Σ(y_true - y_pred)^2\n    // Manually compute to avoid dtype mismatch\n    let ss_res = 0;\n    for (let i = 0; i < y.size; i++) {\n      const diff = Number(y.data[y.offset + i] ?? 0) - Number(y_pred.data[y_pred.offset + i] ?? 0);\n      ss_res += diff * diff;\n    }\n\n    // Compute total sum of squares: SS_tot = Σ(y_true - y_mean)^2\n    const y_mean_tensor = mean(y);\n    const y_mean_val = Number(y_mean_tensor.data[y_mean_tensor.offset] ?? 0);\n    let ss_tot = 0;\n    for (let i = 0; i < y.size; i++) {\n      const diff = Number(y.data[y.offset + i] ?? 0) - y_mean_val;\n      ss_tot += diff * diff;\n    }\n\n    // R^2 = 1 - (SS_res / SS_tot)\n    // Handle edge case where ss_tot is 0 (constant y)\n    if (ss_tot === 0) {\n      return ss_res === 0 ? 1.0 : 0.0;\n    }\n\n    return 1 - ss_res / ss_tot;\n  }\n\n  /**\n   * Get the model coefficients (weights).\n   *\n   * @returns Coefficient tensor of shape (n_features,) or (n_features, n_targets)\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get coef(): Tensor {\n    if (!this.fitted || !this.coef_) {\n      throw new NotFittedError(\"LinearRegression must be fitted to access coefficients\");\n    }\n    return this.coef_;\n  }\n\n  /**\n   * Get the intercept (bias term).\n   *\n   * @returns Intercept value or tensor\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get intercept(): Tensor | undefined {\n    if (!this.fitted) {\n      throw new NotFittedError(\"LinearRegression must be fitted to access intercept\");\n    }\n    return this.intercept_;\n  }\n\n  /**\n   * Get parameters for this estimator.\n   *\n   * @returns Object containing all parameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      fitIntercept: this.options.fitIntercept ?? true,\n      normalize: this.options.normalize ?? false,\n      copyX: this.options.copyX ?? true,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param params - Parameters to set\n   * @returns this - The estimator\n   */\n  setParams(_params: Record<string, unknown>): this {\n    for (const [key, value] of Object.entries(_params)) {\n      switch (key) {\n        case \"fitIntercept\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\n              `fitIntercept must be a boolean; received ${String(value)}`,\n              \"fitIntercept\",\n              value\n            );\n          }\n          this.options.fitIntercept = value;\n          break;\n        case \"normalize\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\n              `normalize must be a boolean; received ${String(value)}`,\n              \"normalize\",\n              value\n            );\n          }\n          this.options.normalize = value;\n          break;\n        case \"copyX\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\n              `copyX must be a boolean; received ${String(value)}`,\n              \"copyX\",\n              value\n            );\n          }\n          this.options.copyX = value;\n          break;\n        default:\n          throw new InvalidParameterError(`Unknown parameter: ${key}`, key, value);\n      }\n    }\n    return this;\n  }\n}\n","import {\n  DataValidationError,\n  DeepboxError,\n  InvalidParameterError,\n  NotFittedError,\n  ShapeError,\n} from \"../../core\";\nimport { reshape, type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Classifier } from \"../base\";\n\n/**\n * Logistic Regression (Binary and Multiclass Classification).\n *\n * Logistic regression uses the logistic (sigmoid) function to model\n * the probability of class membership.\n *\n * @example\n * ```ts\n * import { LogisticRegression } from 'deepbox/ml';\n *\n * // Binary classification\n * const model = new LogisticRegression({ C: 1.0, maxIter: 100 });\n * model.fit(X_train, y_train);\n *\n * const predictions = model.predict(X_test);\n * const probabilities = model.predictProba(X_test);\n *\n * // Multiclass classification\n * const multiModel = new LogisticRegression({\n *   multiClass: 'multinomial',\n *   solver: 'lbfgs'\n * });\n * multiModel.fit(X_train_multi, y_train_multi);\n * ```\n *\n * @category Linear Models\n * @implements {Classifier}\n */\nexport class LogisticRegression implements Classifier {\n  private options: {\n    penalty?: \"l2\" | \"none\";\n    tol?: number;\n    C?: number;\n    fitIntercept?: boolean;\n    maxIter?: number;\n    learningRate?: number;\n    multiClass?: \"ovr\" | \"auto\";\n  };\n\n  private coef_?: Tensor; // Shape (n_features,) for binary, (n_classes, n_features) for multiclass\n  private intercept_?: number | number[]; // Scalar for binary, Array for multiclass\n  private nFeaturesIn_?: number;\n  private classes_?: Tensor;\n  private fitted = false;\n  private multiclass_ = false;\n\n  /**\n   * Create a new Logistic Regression classifier.\n   *\n   * @param options - Configuration options\n   * @param options.penalty - Regularization type: 'l2' or 'none' (default: 'l2')\n   * @param options.C - Inverse regularization strength (default: 1.0). Must be > 0.\n   * @param options.tol - Tolerance for stopping criterion (default: 1e-4)\n   * @param options.maxIter - Maximum number of iterations (default: 100)\n   * @param options.fitIntercept - Whether to fit intercept (default: true)\n   * @param options.learningRate - Learning rate for gradient descent (default: 0.1)\n   * @param options.multiClass - Multiclass strategy: 'ovr' (One-vs-Rest) or 'auto' (default: 'auto')\n   */\n  constructor(\n    options: {\n      readonly penalty?: \"l2\" | \"none\";\n      readonly tol?: number;\n      readonly C?: number;\n      readonly fitIntercept?: boolean;\n      readonly maxIter?: number;\n      readonly learningRate?: number;\n      readonly multiClass?: \"ovr\" | \"auto\";\n    } = {}\n  ) {\n    this.options = { ...options };\n\n    const penalty = this.options.penalty ?? \"l2\";\n    if (penalty !== \"l2\" && penalty !== \"none\") {\n      throw new InvalidParameterError(\n        `Only penalty='l2' or 'none' is supported; received ${String(penalty)}`,\n        \"penalty\",\n        penalty\n      );\n    }\n    this.options.penalty = penalty;\n\n    const multiClass = this.options.multiClass ?? \"auto\";\n    if (multiClass !== \"ovr\" && multiClass !== \"auto\") {\n      throw new InvalidParameterError(\n        `multiClass must be 'ovr' or 'auto'; received ${String(multiClass)}`,\n        \"multiClass\",\n        multiClass\n      );\n    }\n    this.options.multiClass = multiClass;\n\n    if (this.options.C !== undefined && this.options.C <= 0) {\n      throw new InvalidParameterError(\n        `C must be > 0; received ${this.options.C}`,\n        \"C\",\n        this.options.C\n      );\n    }\n    if (\n      this.options.maxIter !== undefined &&\n      (!Number.isFinite(this.options.maxIter) || this.options.maxIter <= 0)\n    ) {\n      throw new InvalidParameterError(\n        `maxIter must be a positive finite number; received ${this.options.maxIter}`,\n        \"maxIter\",\n        this.options.maxIter\n      );\n    }\n    if (\n      this.options.tol !== undefined &&\n      (!Number.isFinite(this.options.tol) || this.options.tol < 0)\n    ) {\n      throw new InvalidParameterError(\n        `tol must be a finite number >= 0; received ${this.options.tol}`,\n        \"tol\",\n        this.options.tol\n      );\n    }\n    if (\n      this.options.learningRate !== undefined &&\n      (!Number.isFinite(this.options.learningRate) || this.options.learningRate <= 0)\n    ) {\n      throw new InvalidParameterError(\n        `learningRate must be a positive finite number; received ${this.options.learningRate}`,\n        \"learningRate\",\n        this.options.learningRate\n      );\n    }\n  }\n\n  /**\n   * Numerically stable sigmoid function.\n   *\n   * Uses different formulations for positive and negative inputs\n   * to avoid overflow:\n   * - For z >= 0: σ(z) = 1 / (1 + exp(-z))\n   * - For z < 0: σ(z) = exp(z) / (1 + exp(z))\n   *\n   * @param z - Input value\n   * @returns Sigmoid output in [0, 1]\n   */\n  private sigmoid(z: number): number {\n    // Guard against non-finite inputs\n    if (!Number.isFinite(z)) {\n      return z > 0 ? 1 : 0;\n    }\n    if (z >= 0) {\n      const ez = Math.exp(-z);\n      return 1 / (1 + ez);\n    }\n    const ez = Math.exp(z);\n    return ez / (1 + ez);\n  }\n\n  private ensureFitted(): void {\n    if (!this.fitted || !this.coef_) {\n      throw new NotFittedError(\"LogisticRegression must be fitted before using this method\");\n    }\n  }\n\n  private _fitBinary(\n    X: Tensor,\n    y: Tensor,\n    m: number,\n    n: number,\n    lambda: number\n  ): { w: number[]; b: number } {\n    const maxIter = this.options.maxIter ?? 100;\n    const tol = this.options.tol ?? 1e-4;\n    const lr = this.options.learningRate ?? 0.1;\n    const fitIntercept = this.options.fitIntercept ?? true;\n\n    // Initialize weights and bias\n    const w = new Array<number>(n).fill(0);\n    let b = 0;\n\n    // Gradient descent training loop\n    for (let iter = 0; iter < maxIter; iter++) {\n      const gradW = new Array<number>(n).fill(0);\n      let gradB = 0;\n\n      // Compute gradients over all samples\n      for (let i = 0; i < m; i++) {\n        let z = fitIntercept ? b : 0;\n        const rowBase = X.offset + i * n;\n        for (let j = 0; j < n; j++) {\n          z += Number(X.data[rowBase + j] ?? 0) * (w[j] ?? 0);\n        }\n\n        const yi = Number(y.data[y.offset + i] ?? 0);\n        const pi = this.sigmoid(z);\n        const error = pi - yi;\n\n        gradB += error;\n        for (let j = 0; j < n; j++) {\n          gradW[j] = (gradW[j] ?? 0) + error * Number(X.data[rowBase + j] ?? 0);\n        }\n      }\n\n      // Update weights with gradient descent + L2 regularization\n      const invM = m === 0 ? 0 : 1 / m;\n      let maxUpdate = 0;\n      for (let j = 0; j < n; j++) {\n        const g = (gradW[j] ?? 0) * invM + lambda * (w[j] ?? 0);\n        const update = lr * g;\n        w[j] = (w[j] ?? 0) - update;\n        maxUpdate = Math.max(maxUpdate, Math.abs(update));\n      }\n      if (fitIntercept) {\n        const gB = gradB * invM;\n        const updateB = lr * gB;\n        b -= updateB;\n        maxUpdate = Math.max(maxUpdate, Math.abs(updateB));\n      }\n\n      // Check convergence\n      if (maxUpdate < tol) {\n        break;\n      }\n    }\n    return { w, b: fitIntercept ? b : 0 };\n  }\n\n  /**\n   * Fit logistic regression model.\n   *\n   * Uses gradient descent with L2 regularization.\n   * Supports binary and multiclass (One-vs-Rest) classification.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target labels of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {DataValidationError} If X or y are empty\n   * @throws {InvalidParameterError} If C <= 0 or penalty is invalid\n   */\n  fit(X: Tensor, y: Tensor): this {\n    // Validate inputs (dimensions, empty data, NaN/Inf)\n    validateFitInputs(X, y);\n\n    const m = X.shape[0] ?? 0;\n    const n = X.shape[1] ?? 0;\n    this.nFeaturesIn_ = n;\n\n    // Identify unique classes\n    const yData = new Float64Array(m);\n    for (let i = 0; i < m; i++) {\n      yData[i] = Number(y.data[y.offset + i]);\n    }\n    const uniqueClasses = [...new Set(yData)].sort((a, b) => a - b);\n    this.classes_ = tensor(uniqueClasses);\n\n    // Extract and validate hyperparameters\n    const penalty = this.options.penalty ?? \"l2\";\n    const C = this.options.C ?? 1.0;\n    if (!(C > 0)) {\n      throw new InvalidParameterError(`C must be > 0; received ${C}`, \"C\", C);\n    }\n    const lambda = penalty === \"l2\" ? 1 / C : 0;\n    const multiClass = this.options.multiClass ?? \"auto\";\n\n    if (uniqueClasses.length <= 2) {\n      // Binary classification\n      this.multiclass_ = false;\n\n      // Map labels to 0/1 if they are not already\n      let yBinary = y;\n      if (uniqueClasses.length === 2 && (uniqueClasses[0] !== 0 || uniqueClasses[1] !== 1)) {\n        // Map uniqueClasses[0] -> 0, uniqueClasses[1] -> 1\n        const mappedData = new Float64Array(m);\n        for (let i = 0; i < m; i++) {\n          mappedData[i] = yData[i] === uniqueClasses[1] ? 1 : 0;\n        }\n        yBinary = tensor(mappedData);\n      } else if (uniqueClasses.length === 1) {\n        // If only 1 class, we still need 0/1 for the math, though it's degenerate\n        const mappedData = new Float64Array(m);\n        const target = uniqueClasses[0] === 1 ? 1 : 0; // If only class is 1, treat as all 1s\n        mappedData.fill(target);\n        yBinary = tensor(mappedData);\n      } else {\n        // Check if they are 0 and 1\n        for (const val of uniqueClasses) {\n          if (val !== 0 && val !== 1) {\n            // Should be caught above, but as fallback\n            throw new DataValidationError(\"Binary classification expects labels 0 and 1\");\n          }\n        }\n      }\n\n      const { w, b } = this._fitBinary(X, yBinary, m, n, lambda);\n      this.coef_ = tensor(w);\n      this.intercept_ = b;\n    } else {\n      // Multiclass (One-vs-Rest)\n      if (multiClass !== \"ovr\" && multiClass !== \"auto\") {\n        throw new InvalidParameterError(\n          `multiClass must be 'ovr' or 'auto'; received ${String(multiClass)}`,\n          \"multiClass\",\n          multiClass\n        );\n      }\n      this.multiclass_ = true;\n      const nClasses = uniqueClasses.length;\n      const allCoefs: number[] = []; // Flattened (nClasses * nFeatures)\n      const allIntercepts: number[] = [];\n\n      for (let k = 0; k < nClasses; k++) {\n        const targetClass = uniqueClasses[k];\n        // Create binary target for class k\n        const yBinaryData = new Float64Array(m);\n        for (let i = 0; i < m; i++) {\n          yBinaryData[i] = yData[i] === targetClass ? 1 : 0;\n        }\n        const yBinary = tensor(yBinaryData);\n\n        const { w, b } = this._fitBinary(X, yBinary, m, n, lambda);\n        allCoefs.push(...w);\n        allIntercepts.push(b);\n      }\n\n      const allCoefsTensor = tensor(allCoefs);\n      this.coef_ = reshape(allCoefsTensor, [nClasses, n]);\n      this.intercept_ = allIntercepts;\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  get classes(): Tensor | undefined {\n    return this.classes_;\n  }\n\n  /**\n   * Get the model coefficients (weights).\n   *\n   * @returns Coefficient tensor of shape (n_features,)\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get coef(): Tensor {\n    this.ensureFitted();\n    const coef = this.coef_;\n    if (!coef) {\n      throw new DeepboxError(\"Internal error: coef_ is missing after ensureFitted() \");\n    }\n    return coef;\n  }\n\n  /**\n   * Get the intercept (bias term).\n   *\n   * @returns Intercept value (scalar for binary, array for multiclass)\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get intercept(): number | number[] {\n    this.ensureFitted();\n    if (this.intercept_ === undefined) {\n      return 0;\n    }\n    return this.intercept_;\n  }\n\n  predict(X: Tensor): Tensor {\n    const proba = this.predictProba(X);\n    const m = X.shape[0] ?? 0;\n    const pred = new Array<number>(m).fill(0);\n\n    if (this.multiclass_) {\n      const classes = this.classes_;\n      if (!classes) {\n        throw new NotFittedError(\"Model not fitted (classes_ missing)\");\n      }\n      const nClasses = classes.size;\n      for (let i = 0; i < m; i++) {\n        let maxProb = -1;\n        let maxClassIdx = 0;\n        for (let k = 0; k < nClasses; k++) {\n          const p = Number(proba.data[proba.offset + i * nClasses + k]);\n          if (p > maxProb) {\n            maxProb = p;\n            maxClassIdx = k;\n          }\n        }\n        // Map index back to class label\n        pred[i] = Number(classes.data[classes.offset + maxClassIdx]);\n      }\n    } else {\n      const classes = this.classes_;\n      if (!classes) {\n        throw new NotFittedError(\"Model not fitted (classes_ missing)\");\n      }\n\n      // Handle binary classification mapping\n      // If we have 2 classes, map 0->classes[0], 1->classes[1]\n      // If we have 1 class, we always predict that class (degenerate case)\n      if (classes.size === 1) {\n        const cls = Number(classes.data[classes.offset]);\n        pred.fill(cls);\n      } else {\n        const cls0 = Number(classes.data[classes.offset]);\n        const cls1 = Number(classes.data[classes.offset + 1]);\n\n        for (let i = 0; i < m; i++) {\n          const p1 = Number(proba.data[proba.offset + i * 2 + 1]);\n          pred[i] = p1 >= 0.5 ? cls1 : cls0;\n        }\n      }\n    }\n    return tensor(pred);\n  }\n\n  /**\n   * Predict class probabilities for samples.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Probabilities of shape (n_samples, n_classes)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predictProba(X: Tensor): Tensor {\n    this.ensureFitted();\n\n    const coef = this.coef_;\n    if (!coef) {\n      throw new DeepboxError(\"Internal error: coef_ is missing after ensureFitted()\");\n    }\n\n    // Validate input\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"LogisticRegression\");\n\n    const m = X.shape[0] ?? 