deepbox 0.1.0

package/dist/chunk-FJYLIGJX.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/metrics/index.ts","../src/metrics/_internal.ts","../src/metrics/classification.ts","../src/metrics/clustering.ts","../src/metrics/regression.ts"],"names":["recall","precision","contingency"],"mappings":";;;;AAAA,IAAA,eAAA,GAAA;AAAA,QAAA,CAAA,eAAA,EAAA;AAAA,EAAA,QAAA,EAAA,MAAA,QAAA;AAAA,EAAA,uBAAA,EAAA,MAAA,uBAAA;AAAA,EAAA,eAAA,EAAA,MAAA,eAAA;AAAA,EAAA,iBAAA,EAAA,MAAA,iBAAA;AAAA,EAAA,qBAAA,EAAA,MAAA,qBAAA;AAAA,EAAA,qBAAA,EAAA,MAAA,qBAAA;AAAA,EAAA,qBAAA,EAAA,MAAA,qBAAA;AAAA,EAAA,oBAAA,EAAA,MAAA,oBAAA;AAAA,EAAA,eAAA,EAAA,MAAA,eAAA;AAAA,EAAA,iBAAA,EAAA,MAAA,iBAAA;AAAA,EAAA,eAAA,EAAA,MAAA,eAAA;AAAA,EAAA,kBAAA,EAAA,MAAA,kBAAA;AAAA,EAAA,sBAAA,EAAA,MAAA,sBAAA;AAAA,EAAA,OAAA,EAAA,MAAA,OAAA;AAAA,EAAA,UAAA,EAAA,MAAA,UAAA;AAAA,EAAA,mBAAA,EAAA,MAAA,mBAAA;AAAA,EAAA,WAAA,EAAA,MAAA,WAAA;AAAA,EAAA,gBAAA,EAAA,MAAA,gBAAA;AAAA,EAAA,YAAA,EAAA,MAAA,YAAA;AAAA,EAAA,OAAA,EAAA,MAAA,OAAA;AAAA,EAAA,GAAA,EAAA,MAAA,GAAA;AAAA,EAAA,IAAA,EAAA,MAAA,IAAA;AAAA,EAAA,gBAAA,EAAA,MAAA,gBAAA;AAAA,EAAA,QAAA,EAAA,MAAA,QAAA;AAAA,EAAA,mBAAA,EAAA,MAAA,mBAAA;AAAA,EAAA,GAAA,EAAA,MAAA,GAAA;AAAA,EAAA,yBAAA,EAAA,MAAA,yBAAA;AAAA,EAAA,SAAA,EAAA,MAAA,SAAA;AAAA,EAAA,oBAAA,EAAA,MAAA,oBAAA;AAAA,EAAA,OAAA,EAAA,MAAA,OAAA;AAAA,EAAA,MAAA,EAAA,MAAA,MAAA;AAAA,EAAA,IAAA,EAAA,MAAA,IAAA;AAAA,EAAA,WAAA,EAAA,MAAA,WAAA;AAAA,EAAA,QAAA,EAAA,MAAA,QAAA;AAAA,EAAA,iBAAA,EAAA,MAAA,iBAAA;AAAA,EAAA,eAAA,EAAA,MAAA,eAAA;AAAA,EAAA,aAAA,EAAA,MAAA;AAAA,CAAA,CAAA;;;ACwBO,SAAS,sBAAsB,KAAA,EAAoC;AACxE,EAAA,MAAM,OAAA,GAAU,IAAI,KAAA,CAAc,KAAA,CAAM,MAAM,CAAA;AAC9C,EAAA,IAAI,MAAA,GAAS,CAAA;AACb,EAAA,KAAA,IAAS,IAAI,KAAA,CAAM,MAAA,GAAS,CAAA,EAAG,CAAA,IAAK,GAAG,CAAA,EAAA,EAAK;AAC1C,IAAA,MAAM,GAAA,GAAM,MAAM,CAAC,CAAA;AACnB,IAAA,IAAI,QAAQ,MAAA,EAAW;AACrB,MAAA,MAAM,IAAI,WAAW,oCAAoC,CAAA;AAAA,IAC3D;AACA,IAAA,OAAA,CAAQ,CAAC,CAAA,GAAI,MAAA;AACb,IAAA,MAAA,IAAU,GAAA;AAAA,EACZ;AACA,EAAA,OAAO,OAAA;AACT;AASO,SAAS,oBAAoB,CAAA,EAA0B;AAC5D,EAAA,MAAM,OAAO,CAAA,CAAE,MAAA;AAEf,EAAA,IAAI,CAAA,CAAE,QAAQ,CAAA,EAAG;AACf,IAAA,MAAM,OAAA,GAAU,CAAA,CAAE,OAAA,CAAQ,CAAC,CAAA,IAAK,CAAA;AAChC,IAAA,OAAO,CAAC,SAAA,KAAsB,IAAA,GAAO,SAAA,GAAY,OAAA;AAAA,EACnD;AAEA,EAAA,MAAM,cAAA,GAAiB,qBAAA,CAAsB,CAAA,CAAE,KAAK,CAAA;AACpD,EAAA,MAAM,UAAU,CAAA,CAAE,OAAA;AAElB,EAAA,OAAO,CAAC,SAAA,KAAsB;AAC5B,IAAA,IAAI,GAAA,GAAM,SAAA;AACV,IAAA,IAAI,MAAA,GAAS,IAAA;AAEb,IAAA,KAAA,IAAS,IAAA,GAAO,CAAA,EAAG,IAAA,GAAO,cAAA,CAAe,QAAQ,IAAA,EAAA,EAAQ;AACvD,MAAA,MAAM,iBAAA,GAAoB,cAAA,CAAe,IAAI,CAAA,IAAK,CAAA;AAClD,MAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,KAAA,CAAM,GAAA,GAAM,iBAAiB,CAAA;AAChD,MAAA,GAAA,IAAO,KAAA,GAAQ,iBAAA;AACf,MAAA,MAAA,IAAU,KAAA,IAAS,OAAA,CAAQ,IAAI,CAAA,IAAK,CAAA,CAAA;AAAA,IACtC;AAEA,IAAA,OAAO,MAAA;AAAA,EACT,CAAA;AACF;AAWO,SAAS,kBAAA,CAAmB,KAAA,EAAe,IAAA,EAAc,MAAA,EAAsB;AACpF,EAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,KAAK,CAAA,EAAG;AAC3B,IAAA,MAAM,IAAI,mBAAA;AAAA,MACR,GAAG,IAAI,CAAA,yCAAA,EAA4C,OAAO,KAAK,CAAC,OAAO,MAAM,CAAA;AAAA,KAC/E;AAAA,EACF;AACF;AAOO,SAAS,gBAAA,CAAiB,GAAW,IAAA,EAAoB;AAC9D,EAAA,IAAI,CAAA,CAAE,QAAQ,CAAA,EAAG;AACjB,EAAA,IAAI,CAAA,CAAE,SAAS,CAAA,IAAA,CAAM,CAAA,CAAE,MAAM,CAAC,CAAA,IAAK,OAAO,CAAA,EAAG;AAC7C,EAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,8BAAA,CAAgC,CAAA;AAC9D;AAOO,SAAS,qBAAA,CAAsB,CAAA,EAAW,CAAA,EAAW,KAAA,EAAe,KAAA,EAAqB;AAC9F,EAAA,IAAI,CAAA,CAAE,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACrB,IAAA,MAAM,IAAI,UAAA;AAAA,MACR,CAAA,EAAG,KAAK,CAAA,OAAA,EAAU,CAAA,CAAE,IAAI,CAAA,MAAA,EAAS,KAAK,CAAA,OAAA,EAAU,CAAA,CAAE,IAAI,CAAA,qBAAA;AAAA,KACxD;AAAA,EACF;AACA,EAAA,gBAAA,CAAiB,GAAG,KAAK,CAAA;AACzB,EAAA,gBAAA,CAAiB,GAAG,KAAK,CAAA;AAC3B;AAOO,SAAS,cAAA,CAAe,CAAA,EAAW,CAAA,EAAW,KAAA,EAAe,KAAA,EAAqB;AACvF,EAAA,IAAI,CAAA,CAAE,IAAA,KAAS,CAAA,CAAE,IAAA,EAAM;AACrB,IAAA,MAAM,IAAI,UAAA;AAAA,MACR,CAAA,EAAG,KAAK,CAAA,OAAA,EAAU,CAAA,CAAE,IAAI,CAAA,MAAA,EAAS,KAAK,CAAA,OAAA,EAAU,CAAA,CAAE,IAAI,CAAA,qBAAA;AAAA,KACxD;AAAA,EACF;AACF;;;AC5GA,SAAS,oBAAoB,CAAA,EAA8B;AACzD,EAAA,IAAI,CAAA,CAAE,UAAU,QAAA,EAAU;AACxB,IAAA,MAAM,IAAI,WAAW,sCAAsC,CAAA;AAAA,EAC7D;AACA,EAAA,IAAI,CAAA,CAAE,UAAU,OAAA,EAAS;AACvB,IAAA,MAAM,IAAI,WAAW,sCAAsC,CAAA;AAAA,EAC7D;AAEA,EAAA,MAAM,OAAO,CAAA,CAAE,IAAA;AACf,EAAA