0;\n    const n = X.shape[1] ?? 0;\n\n    if (this.multiclass_) {\n      const nClasses = this.classes_?.size ?? 0;\n      const proba = new Float64Array(m * nClasses);\n      const interceptValue = this.intercept_;\n      if (!Array.isArray(interceptValue)) {\n        throw new DeepboxError(\"Internal error: intercept_ must be an array for multiclass\");\n      }\n\n      for (let i = 0; i < m; i++) {\n        const rowBase = X.offset + i * n;\n        let sumExp = 0;\n        const scores = new Array<number>(nClasses).fill(0);\n\n        // Compute scores for each class\n        for (let k = 0; k < nClasses; k++) {\n          let z = interceptValue[k] ?? 0;\n          const coefRowBase = coef.offset + k * n;\n          for (let j = 0; j < n; j++) {\n            z += Number(X.data[rowBase + j] ?? 0) * Number(coef.data[coefRowBase + j] ?? 0);\n          }\n          // For OvR, we apply sigmoid to get probability of class k vs rest\n          // Then we normalize these probabilities to sum to 1 (scikit-learn approach)\n          scores[k] = this.sigmoid(z);\n          sumExp += scores[k] ?? 0;\n        }\n\n        // Normalize\n        for (let k = 0; k < nClasses; k++) {\n          // If sum is 0 (unlikely with sigmoid), avoid NaN\n          proba[i * nClasses + k] = sumExp > 0 ? (scores[k] ?? 0) / sumExp : 1.0 / nClasses;\n        }\n      }\n\n      return tensor(proba, { dtype: \"float64\" }).reshape([m, nClasses]);\n    } else {\n      // Binary case\n      const proba = new Array<number>(m * 2).fill(0);\n      const interceptValue = this.intercept_;\n      if (Array.isArray(interceptValue) || typeof interceptValue !== \"number\") {\n        throw new DeepboxError(\"Internal error: intercept_ must be a number for binary case\");\n      }\n\n      for (let i = 0; i < m; i++) {\n        let z = interceptValue;\n        const rowBase = X.offset + i * n;\n        for (let j = 0; j < n; j++) {\n          z += Number(X.data[rowBase + j] ?? 0) * Number(coef.data[coef.offset + j] ?? 0);\n        }\n\n        const p1 = this.sigmoid(z);\n        proba[i * 2 + 0] = 1 - p1;\n        proba[i * 2 + 1] = p1;\n      }\n\n      return tensor(proba).reshape([m, 2]);\n    }\n  }\n\n  /**\n   * Return the mean accuracy on the given test data and labels.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True labels of shape (n_samples,)\n   * @returns Accuracy score in range [0, 1]\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const pred = this.predict(X);\n    if (pred.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${pred.size}, y=${y.size}`\n      );\n    }\n    let correct = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      if (Number(pred.data[pred.offset + i] ?? 0) === Number(y.data[y.offset + i] ?? 0)) {\n        correct++;\n      }\n    }\n    return correct / y.size;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return { ...this.options };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param params - Parameters to set (maxIter, tol, C, learningRate, penalty, fitIntercept)\n   * @returns this\n   * @throws {InvalidParameterError} If any parameter value is invalid\n   */\n  setParams(params: Record<string, unknown>): this {\n    for (const [key, value] of Object.entries(params)) {\n      switch (key) {\n        case \"maxIter\":\n          if (typeof value !== \"number\" || !Number.isFinite(value) || value <= 0) {\n            throw new InvalidParameterError(\n              \"maxIter must be a positive finite number\",\n              \"maxIter\",\n              value\n            );\n          }\n          this.options.maxIter = value;\n          break;\n        case \"tol\":\n          if (typeof value !== \"number\" || !Number.isFinite(value) || value < 0) {\n            throw new InvalidParameterError(\"tol must be a finite number >= 0\", \"tol\", value);\n          }\n          this.options.tol = value;\n          break;\n        case \"C\":\n          if (typeof value !== \"number\" || !Number.isFinite(value) || value <= 0) {\n            throw new InvalidParameterError(\"C must be a positive finite number\", \"C\", value);\n          }\n          this.options.C = value;\n          break;\n        case \"learningRate\":\n          if (typeof value !== \"number\" || !Number.isFinite(value) || value <= 0) {\n            throw new InvalidParameterError(\n              \"learningRate must be a positive finite number\",\n              \"learningRate\",\n              value\n            );\n          }\n          this.options.learningRate = value;\n          break;\n        case \"fitIntercept\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\n              \"fitIntercept must be a boolean\",\n              \"fitIntercept\",\n              value\n            );\n          }\n          this.options.fitIntercept = value;\n          break;\n        case \"penalty\":\n          if (value !== \"none\" && value !== \"l2\") {\n            throw new InvalidParameterError(\n              `Only penalty='l2' or 'none' is supported; received ${String(value)}`,\n              \"penalty\",\n              value\n            );\n          }\n          this.options.penalty = value;\n          break;\n        case \"multiClass\":\n          if (value !== \"ovr\" && value !== \"auto\") {\n            throw new InvalidParameterError(\n              `multiClass must be 'ovr' or 'auto'; received ${String(value)}`,\n              \"multiClass\",\n              value\n            );\n          }\n          this.options.multiClass = value;\n          break;\n        default:\n          throw new InvalidParameterError(`Unknown parameter: ${key}`, key, value);\n      }\n    }\n    return this;\n  }\n}\n","import { DataValidationError, InvalidParameterError, NotFittedError, ShapeError } from \"../../core\";\nimport { cholesky } from \"../../linalg/decomposition/cholesky\";\nimport { svd } from \"../../linalg/decomposition/svd\";\nimport { solveTriangular } from \"../../linalg/solvers/solve\";\nimport { dot, type Tensor, tensor, transpose } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Regressor } from \"../base\";\n\n/**\n * Ridge Regression (L2 Regularized Linear Regression).\n *\n * Ridge regression addresses multicollinearity by adding a penalty term\n * (L2 regularization) to the loss function.\n *\n * @example\n * ```ts\n * import { Ridge } from 'deepbox/ml';\n *\n * const model = new Ridge({ alpha: 0.5 });\n * model.fit(X_train, y_train);\n * const predictions = model.predict(X_test);\n * ```\n *\n * @category Linear Models\n * @implements {Regressor}\n */\nexport class Ridge implements Regressor {\n  /** Configuration options for the Ridge regression model */\n  private options: {\n    alpha?: number;\n    fitIntercept?: boolean;\n    normalize?: boolean;\n    solver?: \"auto\" | \"svd\" | \"cholesky\" | \"lsqr\" | \"sag\";\n    maxIter?: number;\n    tol?: number;\n  };\n\n  /** Model coefficients (weights) after fitting - shape (n_features,) */\n  private coef_?: Tensor;\n\n  /** Intercept (bias) term after fitting */\n  private intercept_?: number;\n\n  /** Number of features seen during fit - used for validation */\n  private nFeaturesIn_?: number;\n\n  /** Number of iterations run by the solver (for iterative solvers) */\n  private nIter_: number | undefined;\n\n  /** Whether the model has been fitted to data */\n  private fitted = false;\n\n  /**\n   * Create a new Ridge Regression model.\n   *\n   * @param options - Configuration options\n   * @param options.alpha - Regularization strength (default: 1.0). Must be >= 0.\n   * @param options.fitIntercept - Whether to calculate the intercept (default: true).\n   * @param options.normalize - Whether to normalize features before regression (default: false).\n   * @param options.solver - Solver to use (default: 'auto'). Options: 'auto', 'svd', 'cholesky', 'lsqr', 'sag'.\n   * @param options.maxIter - Maximum number of iterations for iterative solvers (default: 1000)\n   * @param options.tol - Tolerance for stopping criterion (default: 1e-4)\n   */\n  constructor(\n    options: {\n      readonly alpha?: number;\n      readonly fitIntercept?: boolean;\n      readonly normalize?: boolean;\n      readonly solver?: \"auto\" | \"svd\" | \"cholesky\" | \"lsqr\" | \"sag\";\n      readonly maxIter?: number;\n      readonly tol?: number;\n    } = {}\n  ) {\n    this.options = { ...options };\n  }\n\n  /**\n   * Fit Ridge regression model.\n   *\n   * Solves the regularized least squares problem:\n   * minimize ||y - Xw||² + α||w||²\n   *\n   * Uses the closed-form solution:\n   * w = (X^T X + αI)^(-1) X^T y\n   *\n   * **Time Complexity**: O(n²p + p³) where n = samples, p = features\n   * **Space Complexity**: O(p²)\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,)\n   * @returns this - The fitted estimator for method chaining\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {DataValidationError} If X or y are empty\n   * @throws {InvalidParameterError} If alpha < 0\n   */\n  fit(X: Tensor, y: Tensor): this {\n    // Validate inputs (dimensions, empty data, NaN/Inf)\n    validateFitInputs(X, y);\n    this.nIter_ = undefined;\n\n    // Extract and validate regularization parameter\n    const alpha = this.options.alpha ?? 1.0;\n    if (!(alpha >= 0)) {\n      throw new InvalidParameterError(`alpha must be >= 0; received ${alpha}`, \"alpha\", alpha);\n    }\n\n    // Determine whether to fit intercept\n    const fitIntercept = this.options.fitIntercept ?? true;\n\n    // Extract dimensions: m = number of samples, n = number of features\n    const m = X.shape[0] ?? 0;\n    const n = X.shape[1] ?? 0;\n\n    // Store number of features for prediction validation\n    this.nFeaturesIn_ = n;\n\n    // Compute means for centering (if fitIntercept is true)\n    // Centering improves numerical stability and allows intercept calculation\n    // Note: By centering X and y, we ensure the intercept is not regularized.\n    // The regularization penalty α||w||² only applies to the coefficients,\n    // not the intercept term. This is the standard Ridge regression behavior.\n    let yMean = 0;\n    const xMean = new Array<number>(n).fill(0);\n\n    if (fitIntercept) {\n      // Compute sum of y values\n      for (let i = 0; i < m; i++) {\n        yMean += Number(y.data[y.offset + i] ?? 0);\n      }\n\n      // Compute sum of each feature column\n      for (let i = 0; i < m; i++) {\n        const rowBase = X.offset + i * n;\n        for (let j = 0; j < n; j++) {\n          xMean[j] = (xMean[j] ?? 0) + Number(X.data[rowBase + j] ?? 0);\n        }\n      }\n\n      // Convert sums to means by dividing by number of samples\n      const invM = m === 0 ? 0 : 1 / m;\n      yMean *= invM;\n      for (let j = 0; j < n; j++) {\n        xMean[j] = (xMean[j] ?? 0) * invM;\n      }\n    }\n\n    const normalize = this.options.normalize ?? false;\n    const maxIter = this.options.maxIter ?? 1000;\n    const tol = this.options.tol ?? 1e-4;\n\n    let xScale: number[] | undefined;\n    if (normalize) {\n      xScale = new Array<number>(n).fill(0);\n      for (let i = 0; i < m; i++) {\n        const rowBase = X.offset + i * n;\n        for (let j = 0; j < n; j++) {\n          const centered = Number(X.data[rowBase + j] ?? 0) - (fitIntercept ? (xMean[j] ?? 0) : 0);\n          xScale[j] = (xScale[j] ?? 0) + centered * centered;\n        }\n      }\n      for (let j = 0; j < n; j++) {\n        xScale[j] = Math.sqrt(xScale[j] ?? 0);\n      }\n    }\n\n    const getX = (sampleIndex: number, featureIndex: number): number => {\n      const raw = Number(X.data[X.offset + sampleIndex * n + featureIndex] ?? 0);\n      const centered = raw - (fitIntercept ? (xMean[featureIndex] ?? 0) : 0);\n      if (normalize && xScale) {\n        const s = xScale[featureIndex] ?? 0;\n        return s === 0 ? 0 : centered / s;\n      }\n      return centered;\n    };\n\n    const getY = (sampleIndex: number): number => {\n      const raw = Number(y.data[y.offset + sampleIndex] ?? 0);\n      return fitIntercept ? raw - yMean : raw;\n    };\n\n    // Solve the linear system (X^T X + αI) w = X^T y\n    // This gives us the optimal coefficients w\n    let coefTensor: Tensor;\n    const solver = this.options.solver ?? \"auto\";\n\n    if (solver === \"sag\") {\n      const res = this.solveSag(getX, getY, m, n, alpha, maxIter, tol);\n      coefTensor = tensor(res.x);\n      this.nIter_ = res.nIter;\n    } else {\n      // Compute X^T X + αI (Gram matrix with regularization)\n      // This is the core of the Ridge regression solution\n      // Time complexity: O(n²m) for computing X^T X\n      const XTX = Array(n)\n        .fill(0)\n        .map(() => Array(n).fill(0));\n\n      for (let i = 0; i < n; i++) {\n        for (let j = 0; j < n; j++) {\n          let sum = 0;\n\n          // Compute (X^T X)[i,j] = Σ_k X[k,i] * X[k,j]\n          for (let k = 0; k < m; k++) {\n            const xi = getX(k, i);\n            const xj = getX(k, j);\n            sum += xi * xj;\n          }\n\n          // Add regularization term αI to diagonal\n          // This ensures the matrix is positive definite and invertible\n          const xtxRow = XTX[i];\n          if (xtxRow) xtxRow[j] = sum + (i === j ? alpha : 0);\n        }\n      }\n\n      // Compute X^T y (feature-target correlation vector)\n      // Time complexity: O(nm)\n      const XTy = new Array<number>(n).fill(0);\n\n      for (let i = 0; i < n; i++) {\n        let sum = 0;\n\n        // Compute (X^T y)[i] = Σ_j X[j,i] * y[j]\n        for (let j = 0; j < m; j++) {\n          const yVal = getY(j);\n          const xVal = getX(j, i);\n          sum += xVal * yVal;\n        }\n        XTy[i] = sum;\n      }\n\n      if (solver === \"lsqr\") {\n        const res = this.solveConjugateGradient(XTX, XTy, maxIter, tol);\n        coefTensor = tensor(res.x);\n        this.nIter_ = res.nIter;\n      } else if (solver === \"cholesky\" || solver === \"auto\") {\n        try {\n          const xtxTensor = tensor(XTX);\n          const xtyTensor = tensor(XTy);\n          const L = cholesky(xtxTensor);\n          const y_ = solveTriangular(L, xtyTensor, true);\n          coefTensor = solveTriangular(transpose(L), y_, false);\n        } catch (e) {\n          if (solver === \"auto\") {\n            // Fallback to Gaussian elimination\n            const res = this.solveLinearSystem(XTX, XTy);\n            coefTensor = tensor(res);\n          } else {\n            throw e;\n          }\n        }\n      } else if (solver === \"svd\") {\n        const xtxTensor = tensor(XTX);\n        const xtyTensor = tensor(XTy);\n        const [U, s, Vt] = svd(xtxTensor);\n\n        // w = V * S^-1 * U^T * y\n        const Ut = transpose(U);\n        const Uty = dot(Ut, xtyTensor);\n\n        const sData = s.data;\n        if (!(sData instanceof Float64Array)) {\n          throw new DataValidationError(\"svd returned non-float64 singular values\");\n        }\n        const scaledData = new Float64Array(Uty.size);\n        for (let i = 0; i < Uty.size; i++) {\n          const val = Number(Uty.data[Uty.offset + i]);\n          const sigma = sData[i] ?? 0;\n          scaledData[i] = Math.abs(sigma) > 1e-15 ? val / sigma : 0;\n        }\n        const scaled = tensor(scaledData);\n\n        const V = transpose(Vt);\n        coefTensor = dot(V, scaled);\n      } else {\n        const res = this.solveLinearSystem(XTX, XTy);\n        coefTensor = tensor(res);\n      }\n    }\n\n    if (normalize && xScale) {\n      coefTensor = this.rescaleCoefs(coefTensor, xScale);\n    }\n\n    this.coef_ = coefTensor;\n\n    // Compute intercept if needed\n    // intercept = mean(y) - mean(X) @ coef\n    // This accounts for the centering we did earlier\n    if (fitIntercept) {\n      let xMeanDotW = 0;\n      for (let j = 0; j < n; j++) {\n        const wj = Number(coefTensor.data[coefTensor.offset + j] ?? 0);\n        xMeanDotW += (xMean[j] ?? 0) * wj;\n      }\n      this.intercept_ = yMean - xMeanDotW;\n    } else {\n      this.intercept_ = 0;\n    }\n\n    // Mark model as fitted\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Solve linear system Ax = b using Gaussian elimination with partial pivoting.\n   *\n   * This is a numerically stable method for solving dense linear systems.\n   * For Ridge regression, A = X^T X + αI is symmetric positive definite,\n   * so Cholesky decomposition would be more efficient, but Gaussian elimination\n   * is more general and still provides good numerical stability.\n   *\n   * **Algorithm**:\n   * 1. Forward elimination: Convert A to upper triangular form\n   * 2. Partial pivoting: Swap rows to avoid division by small numbers\n   * 3. Back substitution: Solve for x from bottom to top\n   *\n   * **Time Complexity**: O(n³)\n   * **Space Complexity**: O(n²)\n   *\n   * @param A - Coefficient matrix (n × n)\n   * @param b - Right-hand side vector (n × 1)\n   * @returns Solution vector x such that Ax = b\n   */\n  private solveLinearSystem(A: number[][], b: number[]): number[] {\n    const n = A.length;\n\n    // Create augmented matrix [A | b]\n    // This allows us to perform row operations on both A and b simultaneously\n    const aug = A.map((row, i) => [...row, b[i] ?? 0]);\n    let maxAbs = 0;\n    for (let i = 0; i < n; i++) {\n      const row = aug[i];\n      if (!row) continue;\n      for (let j = 0; j < n; j++) {\n        const v = Math.abs(row[j] ?? 0);\n        if (v > maxAbs) maxAbs = v;\n      }\n    }\n    if (maxAbs === 0 || !Number.isFinite(maxAbs)) {\n      throw new DataValidationError(\"Matrix is singular or ill-conditioned\");\n    }\n    const tol = Number.EPSILON * n * maxAbs;\n\n    // Forward elimination with partial pivoting\n    for (let i = 0; i < n; i++) {\n      // Find pivot: row with largest absolute value in column i\n      // This improves numerical stability by avoiding division by small numbers\n      let maxRow = i;\n      for (let k = i + 1; k < n; k++) {\n        if (Math.abs(aug[k]?.[i] ?? 0) > Math.abs(aug[maxRow]?.[i] ?? 0)) {\n          maxRow = k;\n        }\n      }\n\n      // Swap rows i and maxRow\n      const augI = aug[i] ?? [];\n      const augMax = aug[maxRow] ?? [];\n      aug[i] = augMax;\n      aug[maxRow] = augI;\n\n      const pivot = aug[i]?.[i] ?? 0;\n      if (!Number.isFinite(pivot) || Math.abs(pivot) <= tol) {\n        throw new DataValidationError(\"Matrix is singular or ill-conditioned\");\n      }\n\n      // Eliminate column i in rows below i\n      for (let k = i + 1; k < n; k++) {\n        // Compute multiplier: c = A[k,i] / A[i,i]\n        const c = (aug[k]?.[i] ?? 0) / pivot;\n        const augK = aug[k];\n\n        if (augK) {\n          // Subtract c * row_i from row_k\n          for (let j = i; j <= n; j++) {\n            augK[j] = (augK[j] ?? 0) - c * (aug[i]?.[j] ?? 0);\n          }\n        }\n      }\n    }\n\n    // Back substitution: solve upper triangular system\n    const x = Array(n).fill(0);\n    for (let i = n - 1; i >= 0; i--) {\n      // Start with b[i]\n      x[i] = aug[i]?.[n] ?? 0;\n\n      // Subtract contributions from already-solved variables\n      for (let j = i + 1; j < n; j++) {\n        x[i] = (x[i] ?? 0) - (aug[i]?.[j] ?? 0) * (x[j] ?? 0);\n      }\n\n      // Divide by diagonal element\n      const diag = aug[i]?.[i] ?? 0;\n      if (!Number.isFinite(diag) || Math.abs(diag) <= tol) {\n        throw new DataValidationError(\"Matrix is singular or ill-conditioned\");\n      }\n      x[i] = (x[i] ?? 0) / diag;\n    }\n\n    return x;\n  }\n\n  private solveConjugateGradient(\n    A: number[][],\n    b: number[],\n    maxIter: number,\n    tol: number\n  ): { x: number[]; nIter: number } {\n    const n = A.length;\n    const x = new Array<number>(n).fill(0);\n    const r = new Array<number>(n).fill(0);\n\n    let rsOld = 0;\n    for (let i = 0; i < n; i++) {\n      const bi = b[i] ?? 0;\n      r[i] = bi;\n      rsOld += bi * bi;\n    }\n\n    if (rsOld === 0) {\n      return { x, nIter: 0 };\n    }\n\n    const p = r.slice();\n    const tolSq = tol * tol;\n    let nIter = 0;\n\n    for (let iter = 0; iter < maxIter; iter++) {\n      const Ap = new Array<number>(n).fill(0);\n      for (let i = 0; i < n; i++) {\n        let sum = 0;\n        const row = A[i];\n        if (!row) continue;\n        for (let j = 0; j < n; j++) {\n          sum += (row[j] ?? 0) * (p[j] ?? 0);\n        }\n        Ap[i] = sum;\n      }\n\n      let denom = 0;\n      for (let i = 0; i < n; i++) {\n        denom += (p[i] ?? 0) * (Ap[i] ?? 0);\n      }\n      if (!Number.isFinite(denom) || denom === 0) {\n        throw new DataValidationError(\n          \"Conjugate gradient failed: denominator is zero or non-finite\"\n        );\n      }\n\n      const alpha = rsOld / denom;\n      for (let i = 0; i < n; i++) {\n        x[i] = (x[i] ?? 0) + alpha * (p[i] ?? 0);\n        r[i] = (r[i] ?? 0) - alpha * (Ap[i] ?? 0);\n      }\n\n      let rsNew = 0;\n      for (let i = 0; i < n; i++) {\n        const ri = r[i] ?? 0;\n        rsNew += ri * ri;\n      }\n      nIter = iter + 1;\n      if (rsNew < tolSq) {\n        break;\n      }\n\n      const beta = rsNew / rsOld;\n      for (let i = 0; i < n; i++) {\n        p[i] = (r[i] ?? 0) + beta * (p[i] ?? 0);\n      }\n      rsOld = rsNew;\n    }\n\n    return { x, nIter };\n  }\n\n  private solveSag(\n    getX: (sampleIndex: number, featureIndex: number) => number,\n    getY: (sampleIndex: number) => number,\n    nSamples: number,\n    nFeatures: number,\n    alpha: number,\n    maxIter: number,\n    tol: number\n  ): { x: number[]; nIter: number } {\n    const w = new Array<number>(nFeatures).fill(0);\n    const avgGrad = new Array<number>(nFeatures).fill(0);\n    const residuals = new Array<number>(nSamples).fill(0);\n\n    let maxNormSq = 0;\n    for (let i = 0; i < nSamples; i++) {\n      let normSq = 0;\n      for (let j = 0; j < nFeatures; j++) {\n        const xij = getX(i, j);\n        normSq += xij * xij;\n      }\n      if (normSq > maxNormSq) {\n        maxNormSq = normSq;\n      }\n    }\n\n    const scale = nSamples === 0 ? 1 : nSamples;\n    const L = maxNormSq * scale + alpha;\n    const step = L > 0 ? 1 / L : 1;\n\n    let nIter = 0;\n    for (let iter = 0; iter < maxIter; iter++) {\n      let maxUpdate = 0;\n\n      for (let i = 0; i < nSamples; i++) {\n        let dotProd = 0;\n        for (let j = 0; j < nFeatures; j++) {\n          dotProd += (w[j] ?? 0) * getX(i, j);\n        }\n\n        const yi = getY(i);\n        const newResidual = dotProd - yi;\n        const delta = newResidual - (residuals[i] ?? 0);\n        residuals[i] = newResidual;\n\n        if (delta !== 0) {\n          for (let j = 0; j < nFeatures; j++) {\n            avgGrad[j] = (avgGrad[j] ?? 0) + delta * getX(i, j);\n          }\n        }\n\n        for (let j = 0; j < nFeatures; j++) {\n          const grad = (avgGrad[j] ?? 0) + alpha * (w[j] ?? 0);\n          const update = step * grad;\n          w[j] = (w[j] ?? 0) - update;\n          if (Math.abs(update) > maxUpdate) {\n            maxUpdate = Math.abs(update);\n          }\n        }\n      }\n\n      nIter = iter + 1;\n      if (maxUpdate < tol) {\n        break;\n      }\n    }\n\n    return { x: w, nIter };\n  }\n\n  private rescaleCoefs(coef: Tensor, scale: number[]): Tensor {\n    const nFeatures = coef.shape[0] ?? 0;\n    const result: number[] = [];\n    for (let j = 0; j < nFeatures; j++) {\n      const c = Number(coef.data[coef.offset + j] ?? 0);\n      const s = scale[j] ?? 1;\n      result.push(s === 0 ? 0 : c / s);\n    }\n    return tensor(result);\n  }\n\n  /**\n   * Predict using the Ridge regression model.\n   *\n   * Computes predictions as: ŷ = X @ coef + intercept\n   *\n   * **Time Complexity**: O(nm) where n = samples, m = features\n   * **Space Complexity**: O(n)\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    // Check if model has been fitted\n    if (!this.fitted || !this.coef_) {\n      throw new NotFittedError(\"Ridge must be fitted before prediction\");\n    }\n\n    // Validate input\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"Ridge\");\n\n    const m = X.shape[0] ?? 0; // Number of samples to predict\n    const n = X.shape[1] ?? 0; // Number of features\n    const pred = Array(m).fill(0);\n\n    // Compute predictions: ŷ[i] = Σ_j X[i,j] * coef[j] + intercept\n    for (let i = 0; i < m; i++) {\n      let sum = this.intercept_ ?? 0; // Start with intercept\n\n      // Add weighted sum of features\n      for (let j = 0; j < n; j++) {\n        sum +=\n          Number(X.data[X.offset + i * n + j] ?? 0) *\n          Number(this.coef_.data[this.coef_.offset + j] ?? 0);\n      }\n      pred[i] = sum;\n    }\n\n    return tensor(pred);\n  }\n\n  /**\n   * Return the coefficient of determination R² of the prediction.\n   *\n   * R² (R-squared) measures the proportion of variance in y explained by the model.\n   * Formula: R² = 1 - (SS_res / SS_tot)\n   *\n   * Where:\n   * - SS_res = Σ(y_true - y_pred)² (residual sum of squares)\n   * - SS_tot = Σ(y_true - y_mean)² (total sum of squares)\n   *\n   * **Interpretation**:\n   * - R² = 1: Perfect predictions\n   * - R² = 0: Model performs as well as predicting the mean\n   * - R² < 0: Model performs worse than predicting the mean\n   *\n   * **Time Complexity**: O(n) where n = number of samples\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True values of shape (n_samples,)\n   * @returns R² score (best possible score is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional\n   */\n  score(X: Tensor, y: Tensor): number {\n    // Check if model has been fitted\n    if (!this.fitted) {\n      throw new NotFittedError(\"Ridge must be fitted before scoring\");\n    }\n\n    // Validate y dimensions\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n\n    // Get predictions\n    const pred = this.predict(X);\n    if (pred.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${pred.size}, y=${y.size}`\n      );\n    }\n\n    let ssRes = 0; // Residual sum of squares\n    let ssTot = 0; // Total sum of squares\n    let yMean = 0; // Mean of y\n\n    // Compute mean of y\n    for (let i = 0; i < y.size; i++) {\n      yMean += Number(y.data[y.offset + i] ?? 0);\n    }\n    yMean /= y.size;\n\n    // Compute SS_res and SS_tot\n    for (let i = 0; i < y.size; i++) {\n      const yVal = Number(y.data[y.offset + i] ?? 0);\n      const predVal = Number(pred.data[pred.offset + i] ?? 0);\n\n      // Residual sum of squares: measures prediction error\n      ssRes += (yVal - predVal) ** 2;\n\n      // Total sum of squares: measures total variance in y\n      ssTot += (yVal - yMean) ** 2;\n    }\n\n    // Handle edge case: constant y (zero variance)\n    // If y is constant and predictions match, R² = 1\n    // If y is constant but predictions don't match, R² = 0\n    if (ssTot === 0) {\n      return ssRes === 0 ? 1.0 : 0.0;\n    }\n\n    // Compute R² = 1 - (SS_res / SS_tot)\n    return 1 - ssRes / ssTot;\n  }\n\n  /**\n   * Get parameters for this estimator.\n   *\n   * Returns a copy of all hyperparameters set during construction or via setParams.\n   *\n   * @returns Object containing all parameters with their current values\n   */\n  getParams(): Record<string, unknown> {\n    return { ...this.options };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * Allows modifying hyperparameters after construction.\n   * Note: Changing parameters requires refitting the model.\n   *\n   * @param params - Dictionary of parameters to set\n   * @returns this - The estimator for method chaining\n   * @throws {TypeError} If parameter value has wrong type\n   * @throws {Error} If parameter name is unknown or value is invalid\n   */\n  setParams(params: Record<string, unknown>): this {\n    for (const [key, value] of Object.entries(params)) {\n      switch (key) {\n        case \"alpha\":\n          if (typeof value !== \"number\" || !Number.isFinite(value)) {\n            throw new InvalidParameterError(\n              `alpha must be a finite number; received ${String(value)}`,\n              \"alpha\",\n              value\n            );\n          }\n          this.options.alpha = value;\n          break;\n\n        case \"maxIter\":\n          if (typeof value !== \"number\" || !Number.isFinite(value)) {\n            throw new InvalidParameterError(\n              `maxIter must be a finite number; received ${String(value)}`,\n              \"maxIter\",\n              value\n            );\n          }\n          this.options.maxIter = value;\n          break;\n\n        case \"tol\":\n          if (typeof value !== \"number\" || !Number.isFinite(value)) {\n            throw new InvalidParameterError(\n              `tol must be a finite number; received ${String(value)}`,\n              \"tol\",\n              value\n            );\n          }\n          this.options.tol = value;\n          break;\n\n        case \"fitIntercept\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\n              `fitIntercept must be a boolean; received ${String(value)}`,\n              \"fitIntercept\",\n              value\n            );\n          }\n          this.options.fitIntercept = value;\n          break;\n\n        case \"normalize\":\n          if (typeof value !== \"boolean\") {\n            throw new InvalidParameterError(\n              `normalize must be a boolean; received ${String(value)}`,\n              \"normalize\",\n              value\n            );\n          }\n          this.options.normalize = value;\n          break;\n\n        case \"solver\":\n          if (\n            value !== \"auto\" &&\n            value !== \"svd\" &&\n            value !== \"cholesky\" &&\n            value !== \"lsqr\" &&\n            value !== \"sag\"\n          ) {\n            throw new InvalidParameterError(`Invalid solver: ${String(value)}`, \"solver\", value);\n          }\n          this.options.solver = value;\n          break;\n\n        default:\n          throw new InvalidParameterError(`Unknown parameter: ${key}`, key, value);\n      }\n    }\n    return this;\n  }\n\n  /**\n   * Get the model coefficients (weights).\n   *\n   * @returns Coefficient tensor of shape (n_features,)\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get coef(): Tensor {\n    if (!this.fitted || !this.coef_) {\n      throw new NotFittedError(\"Ridge must be fitted to access coefficients\");\n    }\n    return this.coef_;\n  }\n\n  /**\n   * Get the intercept (bias term).\n   *\n   * @returns Intercept value\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get intercept(): number {\n    if (!this.fitted) {\n      throw new NotFittedError(\"Ridge must be fitted to access intercept\");\n    }\n    return this.intercept_ ?? 0;\n  }\n\n  /**\n   * Get the number of iterations run by the solver.\n   *\n   * @returns Number of iterations (undefined for direct solvers)\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get nIter(): number | undefined {\n    if (!this.fitted) {\n      throw new NotFittedError(\"Ridge must be fitted to access nIter\");\n    }\n    return this.nIter_;\n  }\n}\n","import { InvalidParameterError, NotFittedError } from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { validateUnsupervisedFitInputs } from \"../_validation\";\n\ntype SparseRow = { indices: number[]; values: number[] };\ntype SparseMatrix = SparseRow[];\ntype SampleRow = { indices: number[]; qValues: number[] };\n\n/**\n * t-Distributed Stochastic Neighbor Embedding (t-SNE).\n *\n * A nonlinear dimensionality reduction technique for embedding high-dimensional\n * data into a low-dimensional space (typically 2D or 3D) for visualization.\n *\n * **Algorithm**: Exact t-SNE with an optional sampling-based approximation\n * - Computes pairwise affinities in high-dimensional space using Gaussian kernel (exact)\n * - Computes pairwise affinities in low-dimensional space using Student-t distribution\n * - Minimizes KL divergence between the two distributions\n *\n * **Scalability Note**:\n * Exact t-SNE is O(n^2) in time and memory. For large datasets, use\n * `method: \"approximate\"` (sampled neighbors + negative sampling) or reduce samples.\n *\n * @example\n * ```ts\n * import { TSNE } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1, 2, 3], [4, 5, 6], [7, 8, 9], [10, 11, 12]]);\n *\n * const tsne = new TSNE({ nComponents: 2, perplexity: 5 });\n * const embedding = tsne.fitTransform(X);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.manifold.TSNE.html | scikit-learn TSNE}\n * @see van der Maaten, L.J.P.; Hinton, G.E. (2008). \"Visualizing High-Dimensional Data Using t-SNE\"\n */\nexport class TSNE {\n  /** Number of dimensions in the embedding */\n  private readonly nComponents: number;\n\n  /** Perplexity parameter (related to number of nearest neighbors) */\n  private readonly perplexity: number;\n\n  /** Learning rate for gradient descent */\n  private readonly learningRate: number;\n\n  /** Number of iterations */\n  private readonly nIter: number;\n\n  /** Early exaggeration factor */\n  private readonly earlyExaggeration: number;\n\n  /** Number of iterations with early exaggeration */\n  private readonly earlyExaggerationIter: number;\n\n  /** Random seed for reproducibility */\n  private readonly randomState: number | undefined;\n\n  /** Minimum gradient norm for convergence */\n  private readonly minGradNorm: number;\n\n  /** Method for computing affinities and gradients */\n  private readonly method: \"exact\" | \"approximate\";\n\n  /** Maximum samples allowed for exact mode */\n  private readonly maxExactSamples: number;\n\n  /** Neighbor count for approximate mode */\n  private readonly approximateNeighbors: number;\n\n  /** Negative samples per point for approximate mode */\n  private readonly negativeSamples: number;\n\n  /** The embedded points after fitting */\n  private embedding: number[][] = [];\n\n  /** Whether the model has been fitted */\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly nComponents?: number;\n      readonly perplexity?: number;\n      readonly learningRate?: number;\n      readonly nIter?: number;\n      readonly earlyExaggeration?: number;\n      readonly earlyExaggerationIter?: number;\n      readonly randomState?: number;\n      readonly minGradNorm?: number;\n      /** \"exact\" uses full pairwise interactions; \"approximate\" uses sampling for large datasets. */\n      readonly method?: \"exact\" | \"approximate\";\n      /** Maximum samples allowed in exact mode before requiring approximate. */\n      readonly maxExactSamples?: number;\n      /** Number of neighbors to sample per point in approximate mode. */\n      readonly approximateNeighbors?: number;\n      /** Number of negative samples per point in approximate mode. */\n      readonly negativeSamples?: number;\n    } = {}\n  ) {\n    this.nComponents = options.nComponents ?? 