,IAAI,CAAC,YAAA,CAAa,IAAI,KAAK,CAAC,mBAAA,CAAoB,IAAI,CAAA,EAAG;AACrD,IAAA,MAAM,IAAI,WAAW,iCAAiC,CAAA;AAAA,EACxD;AACA,EAAA,OAAO,IAAA;AACT;AAEA,SAAS,gBAAA,CACP,IAAA,EACA,SAAA,EACA,KAAA,EACA,IAAA,EACA;AACA,EAAA,MAAM,KAAA,GAAQ,iBAAA,CAAkB,IAAA,EAAM,SAAA,CAAU,KAAK,CAAC,CAAA;AACtD,EAAA,kBAAA,CAAmB,KAAA,EAAO,IAAA,EAAM,CAAA,MAAA,EAAS,KAAK,CAAA,CAAE,CAAA;AAChD,EAAA,OAAO,KAAA;AACT;AAEA,SAAS,iBAAA,CAAkB,KAAA,EAAe,IAAA,EAAc,KAAA,EAAqB;AAC3E,EAAA,IAAI,KAAA,KAAU,CAAA,IAAK,KAAA,KAAU,CAAA,EAAG;AAC9B,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,GAAG,IAAI,CAAA,iDAAA,EAAoD,OAAO,KAAK,CAAC,aAAa,KAAK,CAAA,CAAA;AAAA,MAC1F,IAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AACF;AAEA,SAAS,kBAAA,CAAmB,OAAe,KAAA,EAAqB;AAC9D,EAAA,IAAI,KAAA,CAAM,KAAA,KAAU,QAAA,IAAY,KAAA,CAAM,UAAU,QAAA,EAAU;AACxD,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,8DAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AACrC,IAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AAAA,EACvC;AACF;AAIA,SAAS,0BAAA,CAA2B,OAAe,KAAA,EAA0B;AAC3E,EAAA,MAAM,QAAA,GACJ,MAAM,KAAA,KAAU,QAAA,GAAW,WAAW,KAAA,CAAM,KAAA,KAAU,UAAU,OAAA,GAAU,SAAA;AAC5E,EAAA,MAAM,QAAA,GACJ,MAAM,KAAA,KAAU,QAAA,GAAW,WAAW,KAAA,CAAM,KAAA,KAAU,UAAU,OAAA,GAAU,SAAA;AAE5E,EAAA,IAAI,aAAa,QAAA,EAAU;AACzB,IAAA,MAAM,IAAI,WAAW,iDAAiD,CAAA;AAAA,EACxE;AAEA,EAAA,OAAO,QAAA;AACT;AAEA,SAAS,mBAAA,CAAoB,CAAA,EAAW,SAAA,EAA0B,KAAA,EAAe,IAAA,EAAc;AAC7F,EAAA,MAAM,MAAA,GAAS,UAAU,KAAK,CAAA;AAC9B,EAAA,MAAM,OAAO,CAAA,CAAE,IAAA;AAEf,EAAA,IAAI,KAAA,CAAM,OAAA,CAAQ,IAAI,CAAA,EAAG;AACvB,IAAA,OAAO,gBAAA,CAAiB,MAAM,MAAM,CAAA;AAAA,EACtC;AAEA,EAAA,IAAI,gBAAgB,aAAA,EAAe;AACjC,IAAA,OAAO,gBAAA,CAAiB,MAAM,MAAM,CAAA;AAAA,EACtC;AAEA,EAAA,IAAI,CAAC,YAAA,CAAa,IAAI,KAAK,CAAC,mBAAA,CAAoB,IAAI,CAAA,EAAG;AACrD,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,iCAAA,CAAmC,CAAA;AAAA,EACjE;AAEA,EAAA,MAAM,KAAA,GAAQ,iBAAA,CAAkB,IAAA,EAAM,MAAM,CAAA;AAC5C,EAAA,kBAAA,CAAmB,KAAA,EAAO,IAAA,EAAM,CAAA,MAAA,EAAS,KAAK,CAAA,CAAE,CAAA;AAChD,EAAA,OAAO,KAAA;AACT;AASA,SAAS,eAAA,CAAgB,OAAe,KAAA,EAAe;AACrD,EAAA,0BAAA,CAA2B,OAAO,KAAK,CAAA;AACvC,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