2;\n    this.perplexity = options.perplexity ?? 30;\n    this.learningRate = options.learningRate ?? 200;\n    this.nIter = options.nIter ?? 1000;\n    this.earlyExaggeration = options.earlyExaggeration ?? 12;\n    const earlyExaggerationIter = options.earlyExaggerationIter ?? 250;\n    this.earlyExaggerationIter = Math.min(earlyExaggerationIter, this.nIter);\n    this.randomState = options.randomState;\n    this.minGradNorm = options.minGradNorm ?? 1e-7;\n    this.method = options.method ?? \"exact\";\n    this.maxExactSamples = options.maxExactSamples ?? 2000;\n    this.approximateNeighbors =\n      options.approximateNeighbors ?? Math.max(5, Math.floor(this.perplexity * 3));\n    this.negativeSamples = options.negativeSamples ?? Math.max(10, Math.floor(this.perplexity * 2));\n\n    if (!Number.isInteger(this.nComponents) || this.nComponents <= 0) {\n      throw new InvalidParameterError(\n        \"nComponents must be positive\",\n        \"nComponents\",\n        this.nComponents\n      );\n    }\n    if (!Number.isFinite(this.perplexity) || this.perplexity <= 0) {\n      throw new InvalidParameterError(\"perplexity must be positive\", \"perplexity\", this.perplexity);\n    }\n    if (!Number.isFinite(this.learningRate) || this.learningRate <= 0) {\n      throw new InvalidParameterError(\n        \"learningRate must be positive\",\n        \"learningRate\",\n        this.learningRate\n      );\n    }\n    if (!Number.isInteger(this.nIter) || this.nIter <= 0) {\n      throw new InvalidParameterError(\"nIter must be a positive integer\", \"nIter\", this.nIter);\n    }\n    if (!Number.isFinite(this.earlyExaggeration) || this.earlyExaggeration <= 0) {\n      throw new InvalidParameterError(\n        \"earlyExaggeration must be positive\",\n        \"earlyExaggeration\",\n        this.earlyExaggeration\n      );\n    }\n    if (!Number.isInteger(earlyExaggerationIter) || earlyExaggerationIter < 0) {\n      throw new InvalidParameterError(\n        \"earlyExaggerationIter must be an integer >= 0\",\n        \"earlyExaggerationIter\",\n        earlyExaggerationIter\n      );\n    }\n    if (!Number.isFinite(this.minGradNorm) || this.minGradNorm <= 0) {\n      throw new InvalidParameterError(\n        \"minGradNorm must be positive\",\n        \"minGradNorm\",\n        this.minGradNorm\n      );\n    }\n    if (this.method !== \"exact\" && this.method !== \"approximate\") {\n      throw new InvalidParameterError(\n        \"method must be 'exact' or 'approximate'\",\n        \"method\",\n        this.method\n      );\n    }\n    if (!Number.isInteger(this.maxExactSamples) || this.maxExactSamples <= 0) {\n      throw new InvalidParameterError(\n        \"maxExactSamples must be a positive integer\",\n        \"maxExactSamples\",\n        this.maxExactSamples\n      );\n    }\n    if (!Number.isInteger(this.approximateNeighbors) || this.approximateNeighbors <= 0) {\n      throw new InvalidParameterError(\n        \"approximateNeighbors must be a positive integer\",\n        \"approximateNeighbors\",\n        this.approximateNeighbors\n      );\n    }\n    if (!Number.isInteger(this.negativeSamples) || this.negativeSamples <= 0) {\n      throw new InvalidParameterError(\n        \"negativeSamples must be a positive integer\",\n        \"negativeSamples\",\n        this.negativeSamples\n      );\n    }\n    if (options.randomState !== undefined && !Number.isFinite(options.randomState)) {\n      throw new InvalidParameterError(\n        \"randomState must be a finite number\",\n        \"randomState\",\n        options.randomState\n      );\n    }\n  }\n\n  /**\n   * Compute pairwise squared Euclidean distances.\n   */\n  private computeDistances(X: number[][]): number[][] {\n    const n = X.length;\n    const distances: number[][] = [];\n\n    for (let i = 0; i < n; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < n; j++) {\n        if (i === j) {\n          row.push(0);\n        } else {\n          let dist = 0;\n          const xi = X[i];\n          const xj = X[j];\n          if (xi && xj) {\n            for (let k = 0; k < xi.length; k++) {\n              const diff = (xi[k] ?? 0) - (xj[k] ?? 0);\n              dist += diff * diff;\n            }\n          }\n          row.push(dist);\n        }\n      }\n      distances.push(row);\n    }\n\n    return distances;\n  }\n\n  /**\n   * Compute squared Euclidean distance between two vectors.\n   */\n  private computeSquaredDistance(a: number[], b: number[]): number {\n    let dist = 0;\n    for (let k = 0; k < a.length; k++) {\n      const diff = (a[k] ?? 0) - (b[k] ?? 0);\n      dist += diff * diff;\n    }\n    return dist;\n  }\n\n  /**\n   * Sample unique indices excluding a single index.\n   */\n  private sampleIndices(n: number, exclude: number, k: number, rng: () => number): number[] {\n    const result = new Set<number>();\n    const maxAttempts = k * 10 + 100;\n    let attempts = 0;\n\n    while (result.size < k && attempts < maxAttempts) {\n      const raw = Math.floor(rng() * (n - 1));\n      const idx = raw >= exclude ? raw + 1 : raw;\n      result.add(idx);\n      attempts += 1;\n    }\n\n    if (result.size < k) {\n      for (let i = 0; i < n && result.size < k; i++) {\n        if (i !== exclude) {\n          result.add(i);\n        }\n      }\n    }\n\n    return Array.from(result);\n  }\n\n  /**\n   * Sample unique indices excluding a set of indices.\n   */\n  private sampleIndicesWithExclusions(\n    n: number,\n    exclusions: ReadonlySet<number>,\n    k: number,\n    rng: () => number\n  ): number[] {\n    const result = new Set<number>();\n    const maxAttempts = k * 12 + 200;\n    let attempts = 0;\n\n    while (result.size < k && attempts < maxAttempts) {\n      const idx = Math.floor(rng() * n);\n      if (!exclusions.has(idx)) {\n        result.add(idx);\n      }\n      attempts += 1;\n    }\n\n    if (result.size < k) {\n      for (let i = 0; i < n && result.size < k; i++) {\n        if (!exclusions.has(i)) {\n          result.add(i);\n        }\n      }\n    }\n\n    return Array.from(result);\n  }\n\n  /**\n   * Compute conditional probabilities P(j|i) using binary search for sigma.\n   */\n  private computeProbabilities(distances: number[][]): number[][] {\n    const n = distances.length;\n    const targetEntropy = Math.log(this.perplexity);\n    const P: number[][] = [];\n\n    for (let i = 0; i < n; i++) {\n      const row: number[] = new Array(n).fill(0);\n\n      // Binary search for sigma\n      let sigmaMin = 1e-10;\n      let sigmaMax = 1e10;\n      let sigma = 1.0;\n\n      for (let iter = 0; iter < 50; iter++) {\n        // Compute probabilities with current sigma\n        let sumExp = 0;\n        for (let j = 0; j < n; j++) {\n          if (i !== j) {\n            const distRow = distances[i];\n            const dist = distRow ? (distRow[j] ?? 0) : 0;\n            sumExp += Math.exp(-dist / (2 * sigma * sigma));\n          }\n        }\n\n        // Compute entropy\n        let entropy = 0;\n        for (let j = 0; j < n; j++) {\n          if (i !== j) {\n            const distRow = distances[i];\n            const dist = distRow ? (distRow[j] ?? 0) : 0;\n            const pij = Math.exp(-dist / (2 * sigma * sigma)) / (sumExp + 1e-10);\n            if (pij > 1e-10) {\n              entropy -= pij * Math.log(pij);\n            }\n          }\n        }\n\n        // Binary search adjustment\n        if (Math.abs(entropy - targetEntropy) < 1e-5) {\n          break;\n        }\n\n        if (entropy > targetEntropy) {\n          sigmaMax = sigma;\n          sigma = (sigma + sigmaMin) / 2;\n        } else {\n          sigmaMin = sigma;\n          sigma = (sigma + sigmaMax) / 2;\n        }\n      }\n\n      // Compute final probabilities\n      let sumExp = 0;\n      for (let j = 0; j < n; j++) {\n        if (i !== j) {\n          const distRow = distances[i];\n          const dist = distRow ? (distRow[j] ?? 0) : 0;\n          sumExp += Math.exp(-dist / (2 * sigma * sigma));\n        }\n      }\n\n      for (let j = 0; j < n; j++) {\n        if (i !== j) {\n          const distRow = distances[i];\n          const dist = distRow ? (distRow[j] ?? 0) : 0;\n          row[j] = Math.exp(-dist / (2 * sigma * sigma)) / (sumExp + 1e-10);\n        }\n      }\n\n      P.push(row);\n    }\n\n    // Symmetrize: P = (P + P') / (2n)\n    const Psym: number[][] = [];\n    for (let i = 0; i < n; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < n; j++) {\n        const pij = P[i]?.[j] ?? 0;\n        const pji = P[j]?.[i] ?? 0;\n        row.push((pij + pji) / (2 * n));\n      }\n      Psym.push(row);\n    }\n\n    return Psym;\n  }\n\n  /**\n   * Compute sparse joint probabilities using sampled neighbors (approximate).\n   */\n  private computeProbabilitiesSparse(\n    X: number[][],\n    neighbors: number,\n    rng: () => number\n  ): SparseMatrix {\n    const n = X.length;\n    const targetEntropy = Math.log(this.perplexity);\n    const rows: SparseMatrix = [];\n\n    for (let i = 0; i < n; i++) {\n      const neighborIndices = this.sampleIndices(n, i, neighbors, rng);\n      const distances = neighborIndices.map((j) =>\n        this.computeSquaredDistance(X[i] ?? [], X[j] ?? [])\n      );\n\n      let sigmaMin = 1e-10;\n      let sigmaMax = 1e10;\n      let sigma = 1.0;\n\n      for (let iter = 0; iter < 50; iter++) {\n        let sumExp = 0;\n        for (const dist of distances) {\n          sumExp += Math.exp(-dist / (2 * sigma * sigma));\n        }\n\n        let entropy = 0;\n        for (const dist of distances) {\n          const pij = Math.exp(-dist / (2 * sigma * sigma)) / (sumExp + 1e-10);\n          if (pij > 1e-10) {\n            entropy -= pij * Math.log(pij);\n          }\n        }\n\n        if (Math.abs(entropy - targetEntropy) < 1e-5) {\n          break;\n        }\n\n        if (entropy > targetEntropy) {\n          sigmaMax = sigma;\n          sigma = (sigma + sigmaMin) / 2;\n        } else {\n          sigmaMin = sigma;\n          sigma = (sigma + sigmaMax) / 2;\n        }\n      }\n\n      let sumExp = 0;\n      for (const dist of distances) {\n        sumExp += Math.exp(-dist / (2 * sigma * sigma));\n      }\n\n      const values: number[] = distances.map(\n        (dist) => Math.exp(-dist / (2 * sigma * sigma)) / (sumExp + 1e-10)\n      );\n\n      rows.push({ indices: neighborIndices, values });\n    }\n\n    return this.symmetrizeSparse(rows, n);\n  }\n\n  /**\n   * Symmetrize sparse probabilities: P = (P + P^T) / (2n).\n   */\n  private symmetrizeSparse(rows: SparseMatrix, n: number): SparseMatrix {\n    const maps: Array<Map<number, number>> = rows.map((row) => {\n      const map = new Map<number, number>();\n      for (let k = 0; k < row.indices.length; k++) {\n        const j = row.indices[k];\n        if (j === undefined) continue;\n        const val = row.values[k] ?? 0;\n        map.set(j, val);\n      }\n      return map;\n    });\n\n    const sym: SparseMatrix = [];\n    for (let i = 0; i < n; i++) {\n      const row = rows[i];\n      if (!row) {\n        sym.push({ indices: [], values: [] });\n        continue;\n      }\n      const indices = row.indices.slice();\n      const values: number[] = [];\n      for (let k = 0; k < indices.length; k++) {\n        const j = indices[k];\n        if (j === undefined) continue;\n        const pij = maps[i]?.get(j) ?? 0;\n        const pji = maps[j]?.get(i) ?? 0;\n        values.push((pij + pji) / (2 * n));\n      }\n      sym.push({ indices, values });\n    }\n\n    return sym;\n  }\n\n  /**\n   * Initialize embedding with small random values.\n   */\n  private initializeEmbedding(n: number): number[][] {\n    const embedding: number[][] = [];\n\n    // Simple LCG random number generator for reproducibility\n    let seed = this.randomState ?? Date.now();\n    const random = (): number => {\n      seed = (seed * 1103515245 + 12345) & 0x7fffffff;\n      return seed / 0x7fffffff;\n    };\n\n    for (let i = 0; i < n; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < this.nComponents; j++) {\n        // Initialize with small random values (normal-ish distribution)\n        const u1 = random() || 0.0001;\n        const u2 = random();\n        const z = Math.sqrt(-2 * Math.log(u1)) * Math.cos(2 * Math.PI * u2);\n        row.push(z * 0.0001);\n      }\n      embedding.push(row);\n    }\n\n    return embedding;\n  }\n\n  /**\n   * Compute Q distribution (Student-t with 1 degree of freedom).\n   */\n  private computeQ(Y: number[][]): { Q: number[][]; sumQ: number } {\n    const n = Y.length;\n    const Q: number[][] = [];\n    let sumQ = 0;\n\n    for (let i = 0; i < n; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < n; j++) {\n        if (i === j) {\n          row.push(0);\n        } else {\n          let dist = 0;\n          const yi = Y[i];\n          const yj = Y[j];\n          if (yi && yj) {\n            for (let k = 0; k < this.nComponents; k++) {\n              const diff = (yi[k] ?? 0) - (yj[k] ?? 0);\n              dist += diff * diff;\n            }\n          }\n          // Student-t distribution with 1 degree of freedom\n          const qij = 1 / (1 + dist);\n          row.push(qij);\n          sumQ += qij;\n        }\n      }\n      Q.push(row);\n    }\n\n    return { Q, sumQ };\n  }\n\n  /**\n   * Compute approximate Q using sampled pairs.\n   */\n  private computeQApprox(\n    Y: number[][],\n    neighbors: SparseMatrix,\n    negativeSamples: number,\n    rng: () => number\n  ): { rows: SampleRow[]; sumQ: number } {\n    const n = Y.length;\n    const rows: SampleRow[] = [];\n    let sumQ = 0;\n\n    for (let i = 0; i < n; i++) {\n      const neighborIndices = neighbors[i]?.indices ?? [];\n      const exclusion = new Set<number>(neighborIndices);\n      exclusion.add(i);\n\n      const negatives = this.sampleIndicesWithExclusions(n, exclusion, negativeSamples, rng);\n      const indices = neighborIndices.concat(negatives);\n      const qValues: number[] = [];\n\n      const yi = Y[i] ?? [];\n      for (let k = 0; k < indices.length; k++) {\n        const j = indices[k] ?? 0;\n        const yj = Y[j] ?? [];\n        const dist = this.computeSquaredDistance(yi, yj);\n        const qij = 1 / (1 + dist);\n        qValues.push(qij);\n        sumQ += qij;\n      }\n\n      rows.push({ indices, qValues });\n    }\n\n    return { rows, sumQ };\n  }\n\n  /**\n   * Compute gradients of KL divergence.\n   */\n  private computeGradients(P: number[][], Q: number[][], sumQ: number, Y: number[][]): number[][] {\n    const n = Y.length;\n    const gradients: number[][] = [];\n\n    for (let i = 0; i < n; i++) {\n      const grad: number[] = new Array(this.nComponents).fill(0);\n\n      for (let j = 0; j < n; j++) {\n        if (i !== j) {\n          const pij = P[i]?.[j] ?? 0;\n          const qij = (Q[i]?.[j] ?? 0) / (sumQ + 1e-10);\n\n          const yi = Y[i];\n          const yj = Y[j];\n          if (yi && yj) {\n            // Compute (1 + ||y_i - y_j||^2)^-1\n            let dist = 0;\n            for (let k = 0; k < this.nComponents; k++) {\n              const diff = (yi[k] ?? 0) - (yj[k] ?? 0);\n              dist += diff * diff;\n            }\n            const mult = (pij - qij) * (1 / (1 + dist));\n\n            for (let k = 0; k < this.nComponents; k++) {\n              grad[k] = (grad[k] ?? 0) + 4 * mult * ((yi[k] ?? 0) - (yj[k] ?? 0));\n            }\n          }\n        }\n      }\n\n      gradients.push(grad);\n    }\n\n    return gradients;\n  }\n\n  /**\n   * Compute approximate gradients of KL divergence using sampled pairs.\n   */\n  private computeGradientsApprox(\n    PMaps: Array<Map<number, number>>,\n    QRows: SampleRow[],\n    sumQ: number,\n    Y: number[][],\n    exaggeration: number\n  ): number[][] {\n    const n = Y.length;\n    const gradients: number[][] = [];\n    const denom = sumQ + 1e-10;\n\n    for (let i = 0; i < n; i++) {\n      const grad: number[] = new Array(this.nComponents).fill(0);\n      const yi = Y[i] ?? [];\n      const pMap = PMaps[i];\n      const qRow = QRows[i];\n\n      if (qRow) {\n        for (let idx = 0; idx < qRow.indices.length; idx++) {\n          const j = qRow.indices[idx] ?? 0;\n          if (i === j) continue;\n          const yj = Y[j] ?? [];\n\n          const pijBase = pMap?.get(j) ?? 0;\n          const pij = pijBase * exaggeration;\n          const qUnnorm = qRow.qValues[idx] ?? 0;\n          const qij = qUnnorm / denom;\n\n          const mult = (pij - qij) * qUnnorm;\n          for (let k = 0; k < this.nComponents; k++) {\n            grad[k] = (grad[k] ?? 0) + 4 * mult * ((yi[k] ?? 0) - (yj[k] ?? 0));\n          }\n        }\n      }\n\n      gradients.push(grad);\n    }\n\n    return gradients;\n  }\n\n  /**\n   * Fit the t-SNE model and return the embedding.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @returns Low-dimensional embedding of shape (n_samples, n_components)\n   */\n  fitTransform(X: Tensor): Tensor {\n    validateUnsupervisedFitInputs(X);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    if (nSamples < 4) {\n      throw new InvalidParameterError(\"t-SNE requires at least 4 samples\", \"nSamples\", nSamples);\n    }\n    if (this.perplexity >= nSamples) {\n      throw new InvalidParameterError(\n        `perplexity must be less than n_samples; received perplexity=${this.perplexity}, n_samples=${nSamples}`,\n        \"perplexity\",\n        this.perplexity\n      );\n    }\n    if (this.method === \"exact\" && nSamples > this.maxExactSamples) {\n      throw new InvalidParameterError(\n        `Exact t-SNE is O(n^2) and limited to n_samples <= ${this.maxExactSamples}; received n_samples=${nSamples}. Use method=\"approximate\" or increase maxExactSamples.`,\n        \"nSamples\",\n        nSamples\n      );\n    }\n\n    // Extract data\n    const XData: number[][] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      XData.push(row);\n    }\n\n    const baseSeed = this.randomState ?? Date.now();\n    const rng = (() => {\n      let state = baseSeed >>> 0;\n      return (): number => {\n        state = (state * 1664525 + 1013904223) >>> 0;\n        return state / 2 ** 32;\n      };\n    })();\n\n    const useApproximate = this.method === \"approximate\";\n    const neighborCount = Math.min(this.approximateNeighbors, nSamples - 1);\n    const availableNegatives = Math.max(0, nSamples - 1 - neighborCount);\n    const negativeCount = Math.min(this.negativeSamples, availableNegatives);\n    if (useApproximate && neighborCount < 2) {\n      throw new InvalidParameterError(\n        \"approximateNeighbors must be at least 2 for approximate mode\",\n        \"approximateNeighbors\",\n        neighborCount\n      );\n    }\n    if (useApproximate && this.perplexity >= neighborCount) {\n      throw new InvalidParameterError(\n        `perplexity must be less than approximateNeighbors; received perplexity=${this.perplexity}, approximateNeighbors=${neighborCount}`,\n        \"perplexity\",\n        this.perplexity\n      );\n    }\n\n    // Compute joint probabilities P\n    const PExact = useApproximate ? null : this.computeProbabilities(this.computeDistances(XData));\n    const PSparse = useApproximate\n      ? this.computeProbabilitiesSparse(XData, neighborCount, rng)\n      : null;\n\n    // Initialize embedding\n    const Y = this.initializeEmbedding(nSamples);\n\n    // Momentum terms\n    const velocities: number[][] = [];\n    for (let i = 0; i < nSamples; i++) {\n      velocities.push(new Array(this.nComponents).fill(0));\n    }\n\n    const momentum = 0.5;\n    const finalMomentum = 0.8;\n\n    const PMaps =\n      useApproximate && PSparse\n        ? PSparse.map((row) => {\n            const map = new Map<number, number>();\n            for (let k = 0; k < row.indices.length; k++) {\n              const j = row.indices[k];\n              if (j === undefined) continue;\n              const val = row.values[k] ?? 0;\n              map.set(j, val);\n            }\n            return map;\n          })\n        : [];\n\n    // Gradient descent\n    for (let iter = 0; iter < this.nIter; iter++) {\n      // Early exaggeration\n      const exaggeration = iter < this.earlyExaggerationIter ? this.earlyExaggeration : 1;\n\n      let gradients: number[][];\n      if (useApproximate && PSparse) {\n        const { rows: QRows, sumQ } = this.computeQApprox(Y, PSparse, negativeCount, rng);\n        gradients = this.computeGradientsApprox(PMaps, QRows, sumQ, Y, exaggeration);\n      } else {\n        // Apply exaggeration to P\n        const Pexag: number[][] = [];\n        for (let i = 0; i < nSamples; i++) {\n          const row: number[] = [];\n          for (let j = 0; j < nSamples; j++) {\n            row.push((PExact?.[i]?.[j] ?? 0) * exaggeration);\n          }\n          Pexag.push(row);\n        }\n\n        // Compute Q\n        const { Q, sumQ } = this.computeQ(Y);\n\n        // Compute gradients\n        gradients = this.computeGradients(Pexag, Q, sumQ, Y);\n      }\n\n      // Check convergence\n      let gradNorm = 0;\n      for (let i = 0; i < nSamples; i++) {\n        const grad = gradients[i];\n        if (grad) {\n          for (let k = 0; k < this.nComponents; k++) {\n            gradNorm += (grad[k] ?? 0) ** 2;\n          }\n        }\n      }\n      gradNorm = Math.sqrt(gradNorm);\n\n      if (gradNorm < this.minGradNorm) {\n        break;\n      }\n\n      // Update momentum\n      const currentMomentum = iter < this.earlyExaggerationIter ? momentum : finalMomentum;\n\n      // Update embedding with momentum\n      for (let i = 0; i < nSamples; i++) {\n        const yi = Y[i];\n        const grad = gradients[i];\n        const vel = velocities[i];\n\n        if (yi && grad && vel) {\n          for (let k = 0; k < this.nComponents; k++) {\n            vel[k] = currentMomentum * (vel[k] ?? 0) - this.learningRate * (grad[k] ?? 0);\n            yi[k] = (yi[k] ?? 0) + (vel[k] ?? 0);\n          }\n        }\n      }\n\n      // Center the embedding\n      const center: number[] = new Array(this.nComponents).fill(0);\n      for (let i = 0; i < nSamples; i++) {\n        const yi = Y[i];\n        if (yi) {\n          for (let k = 0; k < this.nComponents; k++) {\n            center[k] = (center[k] ?? 0) + (yi[k] ?? 0);\n          }\n        }\n      }\n      for (let k = 0; k < this.nComponents; k++) {\n        center[k] = (center[k] ?? 0) / nSamples;\n      }\n      for (let i = 0; i < nSamples; i++) {\n        const yi = Y[i];\n        if (yi) {\n          for (let k = 0; k < this.nComponents; k++) {\n            yi[k] = (yi[k] ?? 0) - (center[k] ?? 0);\n          }\n        }\n      }\n    }\n\n    this.embedding = Y;\n    this.fitted = true;\n\n    return tensor(Y);\n  }\n\n  /**\n   * Fit the model (same as fitTransform for t-SNE).\n   */\n  fit(X: Tensor): this {\n    this.fitTransform(X);\n    return this;\n  }\n\n  /**\n   * Get the embedding.\n   */\n  get embeddingResult(): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"TSNE must be fitted before accessing embedding\");\n    }\n    return tensor(this.embedding);\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nComponents: this.nComponents,\n      perplexity: this.perplexity,\n      learningRate: this.learningRate,\n      nIter: this.nIter,\n      earlyExaggeration: this.earlyExaggeration,\n      earlyExaggerationIter: this.earlyExaggerationIter,\n      randomState: this.randomState,\n      minGradNorm: this.minGradNorm,\n      method: this.method,\n      maxExactSamples: this.maxExactSamples,\n      approximateNeighbors: this.approximateNeighbors,\n      negativeSamples: this.negativeSamples,\n    };\n  }\n}\n","import {\n  DataValidationError,\n  InvalidParameterError,\n  NotFittedError,\n  NotImplementedError,\n  ShapeError,\n} from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Classifier } from \"../base\";\n\n/**\n * Gaussian Naive Bayes classifier.\n *\n * Implements the Gaussian Naive Bayes algorithm for classification.\n * Assumes features follow a Gaussian (normal) distribution.\n *\n * **Algorithm**:\n * 1. Calculate mean and variance for each feature per class\n * 2. For prediction, calculate likelihood using Gaussian PDF\n * 3. Apply Bayes' theorem to get posterior probabilities\n * 4. Predict class with highest posterior probability\n *\n * **Time Complexity**:\n * - Training: O(n * d) where n=samples, d=features\n * - Prediction: O(k * d) per sample where k=classes\n *\n * @example\n * ```ts\n * import { GaussianNB } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1, 2], [2, 3], [3, 4], [4, 5]]);\n * const y = tensor([0, 0, 1, 1]);\n *\n * const nb = new GaussianNB();\n * nb.fit(X, y);\n *\n * const predictions = nb.predict(tensor([[2.5, 3.5]]));\n * ```\n *\n * @see {@link https://en.wikipedia.org/wiki/Naive_Bayes_classifier | Wikipedia: Naive Bayes}\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.GaussianNB.html | scikit-learn GaussianNB}\n */\nexport class GaussianNB implements Classifier {\n  private readonly varSmoothing: number;\n\n  private classes_?: number[];\n  private classPrior_?: Map<number, number>;\n  private theta_?: Map<number, number[]>; // Mean for each class and feature\n  private var_?: Map<number, number[]>; // Variance for each class and feature\n  private nFeaturesIn_?: number;\n  private fitted = false;\n\n  /**\n   * Create a new Gaussian Naive Bayes classifier.\n   *\n   * @param options - Configuration options\n   * @param options.varSmoothing - Portion of largest variance added to variances for stability (default: 1e-9)\n   */\n  constructor(\n    options: {\n      readonly varSmoothing?: number;\n    } = {}\n  ) {\n    this.varSmoothing = options.varSmoothing ?? 1e-9;\n    if (!Number.isFinite(this.varSmoothing) || this.varSmoothing < 0) {\n      throw new InvalidParameterError(\n        \"varSmoothing must be a finite number >= 0\",\n        \"varSmoothing\",\n        this.varSmoothing\n      );\n    }\n  }\n\n  /**\n   * Fit Gaussian Naive Bayes classifier from the training set.\n   *\n   * Computes per-class mean, variance, and prior probabilities.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target class labels of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {DataValidationError} If zero variance encountered with varSmoothing=0\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    this.nFeaturesIn_ = nFeatures;\n\n    // Get unique classes\n    const classSet = new Set<number>();\n    for (let i = 0; i < y.size; i++) {\n      classSet.add(Number(y.data[y.offset + i]));\n    }\n    this.classes_ = Array.from(classSet).sort((a, b) => a - b);\n\n    // Calculate class priors\n    this.classPrior_ = new Map();\n    for (const cls of this.classes_) {\n      let count = 0;\n      for (let i = 0; i < nSamples; i++) {\n        if (Number(y.data[y.offset + i]) === cls) {\n          count++;\n        }\n      }\n      this.classPrior_.set(cls, count / nSamples);\n    }\n\n    // Calculate mean and variance for each class and feature\n    this.theta_ = new Map();\n    this.var_ = new Map();\n\n    for (const cls of this.classes_) {\n      const classSamples: number[][] = [];\n\n      for (let i = 0; i < nSamples; i++) {\n        if (Number(y.data[y.offset + i]) === cls) {\n          const sample: number[] = [];\n          for (let j = 0; j < nFeatures; j++) {\n            sample.push(Number(X.data[X.offset + i * nFeatures + j]));\n          }\n          classSamples.push(sample);\n        }\n      }\n\n      const means: number[] = [];\n      const variances: number[] = [];\n\n      for (let j = 0; j < nFeatures; j++) {\n        // Calculate mean\n        let sum = 0;\n        for (const sample of classSamples) {\n          sum += sample[j] ?? 0;\n        }\n        const mean = sum / classSamples.length;\n        means.push(mean);\n\n        // Calculate variance\n        let varSum = 0;\n        for (const sample of classSamples) {\n          const diff = (sample[j] ?? 0) - mean;\n          varSum += diff * diff;\n        }\n        const variance = varSum / classSamples.length;\n        if (variance === 0 && this.varSmoothing === 0) {\n          throw new DataValidationError(\n            \"Zero variance encountered with varSmoothing=0; set varSmoothing > 0 to avoid degenerate Gaussians\"\n          );\n        }\n        variances.push(variance);\n      }\n\n      this.theta_.set(cls, means);\n      this.var_.set(cls, variances);\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Predict class labels for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted class labels of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"GaussianNB must be fitted before prediction\");\n    }\n\n    const proba = this.predictProba(X);\n    const nSamples = proba.shape[0] ?? 0;\n    const nClasses = proba.shape[1] ?? 0;\n    const predictions: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      let maxProb = -1;\n      let maxClass = 0;\n\n      for (let j = 0; j < nClasses; j++) {\n        const prob = Number(proba.data[proba.offset + i * nClasses + j]);\n        if (prob > maxProb) {\n          maxProb = prob;\n          maxClass = this.classes_?.[j] ?? 0;\n        }\n      }\n\n      predictions.push(maxClass);\n    }\n\n    return tensor(predictions, { dtype: \"int32\" });\n  }\n\n  /**\n   * Predict class probabilities for samples in X.\n   *\n   * Uses Bayes' theorem with Gaussian class-conditional likelihoods.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Class probability matrix of shape (n_samples, n_classes)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predictProba(X: Tensor): Tensor {\n    if (!this.fitted || !this.classes_ || !this.classPrior_ || !this.theta_ || !this.var_) {\n      throw new NotFittedError(\"GaussianNB must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"GaussianNB\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    const probabilities: number[][] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const logProbs: number[] = [];\n\n      for (const cls of this.classes_) {\n        const prior = this.classPrior_.get(cls) ?? 0;\n        const means = this.theta_.get(cls) ?? [];\n        const variances = this.var_.get(cls) ?? [];\n\n        let logProb = Math.log(prior);\n\n        for (let j = 0; j < nFeatures; j++) {\n          const x = Number(X.data[X.offset + i * nFeatures + j]);\n          const mean = means[j] ?? 0;\n          const variance = (variances[j] ?? 0) + this.varSmoothing;\n\n          // Gaussian PDF in log space\n          logProb -= 0.5 * Math.log(2 * Math.PI * variance);\n          logProb -= (x - mean) ** 2 / (2 * variance);\n        }\n\n        logProbs.push(logProb);\n      }\n\n      // Convert log probabilities to probabilities using log-sum-exp trick\n      const maxLogProb = Math.max(...logProbs);\n      const expProbs = logProbs.map((lp) => Math.exp(lp - maxLogProb));\n      const sumExpProbs = expProbs.reduce((a, b) => a + b, 0);\n      const probs = expProbs.map((ep) => ep / sumExpProbs);\n\n      probabilities.push(probs);\n    }\n\n    return tensor(probabilities);\n  }\n\n  /**\n   * Return the mean accuracy on the given test data and labels.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True labels of shape (n_samples,)\n   * @returns Accuracy score in range [0, 1]\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const yPred = this.predict(X);\n    if (yPred.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${yPred.size}, y=${y.size}`\n      );\n    }\n    let correct = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      if (Number(y.data[y.offset + i]) === Number(yPred.data[yPred.offset + i])) {\n        correct++;\n      }\n    }\n\n    return correct / y.size;\n  }\n\n  /**\n   * Get the unique class labels discovered during fitting.\n   *\n   * @returns Tensor of class labels or undefined if not fitted\n   */\n  get classes(): Tensor | undefined {\n    if (!this.fitted || !this.classes_) {\n      return undefined;\n    }\n    return tensor(this.classes_, { dtype: \"int32\" });\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      varSmoothing: this.varSmoothing,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\"GaussianNB does not support setParams after construction\");\n  }\n}\n","import {\n  DataValidationError,\n  InvalidParameterError,\n  NotFittedError,\n  NotImplementedError,\n  ShapeError,\n} from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Classifier, Regressor } from \"../base\";\n\n/**\n * K-Nearest Neighbors base class.\n */\nabstract class KNeighborsBase {\n  protected readonly nNeighbors: number;\n  protected readonly weights: \"uniform\" | \"distance\";\n  protected readonly metric: \"euclidean\" | \"manhattan\";\n\n  protected XTrain_?: Tensor;\n  protected yTrain_?: Tensor;\n  protected nFeaturesIn_?: number;\n  protected fitted = false;\n\n  constructor(\n    options: {\n      readonly nNeighbors?: number;\n      readonly weights?: \"uniform\" | \"distance\";\n      readonly metric?: \"euclidean\" | \"manhattan\";\n    } = {}\n  ) {\n    this.nNeighbors = options.nNeighbors ?? 5;\n    this.weights = options.weights ?? \"uniform\";\n    this.metric = options.metric ?? \"euclidean\";\n\n    if (!Number.isInteger(this.nNeighbors) || this.nNeighbors < 1) {\n      throw new InvalidParameterError(\n        \"nNeighbors must be an integer >= 1\",\n        \"nNeighbors\",\n        this.nNeighbors\n      );\n    }\n    if (this.weights !== \"uniform\" && this.weights !== \"distance\") {\n      throw new InvalidParameterError(\n        `weights must be \"uniform\" or \"distance\"; received ${String(this.weights)}`,\n        \"weights\",\n        this.weights\n      );\n    }\n    if (this.metric !== \"euclidean\" && this.metric !== \"manhattan\") {\n      throw new InvalidParameterError(\n        `metric must be \"euclidean\" or \"manhattan\"; received ${String(this.metric)}`,\n        \"metric\",\n        this.metric\n      );\n    }\n  }\n\n  protected calculateDistance(x1: number[], x2: number[]): number {\n    let dist = 0;\n    if (this.metric === \"euclidean\") {\n      for (let i = 0; i < x1.length; i++) {\n        const diff = (x1[i] ?? 0) - (x2[i] ?? 0);\n        dist += diff * diff;\n      }\n      return Math.sqrt(dist);\n    } else {\n      // manhattan\n      for (let i = 0; i < x1.length; i++) {\n        dist += Math.abs((x1[i] ?? 0) - (x2[i] ?? 0));\n      }\n      return dist;\n    }\n  }\n\n  protected findKNearest(sample: number[]): Array<{ index: number; distance: number }> {\n    if (!this.XTrain_) {\n      throw new NotFittedError(\"Model must be fitted before finding neighbors\");\n    }\n\n    const nSamples = this.XTrain_.shape[0] ?? 0;\n    const nFeatures = this.XTrain_.shape[1] ?? 0;\n\n    const distances: Array<{ index: number; distance: number }> = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const trainSample: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        trainSample.push(Number(this.XTrain_.data[this.XTrain_.offset + i * nFeatures + j]));\n      }\n      const dist = this.calculateDistance(sample, trainSample);\n      distances.push({ index: i, distance: dist });\n    }\n\n    distances.sort((a, b) => a.distance - b.distance);\n    return distances.slice(0, this.nNeighbors);\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nNeighbors: this.nNeighbors,\n      weights: this.weights,\n      metric: this.metric,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\"KNeighbors does not support setParams after construction\");\n  }\n}\n\n/**\n * K-Nearest Neighbors Classifier.\n *\n * Classification based on k nearest neighbors. Predicts class by majority vote\n * of k nearest training samples.\n *\n * **Algorithm**: Instance-based learning\n * 1. Store all training data\n * 2. For each test sample, find k nearest training samples\n * 3. Predict class by majority vote (or weighted vote)\n *\n * **Time Complexity**:\n * - Training: O(1) (just stores data)\n * - Prediction: O(n * d) per sample where n=training samples, d=features\n *\n * @example\n * ```ts\n * import { KNeighborsClassifier } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[0, 0], [1, 1], [2, 2], [3, 3]]);\n * const y = tensor([0, 0, 1, 1]);\n *\n * const knn = new KNeighborsClassifier({ nNeighbors: 3 });\n * knn.fit(X, y);\n *\n * const predictions = knn.predict(tensor([[1.5, 1.5]]));\n * ```\n *\n * @see {@link https://en.wikipedia.org/wiki/K-nearest_neighbors_algorithm | Wikipedia: KNN}\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsClassifier.html | scikit-learn KNeighborsClassifier}\n */\nexport class KNeighborsClassifier extends KNeighborsBase implements Classifier {\n  /**\n   * Fit the k-nearest neighbors classifier from the training set.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target class labels of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {InvalidParameterError} If nNeighbors > n_samples\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    if (this.nNeighbors > nSamples) {\n      throw new InvalidParameterError(\n        `nNeighbors must be <= n_samples; received ${this.nNeighbors} > ${nSamples}`,\n        \"nNeighbors\",\n        this.nNeighbors\n      );\n    }\n\n    this.XTrain_ = X;\n    this.yTrain_ = y;\n    this.nFeaturesIn_ = nFeatures;\n    this.fitted = true;\n\n    return this;\n  }\n\n  /**\n   * Predict class labels for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted class labels of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted || !this.XTrain_ || !this.yTrain_) {\n      throw new NotFittedError(\"KNeighborsClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"KNeighborsClassifier\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const predictions: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const sample: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        sample.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n\n      const neighbors = this.findKNearest(sample);\n\n      // Count votes for each class\n      const votes = new Map<number, number>();\n\n      for (const neighbor of neighbors) {\n        const label = Number(this.yTrain_.data[this.yTrain_.offset + neighbor.index]);\n        const weight = this.weights === \"uniform\" ? 1 : 1 / (neighbor.distance + 1e-10);\n        votes.set(label, (votes.get(label) ?? 0) + weight);\n      }\n\n      // Find class with most votes\n      let maxVotes = -1;\n      let predictedClass = 0;\n      for (const [label, voteCount] of votes.entries()) {\n        if (voteCount > maxVotes) {\n          maxVotes = voteCount;\n          predictedClass = label;\n        }\n      }\n\n      predictions.push(predictedClass);\n    }\n\n    return tensor(predictions, { dtype: \"int32\" });\n  }\n\n  /**\n   * Predict class probabilities for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Class probability matrix of shape (n_samples, n_classes)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predictProba(X: Tensor): Tensor {\n    if (!this.fitted || !this.XTrain_ || !this.yTrain_) {\n      throw new NotFittedError(\"KNeighborsClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"KNeighborsClassifier\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    // Get unique classes\n    const classSet = new Set<number>();\n    for (let i = 0; i < this.yTrain_.size; i++) {\n      classSet.add(Number(this.yTrain_.data[this.yTrain_.offset + i]));\n    }\n    const classes = Array.from(classSet).sort((a, b) => a - b);\n\n    const probabilities: number[][] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const sample: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        sample.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n\n      const neighbors = this.findKNearest(sample);\n\n      // Count votes for each class\n      const votes = new Map<number, number>();\n      let totalWeight = 0;\n\n      for (const neighbor of neighbors) {\n        const label = Number(this.yTrain_.data[this.yTrain_.offset + neighbor.index]);\n        const weight = this.weights === \"uniform\" ? 1 : 1 / (neighbor.distance + 1e-10);\n        votes.set(label, (votes.get(label) ?? 0) + weight);\n        totalWeight += weight;\n      }\n\n      // Convert to probabilities\n      const probs: number[] = [];\n      for (const cls of classes) {\n        probs.push((votes.get(cls) ?? 0) / totalWeight);\n      }\n      probabilities.push(probs);\n    }\n\n    return tensor(probabilities);\n  }\n\n  /**\n   * Return the mean accuracy on the given test data and labels.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True labels of shape (n_samples,)\n   * @returns Accuracy score in range [0, 1]\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const yPred = this.predict(X);\n    if (yPred.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${yPred.size}, y=${y.size}`\n      );\n    }\n    let correct = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      if (Number(y.data[y.offset + i]) === Number(yPred.data[yPred.offset + i])) {\n        correct++;\n      }\n    }\n\n    return correct / y.size;\n  }\n}\n\n/**\n * K-Nearest Neighbors Regressor.\n *\n * Regression based on k nearest neighbors. Predicts value as mean (or weighted mean)\n * of k nearest training samples.\n *\n * @example\n * ```ts\n * import { KNeighborsRegressor } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[0], [1], [2], [3]]);\n * const y = tensor([0, 1, 4, 9]);\n *\n * const knn = new KNeighborsRegressor({ nNeighbors: 2 });\n * knn.fit(X, y);\n *\n * const predictions = knn.predict(tensor([[1.5]]));\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsRegressor.html | scikit-learn KNeighborsRegressor}\n */\nexport class KNeighborsRegressor extends KNeighborsBase implements Regressor {\n  /**\n   * Fit the k-nearest neighbors regressor from the training set.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {InvalidParameterError} If nNeighbors > n_samples\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    if (this.nNeighbors > nSamples) {\n      throw new InvalidParameterError(\n        `nNeighbors must be <= n_samples; received ${this.nNeighbors} > ${nSamples}`,\n        \"nNeighbors\",\n        this.nNeighbors\n      );\n    }\n\n    this.XTrain_ = X;\n    this.yTrain_ = y;\n    this.nFeaturesIn_ = nFeatures;\n    this.fitted = true;\n\n    return this;\n  }\n\n  /**\n   * Predict target values for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted || !this.XTrain_ || !this.yTrain_) {\n      throw new NotFittedError(\"KNeighborsRegressor must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeaturesIn_ ?? 0, \"KNeighborsRegressor\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    const predictions: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const sample: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        sample.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n\n      const neighbors = this.findKNearest(sample);\n\n      // Calculate weighted mean\n      let sumValues = 0;\n      let sumWeights = 0;\n\n      for (const neighbor of neighbors) {\n        const value = Number(this.yTrain_.data[this.yTrain_.offset + neighbor.index]);\n        const weight = this.weights === \"uniform\" ? 1 : 1 / (neighbor.distance + 1e-10);\n        sumValues += value * weight;\n        sumWeights += weight;\n      }\n\n      predictions.push(sumValues / sumWeights);\n    }\n\n    return tensor(predictions);\n  }\n\n  /**\n   * Return the R² score on the given test data and target values.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True target values of shape (n_samples,)\n   * @returns R² score (best possible is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const yPred = this.predict(X);\n    if (yPred.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${yPred.size}, y=${y.size}`\n      );\n    }\n\n    let ssRes = 0;\n    let ssTot = 0;\n    let yMean = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      yMean += Number(y.data[y.offset + i]);\n    }\n    yMean /= y.size;\n\n    for (let i = 0; i < y.size; i++) {\n      const yTrue = Number(y.data[y.offset + i]);\n      const yPredVal = Number(yPred.data[yPred.offset + i]);\n      ssRes += (yTrue - yPredVal) ** 2;\n      ssTot += (yTrue - yMean) ** 2;\n    }\n\n    if (ssTot === 0) {\n      return ssRes === 0 ? 1.0 : 0.0;\n    }\n\n    return 1 - ssRes / ssTot;\n  }\n}\n","import {\n  DataValidationError,\n  InvalidParameterError,\n  NotFittedError,\n  NotImplementedError,\n  ShapeError,\n} from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Classifier, Regressor } from \"../base\";\n\n/**\n * Support Vector Machine (SVM) Classifier.\n *\n * Implements a linear SVM using sub-gradient descent on the hinge loss\n * with L2 regularization (soft margin). Suitable for binary classification tasks.\n *\n * **Algorithm**: Sub-gradient descent on hinge loss (linear kernel)\n *\n * **Mathematical Formulation**:\n * - Decision function: f(x) = sign(w · x + b)\n * - Optimization: minimize (1/2)||w||² + C * Σmax(0, 1 - y_i(w · x_i + b))\n *\n * @example\n * ```ts\n * import { LinearSVC } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1, 2], [2, 3], [3, 1], [4, 2]]);\n * const y = tensor([0, 0, 1, 1]);\n *\n * const svm = new LinearSVC({ C: 1.0 });\n * svm.fit(X, y);\n * const predictions = svm.predict(X);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.svm.LinearSVC.html | scikit-learn LinearSVC}\n */\nexport class LinearSVC implements Classifier {\n  /** Regularization parameter (inverse of regularization strength) */\n  private readonly C: number;\n\n  /** Maximum number of iterations for optimization */\n  private readonly maxIter: number;\n\n  /** Tolerance for stopping criterion */\n  private readonly tol: number;\n\n  /** Weight vector of shape (n_features,) */\n  private weights: number[] = [];\n\n  /** Bias term */\n  private bias = 0;\n\n  /** Number of features seen during fit */\n  private nFeatures = 0;\n\n  /** Unique class labels [0, 1] mapped from original labels */\n  private classLabels: number[] = [];\n\n  /** Whether the model has been fitted */\n  private fitted = false;\n\n  /**\n   * Create a new SVM Classifier.\n   *\n   * @param options - Configuration options\n   * @param options.C - Regularization parameter (default: 1.0). Larger C = stronger penalty on errors = harder margin.\n   * @param options.maxIter - Maximum iterations (default: 1000)\n   * @param options.tol - Convergence tolerance (default: 1e-4)\n   */\n  constructor(\n    options: {\n      readonly C?: number;\n      readonly maxIter?: number;\n      readonly tol?: number;\n    } = {}\n  ) {\n    this.C = options.C ?? 1.0;\n    this.maxIter = options.maxIter ?? 1000;\n    this.tol = options.tol ?? 1e-4;\n\n    // Validate parameters\n    if (!Number.isFinite(this.C) || this.C <= 0) {\n      throw new InvalidParameterError(\"C must be positive\", \"C\", this.C);\n    }\n    if (!Number.isInteger(this.maxIter) || this.maxIter <= 0) {\n      throw new InvalidParameterError(\n        \"maxIter must be a positive integer\",\n        \"maxIter\",\n        this.maxIter\n      );\n    }\n    if (!Number.isFinite(this.tol) || this.tol < 0) {\n      throw new InvalidParameterError(\"tol must be >= 0\", \"tol\", this.tol);\n    }\n  }\n\n  /**\n   * Fit the SVM classifier using sub-gradient descent.\n   *\n   * Uses a simplified hinge loss optimization with L2 regularization.\n   * Objective: minimize (1/2)||w||² + C * Σmax(0, 1 - y_i(w · x_i + b))\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target labels of shape (n_samples,). Must contain exactly 2 classes.\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   * @throws {InvalidParameterError} If y does not contain exactly 2 classes\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    this.nFeatures = nFeatures;\n\n    // Extract data\n    const XData: number[][] = [];\n    const yData: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      XData.push(row);\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    // Get unique classes and map to {-1, 1} for SVM\n    this.classLabels = [...new Set(yData)].sort((a, b) => a - b);\n    if (this.classLabels.length !== 2) {\n      throw new InvalidParameterError(\n        \"LinearSVC requires exactly 2 classes for binary classification\",\n        \"y\",\n        this.classLabels.length\n      );\n    }\n\n    // Map labels to {-1, 1}\n    const yMapped = yData.map((label) => (label === this.classLabels[0] ? -1 : 1));\n\n    // Initialize weights and bias\n    this.weights = new Array(nFeatures).fill(0);\n    this.bias = 0;\n\n    // Sub-gradient descent optimization\n    // We use a constant learning rate schedule for simplicity, scaled by 1/(lambda*n)\n    // Here lambda = 1/C. So eta = C/n.\n    // However, to ensure convergence, usually eta decays as 1/t.\n    // For fixed iterations, small constant rate is often sufficient.\n    // We choose eta such that eta * lambda * n = 1 (approx) implies eta = 1/(lambda*n) = C/n?\n    // Let's use eta = 1 / (nSamples * lambda) = C / nSamples.\n    // But if C is large, eta is large, which might be unstable.\n    // Let's use a safe learning rate.\n    const learningRate = 0.01; // Fixed small learning rate for stability\n\n    for (let iter = 0; iter < this.maxIter; iter++) {\n      let maxViolation = 0;\n\n      for (let i = 0; i < nSamples; i++) {\n        const xi = XData[i];\n        const yi = yMapped[i];\n\n        if (xi === undefined || yi === undefined) continue;\n\n        // Compute decision function: w · x + b\n        let decision = this.bias;\n        for (let j = 0; j < nFeatures; j++) {\n          decision += (this.weights[j] ?? 0) * (xi[j] ?? 0);\n        }\n\n        // Hinge loss margin: y * (w · x + b)\n        const margin = yi * decision;\n\n        // Track constraint violation for convergence check\n        if (margin < 1) {\n          maxViolation = Math.max(maxViolation, 1 - margin);\n        }\n\n        // Sub-gradient update\n        // Objective: 0.5*|w|^2 + C * sum(max(0, 1 - y(wx+b)))\n        // Grad w: w - C*y*x (if margin < 1)\n        // Update: w <- w - eta * (w - C*y*x) = w(1-eta) + eta*C*y*x\n\n        // Regularization part (always applied)\n        // Let's trust the user input C implies hard margin.\n        // We will use: w = w - learningRate * (w - C * y * x)\n        // To prevent explosion, learningRate must be < 1/C.\n        // So we adapt LR.\n        const effectiveLR = Math.min(learningRate, 1.0 / (this.C * 10));\n\n        if (margin < 1) {\n          // Misclassified or within margin\n          for (let j = 0; j < nFeatures; j++) {\n            this.weights[j] =\n              (this.weights[j] ?? 0) * (1 - effectiveLR) + effectiveLR * this.C * yi * (xi[j] ?? 0);\n          }\n          this.bias += effectiveLR * this.C * yi;\n        } else {\n          // Correctly classified outside margin: only apply regularization\n          for (let j = 0; j < nFeatures; j++) {\n            this.weights[j] = (this.weights[j] ?? 0) * (1 - effectiveLR);\n          }\n        }\n      }\n\n      // Check convergence\n      if (maxViolation < this.tol) {\n        break;\n      }\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Predict class labels for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted labels of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"SVC must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"LinearSVC\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    const predictions: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      // Compute decision function\n      let decision = this.bias;\n      for (let j = 0; j < nFeatures; j++) {\n        decision += (this.weights[j] ?? 0) * Number(X.data[X.offset + i * nFeatures + j]);\n      }\n\n      // Map back to original labels\n      const predictedClass = decision >= 0 ? this.classLabels[1] : this.classLabels[0];\n      predictions.push(predictedClass ?? 0);\n    }\n\n    return tensor(predictions, { dtype: \"int32\" });\n  }\n\n  /**\n   * Predict class probabilities using Platt scaling approximation.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Probability estimates of shape (n_samples, 2)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predictProba(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"LinearSVC must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"LinearSVC\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    const proba: number[][] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      // Compute decision function\n      let decision = this.bias;\n      for (let j = 0; j < nFeatures; j++) {\n        decision += (this.weights[j] ?? 0) * Number(X.data[X.offset + i * nFeatures + j]);\n      }\n\n      // Use sigmoid for probability approximation (Platt scaling)\n      const p1 = 1 / (1 + Math.exp(-decision));\n      proba.push([1 - p1, p1]);\n    }\n\n    return tensor(proba);\n  }\n\n  /**\n   * Return the mean accuracy on the given test data and labels.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True labels of shape (n_samples,)\n   * @returns Accuracy score in range [0, 1]\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n    let correct = 0;\n    for (let i = 0; i < y.size; i++) {\n      if (Number(predictions.data[predictions.offset + i]) === Number(y.data[y.offset + i])) {\n        correct++;\n      }\n    }\n    return correct / y.size;\n  }\n\n  /**\n   * Get the weight vector.\n   *\n   * @returns Weight vector as tensor of shape (1, n_features)\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get coef(): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"LinearSVC must be fitted to access coefficients\");\n    }\n    return tensor([this.weights]);\n  }\n\n  /**\n   * Get the bias term.\n   *\n   * @returns Bias value\n   * @throws {NotFittedError} If the model has not been fitted\n   */\n  get intercept(): number {\n    if (!this.fitted) {\n      throw new NotFittedError(\"LinearSVC must be fitted to access intercept\");\n    }\n    return this.bias;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      C: this.C,\n      maxIter: this.maxIter,\n      tol: this.tol,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\"LinearSVC does not support setParams after construction\");\n  }\n}\n\n/**\n * Support Vector Machine (SVM) Regressor.\n *\n * Implements epsilon-SVR (Support Vector Regression) using sub-gradient descent.\n *\n * @example\n * ```ts\n * import { LinearSVR } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const X = tensor([[1], [2], [3], [4]]);\n * const y = tensor([1.5, 2.5, 3.5, 4.5]);\n *\n * const svr = new LinearSVR({ C: 1.0, epsilon: 0.1 });\n * svr.fit(X, y);\n * const predictions = svr.predict(X);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.svm.LinearSVR.html | scikit-learn LinearSVR}\n */\nexport class LinearSVR implements Regressor {\n  /** Regularization parameter */\n  private readonly C: number;\n\n  /** Epsilon in the epsilon-SVR model */\n  private readonly epsilon: number;\n\n  /** Maximum number of iterations */\n  private readonly maxIter: number;\n\n  /** Tolerance for stopping criterion */\n  private readonly tol: number;\n\n  /** Weight vector */\n  private weights: number[] = [];\n\n  /** Bias term */\n  private bias = 0;\n\n  /** Number of features */\n  private nFeatures = 0;\n\n  /** Whether the model has been fitted */\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly C?: number;\n      readonly epsilon?: number;\n      readonly maxIter?: number;\n      readonly tol?: number;\n    } = {}\n  ) {\n    this.C = options.C ?? 1.0;\n    this.epsilon = options.epsilon ?? 0.1;\n    this.maxIter = options.maxIter ?? 1000;\n    this.tol = options.tol ?? 1e-4;\n\n    if (!Number.isFinite(this.C) || this.C <= 0) {\n      throw new InvalidParameterError(\"C must be positive\", \"C\", this.C);\n    }\n    if (!Number.isFinite(this.epsilon) || this.epsilon < 0) {\n      throw new InvalidParameterError(\"epsilon must be >= 0\", \"epsilon\", this.epsilon);\n    }\n    if (!Number.isInteger(this.maxIter) || this.maxIter <= 0) {\n      throw new InvalidParameterError(\"maxIter must be positive\", \"maxIter\", this.maxIter);\n    }\n    if (!Number.isFinite(this.tol) || this.tol < 0) {\n      throw new InvalidParameterError(\"tol must be >= 0\", \"tol\", this.tol);\n    }\n  }\n\n  /**\n   * Fit the SVR model using sub-gradient descent on epsilon-insensitive loss.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    this.nFeatures = nFeatures;\n\n    // Extract data\n    const XData: number[][] = [];\n    const yData: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      XData.push(row);\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    // Initialize weights\n    this.weights = new Array(nFeatures).fill(0);\n    this.bias = 0;\n\n    const learningRate = 0.01;\n\n    for (let iter = 0; iter < this.maxIter; iter++) {\n      let totalLoss = 0;\n\n      for (let i = 0; i < nSamples; i++) {\n        const xi = XData[i];\n        const yi = yData[i];\n\n        if (xi === undefined || yi === undefined) continue;\n\n        // Compute prediction\n        let pred = this.bias;\n        for (let j = 0; j < nFeatures; j++) {\n          pred += (this.weights[j] ?? 0) * (xi[j] ?? 0);\n        }\n\n        const error = pred - yi;\n        const absError = Math.abs(error);\n\n        // Epsilon-insensitive loss\n        if (absError > this.epsilon) {\n          totalLoss += absError - this.epsilon;\n\n          // Sub-gradient\n          const sign = error > 0 ? 1 : -1;\n\n          for (let j = 0; j < nFeatures; j++) {\n            this.weights[j] =\n              (this.weights[j] ?? 0) -\n              learningRate * (this.C * sign * (xi[j] ?? 0) + (this.weights[j] ?? 0));\n          }\n          this.bias -= learningRate * this.C * sign;\n        } else {\n          // Only regularization\n          for (let j = 0; j < nFeatures; j++) {\n            this.weights[j] = (this.weights[j] ?? 0) - learningRate * (this.weights[j] ?? 0);\n          }\n        }\n      }\n\n      if (totalLoss / nSamples < this.tol) {\n        break;\n      }\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Predict target values for samples in X.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"SVR must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"SVR\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n\n    const predictions: number[] = [];\n\n    for (let i = 0; i < nSamples; i++) {\n      let pred = this.bias;\n      for (let j = 0; j < nFeatures; j++) {\n        pred += (this.weights[j] ?? 0) * Number(X.data[X.offset + i * nFeatures + j]);\n      }\n      predictions.push(pred);\n    }\n\n    return tensor(predictions);\n  }\n\n  /**\n   * Return the R² score on the given test data and target values.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True target values of shape (n_samples,)\n   * @returns R² score (best possible is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n\n    let ssRes = 0;\n    let ssTot = 0;\n    let yMean = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      yMean += Number(y.data[y.offset + i]);\n    }\n    yMean /= y.size;\n\n    for (let i = 0; i < y.size; i++) {\n      const yTrue = Number(y.data[y.offset + i]);\n      const yPred = Number(predictions.data[predictions.offset + i]);\n      ssRes += (yTrue - yPred) ** 2;\n      ssTot += (yTrue - yMean) ** 2;\n    }\n\n    return ssTot === 0 ? (ssRes === 0 ? 1.0 : 0.0) : 1 - ssRes / ssTot;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      C: this.C,\n      epsilon: this.epsilon,\n      maxIter: this.maxIter,\n      tol: this.tol,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\"LinearSVR does not support setParams after construction\");\n  }\n}\n","import {\n  DataValidationError,\n  InvalidParameterError,\n  NotFittedError,\n  NotImplementedError,\n  ShapeError,\n} from \"../../core\";\nimport { type Tensor, tensor } from \"../../ndarray\";\nimport { assertContiguous, validateFitInputs, validatePredictInputs } from \"../_validation\";\nimport type { Classifier, Regressor } from \"../base\";\nimport { DecisionTreeClassifier, DecisionTreeRegressor } from \"./DecisionTree\";\n\n/**\n * Random Forest Classifier.\n *\n * An ensemble of decision trees trained on random subsets of data and features.\n * Predictions are made by majority voting.\n *\n * **Algorithm**:\n * 1. Create n_estimators bootstrap samples from training data\n * 2. Train a decision tree on each sample with random feature subsets\n * 3. Aggregate predictions via majority voting\n *\n * @example\n * ```ts\n * import { RandomForestClassifier } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const clf = new RandomForestClassifier({ nEstimators: 100 });\n * clf.fit(X_train, y_train);\n * const predictions = clf.predict(X_test);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html | scikit-learn RandomForestClassifier}\n */\nexport class RandomForestClassifier implements Classifier {\n  private readonly nEstimators: number;\n  private readonly maxDepth: number;\n  private readonly minSamplesSplit: number;\n  private readonly minSamplesLeaf: number;\n  private readonly maxFeatures: \"sqrt\" | \"log2\" | number;\n  private readonly bootstrap: boolean;\n  private readonly randomState?: number;\n\n  private trees: DecisionTreeClassifier[] = [];\n  private classLabels?: number[];\n  private nFeatures?: number;\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly nEstimators?: number;\n      readonly maxDepth?: number;\n      readonly minSamplesSplit?: number;\n      readonly minSamplesLeaf?: number;\n      readonly maxFeatures?: \"sqrt\" | \"log2\" | number;\n      readonly bootstrap?: boolean;\n      readonly randomState?: number;\n    } = {}\n  ) {\n    this.nEstimators = options.nEstimators ?? 100;\n    this.maxDepth = options.maxDepth ?? 10;\n    this.minSamplesSplit = options.minSamplesSplit ?? 2;\n    this.minSamplesLeaf = options.minSamplesLeaf ?? 1;\n    this.maxFeatures = options.maxFeatures ?? \"sqrt\";\n    this.bootstrap = options.bootstrap ?? true;\n    if (options.randomState !== undefined) {\n      this.randomState = options.randomState;\n    }\n\n    if (!Number.isInteger(this.nEstimators) || this.nEstimators < 1) {\n      throw new InvalidParameterError(\n        `nEstimators must be an integer >= 1; received ${this.nEstimators}`,\n        \"nEstimators\",\n        this.nEstimators\n      );\n    }\n    if (!Number.isInteger(this.maxDepth) || this.maxDepth < 1) {\n      throw new InvalidParameterError(\n        `maxDepth must be an integer >= 1; received ${this.maxDepth}`,\n        \"maxDepth\",\n        this.maxDepth\n      );\n    }\n    if (!Number.isInteger(this.minSamplesSplit) || this.minSamplesSplit < 2) {\n      throw new InvalidParameterError(\n        `minSamplesSplit must be an integer >= 2; received ${this.minSamplesSplit}`,\n        \"minSamplesSplit\",\n        this.minSamplesSplit\n      );\n    }\n    if (!Number.isInteger(this.minSamplesLeaf) || this.minSamplesLeaf < 1) {\n      throw new InvalidParameterError(\n        `minSamplesLeaf must be an integer >= 1; received ${this.minSamplesLeaf}`,\n        \"minSamplesLeaf\",\n        this.minSamplesLeaf\n      );\n    }\n    if (typeof this.maxFeatures === \"number\") {\n      if (!Number.isInteger(this.maxFeatures) || this.maxFeatures < 1) {\n        throw new InvalidParameterError(\n          `maxFeatures must be an integer >= 1; received ${this.maxFeatures}`,\n          \"maxFeatures\",\n          this.maxFeatures\n        );\n      }\n    } else if (this.maxFeatures !== \"sqrt\" && this.maxFeatures !== \"log2\") {\n      throw new InvalidParameterError(\n        `maxFeatures must be \"sqrt\", \"log2\", or a positive integer; received ${String(this.maxFeatures)}`,\n        \"maxFeatures\",\n        this.maxFeatures\n      );\n    }\n    if (options.randomState !== undefined && !Number.isFinite(options.randomState)) {\n      throw new InvalidParameterError(\n        `randomState must be a finite number; received ${String(options.randomState)}`,\n        \"randomState\",\n        options.randomState\n      );\n    }\n  }\n\n  private createRNG(): () => number {\n    if (this.randomState !== undefined) {\n      let seed = this.randomState;\n      return () => {\n        seed = (seed * 9301 + 49297) % 233280;\n        return seed / 233280;\n      };\n    }\n    return Math.random;\n  }\n\n  /**\n   * Fit the random forest classifier on training data.\n   *\n   * Builds an ensemble of decision trees, each trained on a bootstrapped\n   * sample with random feature subsets.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target class labels of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    this.nFeatures = nFeatures;\n\n    // Extract data\n    const XData: number[][] = [];\n    const yData: number[] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      XData.push(row);\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    this.classLabels = [...new Set(yData)].sort((a, b) => a - b);\n\n    // Determine number of features to select\n    let nSelectFeatures: number;\n    if (typeof this.maxFeatures === \"number\") {\n      nSelectFeatures = Math.min(this.maxFeatures, nFeatures);\n    } else if (this.maxFeatures === \"sqrt\") {\n      nSelectFeatures = Math.floor(Math.sqrt(nFeatures));\n    } else {\n      nSelectFeatures = Math.floor(Math.log2(nFeatures));\n    }\n    nSelectFeatures = Math.max(1, nSelectFeatures);\n\n    const rng = this.createRNG();\n\n    this.trees = [];\n\n    for (let t = 0; t < this.nEstimators; t++) {\n      // Bootstrap sample\n      const sampleIndices: number[] = [];\n      if (this.bootstrap) {\n        for (let i = 0; i < nSamples; i++) {\n          sampleIndices.push(Math.floor(rng() * nSamples));\n        }\n      } else {\n        for (let i = 0; i < nSamples; i++) {\n          sampleIndices.push(i);\n        }\n      }\n\n      // Create subset of data (all features, bootstrapped samples)\n      const XSubset: number[][] = [];\n      const ySubset: number[] = [];\n      for (const sampleIdx of sampleIndices) {\n        // We copy the whole row (all features)\n        XSubset.push(XData[sampleIdx] ?? []);\n        ySubset.push(yData[sampleIdx] ?? 0);\n      }\n\n      // Train tree with maxFeatures set for random feature selection at each split\n      const treeOptions: {\n        maxDepth: number;\n        minSamplesSplit: number;\n        minSamplesLeaf: number;\n        maxFeatures: number;\n        randomState?: number;\n      } = {\n        maxDepth: this.maxDepth,\n        minSamplesSplit: this.minSamplesSplit,\n        minSamplesLeaf: this.minSamplesLeaf,\n        maxFeatures: nSelectFeatures,\n      };\n      if (this.randomState !== undefined) {\n        treeOptions.randomState = this.randomState + t;\n      }\n      const tree = new DecisionTreeClassifier(treeOptions);\n      tree.fit(tensor(XSubset), tensor(ySubset, { dtype: \"int32\" }));\n      this.trees.push(tree);\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Predict class labels for samples in X.\n   *\n   * Aggregates predictions from all trees via majority voting.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted class labels of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"RandomForestClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"RandomForestClassifier\");\n\n    const nSamples = X.shape[0] ?? 0;\n\n    // Get predictions from all trees\n    const allPredictions: number[][] = [];\n\n    for (const tree of this.trees) {\n      const preds = tree.predict(X);\n      const treePreds: number[] = [];\n      for (let i = 0; i < nSamples; i++) {\n        treePreds.push(Number(preds.data[preds.offset + i]));\n      }\n      allPredictions.push(treePreds);\n    }\n\n    // Majority voting\n    const finalPredictions: number[] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const votes = new Map<number, number>();\n      for (const treePreds of allPredictions) {\n        const pred = treePreds[i] ?? 0;\n        votes.set(pred, (votes.get(pred) ?? 0) + 1);\n      }\n\n      let maxVotes = 0;\n      let prediction = 0;\n      for (const [label, count] of votes) {\n        if (count > maxVotes) {\n          maxVotes = count;\n          prediction = label;\n        }\n      }\n      finalPredictions.push(prediction);\n    }\n\n    return tensor(finalPredictions, { dtype: \"int32\" });\n  }\n\n  /**\n   * Predict class probabilities for samples in X.\n   *\n   * Averages the predicted class probabilities from all trees in the ensemble.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Class probability matrix of shape (n_samples, n_classes)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predictProba(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"RandomForestClassifier must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"RandomForestClassifier\");\n    assertContiguous(X, \"X\");\n\n    const nSamples = X.shape[0] ?? 0;\n    const classLabels = this.classLabels ?? [];\n    const nClasses = classLabels.length;\n\n    if (nClasses === 0) {\n      throw new NotFittedError(\"RandomForestClassifier must be fitted before prediction\");\n    }\n\n    const classIndex = new Map<number, number>();\n    for (let i = 0; i < nClasses; i++) {\n      const v = classLabels[i];\n      if (v !== undefined) classIndex.set(v, i);\n    }\n\n    const proba: number[][] = Array.from({ length: nSamples }, () =>\n      new Array<number>(nClasses).fill(0)\n    );\n\n    for (const tree of this.trees) {\n      const treeProba = tree.predictProba(X);\n      const treeClasses = tree.classes;\n      if (!treeClasses) continue;\n      assertContiguous(treeClasses, \"classes\");\n\n      const k = treeClasses.size;\n      for (let j = 0; j < k; j++) {\n        const lbl = Number(treeClasses.data[treeClasses.offset + j] ?? 0);\n        const globalJ = classIndex.get(lbl);\n        if (globalJ === undefined) continue;\n\n        for (let i = 0; i < nSamples; i++) {\n          const row = proba[i];\n          if (row) {\n            row[globalJ] =\n              (row[globalJ] ?? 0) + Number(treeProba.data[treeProba.offset + i * k + j] ?? 0);\n          }\n        }\n      }\n    }\n\n    const invTrees = this.trees.length === 0 ? 0 : 1 / this.trees.length;\n    for (let i = 0; i < nSamples; i++) {\n      const row = proba[i];\n      if (row) {\n        for (let j = 0; j < nClasses; j++) {\n          row[j] = (row[j] ?? 0) * invTrees;\n        }\n      }\n    }\n\n    return tensor(proba);\n  }\n\n  /**\n   * Return the mean accuracy on the given test data and labels.\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True labels of shape (n_samples,)\n   * @returns Accuracy score in range [0, 1]\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (!this.fitted) {\n      throw new NotFittedError(\"RandomForestClassifier must be fitted before scoring\");\n    }\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n    let correct = 0;\n    for (let i = 0; i < y.size; i++) {\n      if (Number(predictions.data[predictions.offset + i]) === Number(y.data[y.offset + i])) {\n        correct++;\n      }\n    }\n    return correct / y.size;\n  }\n\n  /**\n   * Get the unique class labels discovered during fitting.\n   *\n   * @returns Tensor of class labels or undefined if not fitted\n   */\n  get classes(): Tensor | undefined {\n    if (!this.fitted || !this.classLabels) {\n      return undefined;\n    }\n    return tensor(this.classLabels, { dtype: \"int32\" });\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nEstimators: this.nEstimators,\n      maxDepth: this.maxDepth,\n      minSamplesSplit: this.minSamplesSplit,\n      minSamplesLeaf: this.minSamplesLeaf,\n      maxFeatures: this.maxFeatures,\n      bootstrap: this.bootstrap,\n      randomState: this.randomState,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\n      \"RandomForestClassifier does not support setParams after construction\"\n    );\n  }\n}\n\n/**\n * Random Forest Regressor.\n *\n * An ensemble of decision tree regressors trained on random subsets of data\n * and features. Predictions are averaged across all trees.\n *\n * **Algorithm**:\n * 1. Create n_estimators bootstrap samples from training data\n * 2. Train a decision tree on each sample with random feature subsets\n * 3. Aggregate predictions via averaging\n *\n * @example\n * ```ts\n * import { RandomForestRegressor } from 'deepbox/ml';\n * import { tensor } from 'deepbox/ndarray';\n *\n * const reg = new RandomForestRegressor({ nEstimators: 100 });\n * reg.fit(X_train, y_train);\n * const predictions = reg.predict(X_test);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestRegressor.html | scikit-learn RandomForestRegressor}\n */\nexport class RandomForestRegressor implements Regressor {\n  private readonly nEstimators: number;\n  private readonly maxDepth: number;\n  private readonly minSamplesSplit: number;\n  private readonly minSamplesLeaf: number;\n  private readonly maxFeatures: \"sqrt\" | \"log2\" | number;\n  private readonly bootstrap: boolean;\n  private readonly randomState?: number;\n\n  private trees: DecisionTreeRegressor[] = [];\n  private nFeatures?: number;\n  private fitted = false;\n\n  constructor(\n    options: {\n      readonly nEstimators?: number;\n      readonly maxDepth?: number;\n      readonly minSamplesSplit?: number;\n      readonly minSamplesLeaf?: number;\n      readonly maxFeatures?: \"sqrt\" | \"log2\" | number;\n      readonly bootstrap?: boolean;\n      readonly randomState?: number;\n    } = {}\n  ) {\n    this.nEstimators = options.nEstimators ?? 100;\n    this.maxDepth = options.maxDepth ?? 10;\n    this.minSamplesSplit = options.minSamplesSplit ?? 2;\n    this.minSamplesLeaf = options.minSamplesLeaf ?? 1;\n    this.maxFeatures = options.maxFeatures ?? 1.0;\n    this.bootstrap = options.bootstrap ?? true;\n    if (options.randomState !== undefined) {\n      this.randomState = options.randomState;\n    }\n\n    if (!Number.isInteger(this.nEstimators) || this.nEstimators < 1) {\n      throw new InvalidParameterError(\n        `nEstimators must be an integer >= 1; received ${this.nEstimators}`,\n        \"nEstimators\",\n        this.nEstimators\n      );\n    }\n    if (!Number.isInteger(this.maxDepth) || this.maxDepth < 1) {\n      throw new InvalidParameterError(\n        `maxDepth must be an integer >= 1; received ${this.maxDepth}`,\n        \"maxDepth\",\n        this.maxDepth\n      );\n    }\n    if (!Number.isInteger(this.minSamplesSplit) || this.minSamplesSplit < 2) {\n      throw new InvalidParameterError(\n        `minSamplesSplit must be an integer >= 2; received ${this.minSamplesSplit}`,\n        \"minSamplesSplit\",\n        this.minSamplesSplit\n      );\n    }\n    if (!Number.isInteger(this.minSamplesLeaf) || this.minSamplesLeaf < 1) {\n      throw new InvalidParameterError(\n        `minSamplesLeaf must be an integer >= 1; received ${this.minSamplesLeaf}`,\n        \"minSamplesLeaf\",\n        this.minSamplesLeaf\n      );\n    }\n    if (typeof this.maxFeatures === \"number\") {\n      if (\n        this.maxFeatures !== 1.0 &&\n        (!Number.isInteger(this.maxFeatures) || this.maxFeatures < 1)\n      ) {\n        throw new InvalidParameterError(\n          `maxFeatures must be 1.0, an integer >= 1, \"sqrt\", or \"log2\"; received ${this.maxFeatures}`,\n          \"maxFeatures\",\n          this.maxFeatures\n        );\n      }\n    } else if (this.maxFeatures !== \"sqrt\" && this.maxFeatures !== \"log2\") {\n      throw new InvalidParameterError(\n        `maxFeatures must be \"sqrt\", \"log2\", or a positive integer; received ${String(this.maxFeatures)}`,\n        \"maxFeatures\",\n        this.maxFeatures\n      );\n    }\n    if (options.randomState !== undefined && !Number.isFinite(options.randomState)) {\n      throw new InvalidParameterError(\n        `randomState must be a finite number; received ${String(options.randomState)}`,\n        \"randomState\",\n        options.randomState\n      );\n    }\n  }\n\n  private createRNG(): () => number {\n    if (this.randomState !== undefined) {\n      let seed = this.randomState;\n      return () => {\n        seed = (seed * 9301 + 49297) % 233280;\n        return seed / 233280;\n      };\n    }\n    return Math.random;\n  }\n\n  /**\n   * Fit the random forest regressor on training data.\n   *\n   * Builds an ensemble of decision trees, each trained on a bootstrapped\n   * sample with random feature subsets.\n   *\n   * @param X - Training data of shape (n_samples, n_features)\n   * @param y - Target values of shape (n_samples,)\n   * @returns this - The fitted estimator\n   * @throws {ShapeError} If X is not 2D or y is not 1D\n   * @throws {ShapeError} If X and y have different number of samples\n   * @throws {DataValidationError} If X or y contain NaN/Inf values\n   */\n  fit(X: Tensor, y: Tensor): this {\n    validateFitInputs(X, y);\n\n    const nSamples = X.shape[0] ?? 0;\n    const nFeatures = X.shape[1] ?? 0;\n    this.nFeatures = nFeatures;\n\n    const XData: number[][] = [];\n    const yData: number[] = [];\n    for (let i = 0; i < nSamples; i++) {\n      const row: number[] = [];\n      for (let j = 0; j < nFeatures; j++) {\n        row.push(Number(X.data[X.offset + i * nFeatures + j]));\n      }\n      XData.push(row);\n      yData.push(Number(y.data[y.offset + i]));\n    }\n\n    let nSelectFeatures: number;\n    if (typeof this.maxFeatures === \"number\") {\n      if (this.maxFeatures === 1.0) {\n        nSelectFeatures = nFeatures;\n      } else {\n        nSelectFeatures = Math.min(this.maxFeatures, nFeatures);\n      }\n    } else if (this.maxFeatures === \"sqrt\") {\n      nSelectFeatures = Math.floor(Math.sqrt(nFeatures));\n    } else {\n      nSelectFeatures = Math.floor(Math.log2(nFeatures));\n    }\n    nSelectFeatures = Math.max(1, nSelectFeatures);\n\n    const rng = this.createRNG();\n\n    this.trees = [];\n\n    for (let t = 0; t < this.nEstimators; t++) {\n      const sampleIndices: number[] = [];\n      if (this.bootstrap) {\n        for (let i = 0; i < nSamples; i++) {\n          sampleIndices.push(Math.floor(rng() * nSamples));\n        }\n      } else {\n        for (let i = 0; i < nSamples; i++) {\n          sampleIndices.push(i);\n        }\n      }\n\n      // Create subset of data (all features, bootstrapped samples)\n      const XSubset: number[][] = [];\n      const ySubset: number[] = [];\n      for (const sampleIdx of sampleIndices) {\n        XSubset.push(XData[sampleIdx] ?? []);\n        ySubset.push(yData[sampleIdx] ?? 0);\n      }\n\n      const treeOptions: {\n        maxDepth: number;\n        minSamplesSplit: number;\n        minSamplesLeaf: number;\n        maxFeatures: number;\n        randomState?: number;\n      } = {\n        maxDepth: this.maxDepth,\n        minSamplesSplit: this.minSamplesSplit,\n        minSamplesLeaf: this.minSamplesLeaf,\n        maxFeatures: nSelectFeatures,\n      };\n      if (this.randomState !== undefined) {\n        treeOptions.randomState = this.randomState + t;\n      }\n      const tree = new DecisionTreeRegressor(treeOptions);\n      tree.fit(tensor(XSubset), tensor(ySubset));\n      this.trees.push(tree);\n    }\n\n    this.fitted = true;\n    return this;\n  }\n\n  /**\n   * Predict target values for samples in X.\n   *\n   * Averages predictions from all trees in the ensemble.\n   *\n   * @param X - Samples of shape (n_samples, n_features)\n   * @returns Predicted values of shape (n_samples,)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If X has wrong dimensions or feature count\n   * @throws {DataValidationError} If X contains NaN/Inf values\n   */\n  predict(X: Tensor): Tensor {\n    if (!this.fitted) {\n      throw new NotFittedError(\"RandomForestRegressor must be fitted before prediction\");\n    }\n\n    validatePredictInputs(X, this.nFeatures ?? 0, \"RandomForestRegressor\");\n\n    const nSamples = X.shape[0] ?? 0;\n\n    const allPredictions: number[][] = [];\n\n    for (const tree of this.trees) {\n      const preds = tree.predict(X);\n      const treePreds: number[] = [];\n      for (let i = 0; i < nSamples; i++) {\n        treePreds.push(Number(preds.data[preds.offset + i]));\n      }\n      allPredictions.push(treePreds);\n    }\n\n    // Average predictions\n    const finalPredictions: number[] = [];\n    for (let i = 0; i < nSamples; i++) {\n      let sum = 0;\n      for (const treePreds of allPredictions) {\n        sum += treePreds[i] ?? 0;\n      }\n      finalPredictions.push(sum / this.trees.length);\n    }\n\n    return tensor(finalPredictions);\n  }\n\n  /**\n   * Return the R² score on the given test data and target values.\n   *\n   * R² = 1 - SS_res / SS_tot, where SS_res = Σ(y - ŷ)² and SS_tot = Σ(y - ȳ)².\n   *\n   * @param X - Test samples of shape (n_samples, n_features)\n   * @param y - True target values of shape (n_samples,)\n   * @returns R² score (best possible is 1.0, can be negative)\n   * @throws {NotFittedError} If the model has not been fitted\n   * @throws {ShapeError} If y is not 1-dimensional or sample counts mismatch\n   * @throws {DataValidationError} If y contains NaN/Inf values\n   */\n  score(X: Tensor, y: Tensor): number {\n    if (!this.fitted) {\n      throw new NotFittedError(\"RandomForestRegressor must be fitted before scoring\");\n    }\n    if (y.ndim !== 1) {\n      throw new ShapeError(`y must be 1-dimensional; got ndim=${y.ndim}`);\n    }\n    assertContiguous(y, \"y\");\n    for (let i = 0; i < y.size; i++) {\n      const val = y.data[y.offset + i] ?? 0;\n      if (!Number.isFinite(val)) {\n        throw new DataValidationError(\"y contains non-finite values (NaN or Inf)\");\n      }\n    }\n    const predictions = this.predict(X);\n    if (predictions.size !== y.size) {\n      throw new ShapeError(\n        `X and y must have the same number of samples; got X=${predictions.size}, y=${y.size}`\n      );\n    }\n\n    let ssRes = 0;\n    let ssTot = 0;\n    let yMean = 0;\n\n    for (let i = 0; i < y.size; i++) {\n      yMean += Number(y.data[y.offset + i]);\n    }\n    yMean /= y.size;\n\n    for (let i = 0; i < y.size; i++) {\n      const yTrue = Number(y.data[y.offset + i]);\n      const yPred = Number(predictions.data[predictions.offset + i]);\n      ssRes += (yTrue - yPred) ** 2;\n      ssTot += (yTrue - yMean) ** 2;\n    }\n\n    return ssTot === 0 ? (ssRes === 0 ? 1.0 : 0.0) : 1 - ssRes / ssTot;\n  }\n\n  /**\n   * Get hyperparameters for this estimator.\n   *\n   * @returns Object containing all hyperparameters\n   */\n  getParams(): Record<string, unknown> {\n    return {\n      nEstimators: this.nEstimators,\n      maxDepth: this.maxDepth,\n      minSamplesSplit: this.minSamplesSplit,\n      minSamplesLeaf: this.minSamplesLeaf,\n      maxFeatures: this.maxFeatures,\n      bootstrap: this.bootstrap,\n      randomState: this.randomState,\n    };\n  }\n\n  /**\n   * Set the parameters of this estimator.\n   *\n   * @param _params - Parameters to set\n   * @throws {NotImplementedError} Always — parameters cannot be changed after construction\n   */\n  setParams(_params: Record<string, unknown>): this {\n    throw new NotImplementedError(\n      \"RandomForestRegressor does not support setParams after construction\"\n    );\n  }\n}\n"]}