,MAAM,KAAA,uBAAY,GAAA,EAA0C;AAC5D,EAAA,IAAI,OAAA,GAAU,CAAA;AACd,EAAA,IAAI,OAAA,GAAU,CAAA;AACd,EAAA,IAAI,OAAA,GAAU,CAAA;AAEd,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AACjE,IAAA,MAAM,OAAA,GAAU,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AAEjE,IAAA,IAAI,SAAA,GAAY,KAAA,CAAM,GAAA,CAAI,OAAO,CAAA;AACjC,IAAA,IAAI,CAAC,SAAA,EAAW;AACd,MAAA,SAAA,GAAY,EAAE,IAAI,CAAA,EAAG,EAAA,EAAI,GAAG,EAAA,EAAI,CAAA,EAAG,SAAS,CAAA,EAAE;AAC9C,MAAA,KAAA,CAAM,GAAA,CAAI,SAAS,SAAS,CAAA;AAAA,IAC9B;AAEA,IAAA,IAAI,SAAA,GAAY,KAAA,CAAM,GAAA,CAAI,OAAO,CAAA;AACjC,IAAA,IAAI,CAAC,SAAA,EAAW;AACd,MAAA,SAAA,GAAY,EAAE,IAAI,CAAA,EAAG,EAAA,EAAI,GAAG,EAAA,EAAI,CAAA,EAAG,SAAS,CAAA,EAAE;AAC9C,MAAA,KAAA,CAAM,GAAA,CAAI,SAAS,SAAS,CAAA;AAAA,IAC9B;AAEA,IAAA,SAAA,CAAU,OAAA,IAAW,CAAA;AAErB,IAAA,IAAI,YAAY,OAAA,EAAS;AACvB,MAAA,SAAA,CAAU,EAAA,IAAM,CAAA;AAChB,MAAA,OAAA,IAAW,CAAA;AAAA,IACb,CAAA,MAAO;AACL,MAAA,SAAA,CAAU,EAAA,IAAM,CAAA;AAChB,MAAA,SAAA,CAAU,EAAA,IAAM,CAAA;AAChB,MAAA,OAAA,IAAW,CAAA;AACX,MAAA,OAAA,IAAW,CAAA;AAAA,IACb;AAAA,EACF;AAEA,EAAA,MAAM,OAAA,GAAU,KAAA,CAAM,IAAA,CAAK,KAAA,CAAM,IAAA,EAAM,CAAA,CAAE,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,KAAM;AACtD,IAAA,IAAI,OAAO,CAAA,KAAM,QAAA,IAAY,OAAO,CAAA,KAAM,QAAA,SAAiB,CAAA,GAAI,CAAA;AAC/D,IAAA,IAAI,OAAO,MAAM,QAAA,IAAY,OAAO,MAAM,QAAA,EAAU,OAAO,CAAA,CAAE,aAAA,CAAc,CAAC,CAAA;AAC5E,IAAA,IAAI,OAAO,CAAA,KAAM,QAAA,IAAY,OAAO,CAAA,KAAM,QAAA,EAAU,OAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,EAAA,GAAK,CAAA;AACtF,IAAA,OAAO,OAAO,CAAC,CAAA,CAAE,aAAA,CAAc,MAAA,CAAO,CAAC,CAAC,CAAA;AAAA,EAC1C,CAAC,CAAA;AACD,EAAA,OAAO,EAAE,OAAA,EAAS,KAAA,EAAO,OAAA,EAAS,SAAS,OAAA,EAAQ;AACrD;AAiCO,SAAS,QAAA,CAAS,OAAe,KAAA,EAAuB;AAE7D,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAGpD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,0BAAA,CAA2B,OAAO,KAAK,CAAA;AACvC,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAG5C,EAAA,IAAI,OAAA,GAAU,CAAA;AACd,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AACjE,IAAA,MAAM,OAAA,GAAU,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AACjE,IAAA,IAAI,YAAY,OAAA,EAAS;AACvB,MAAA,OAAA,EAAA;AAAA,IACF;AAAA,EACF;AAGA,EAAA,OAAO,UAAU,KAAA,CAAM,IAAA;AACzB;AAqDO,SAAS,SAAA,CACd,KAAA,EACA,KAAA,EACA,OAAA,GAA4D,QAAA,EACzC;AAEnB,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAGpD,EAAA,IAAI,MAAM,IAAA,KAAS,CAAA,SAAU,OAAA,KAAY,IAAA,GAAO,EAAC,GAAI,CAAA;AAGrD,EAAA,IAAI,YAAY,QAAA,EAAU;AACxB,IAAA,IAAI,KAAA,CAAM,KAAA,KAAU,QAAA,IAAY,KAAA,CAAM,UAAU,QAAA,EAAU;AACxD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR;AAAA,OACF;AAAA,IACF;AACA,IAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,IAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,IAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,IAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,IAAA,IAAI,EAAA,GAAK,CAAA;AACT,IAAA,IAAI,EAAA,GAAK,CAAA;AAET,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,MAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,MAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,MAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AACrC,MAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AAErC,MAAA,IAAI,YAAY,CAAA,EAAG;AACjB,QAAA,IAAI,YAAY,CAAA,EAAG;AACjB,UAAA,EAAA,EAAA;AAAA,QACF,CAAA,MAAO;AACL,UAAA,EAAA,EAAA;AAAA,QACF;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,EAAA,GAAK,EAAA,KAAO,CAAA,GAAI,CAAA,GAAI,MAAM,EAAA,GAAK,EAAA,CAAA;AAAA,EACxC;AAEA,EAAA,MAAM,EAAE,SAAS,KAAA,EAAO,OAAA,EAAS,SAAQ,GAAI,eAAA,CAAgB,OAAO,KAAK,CAAA;AAEzE,EAAA,MAAM,aAAuB,EAAC;AAC9B,EAAA,MAAM,WAAqB,EAAC;AAC5B,EAAA,KAAA,MAAW,OAAO,OAAA,EAAS;AACzB,IAAA,MAAM,UAAA,GAAa,KAAA,CAAM,GAAA,CAAI,GAAG,CAAA;AAChC,IAAA,MAAM,EAAA,GAAK,YAAY,EAAA,IAAM,CAAA;AAC7B,IAAA,MAAM,EAAA,GAAK,YAAY,EAAA,IAAM,CAAA;AAC7B,IAAA,MAAM,OAAA,GAAU,YAAY,OAAA,IAAW,CAAA;AACvC,IAAA,UAAA,CAAW,KAAK,EAAA,GAAK,EAAA,KAAO,IAAI,CAAA,GAAI,EAAA,IAAM,KAAK,EAAA,CAAG,CAAA;AAClD,IAAA,QAAA,CAAS,KAAK,OAAO,CAAA;AAAA,EACvB;AAEA,EAAA,IAAI,YAAY,IAAA,EAAM;AACpB,IAAA,OAAO,UAAA;AAAA,EACT;AAEA,EAAA,IAAI,YAAY,OAAA,EAAS;AACvB,IAAA,OAAO,OAAA,GAAU,OAAA,KAAY,CAAA,GAAI,CAAA,GAAI,WAAW,OAAA,GAAU,OAAA,CAAA;AAAA,EAC5D;AAEA,EAAA,IAAI,YAAY,OAAA,EAAS;AACvB,IAAA,MAAM,GAAA,GAAM,WAAW,MAAA,CAAO,CAAC,KAAK,GAAA,KAAQ,GAAA,GAAM,KAAK,CAAC,CAAA;AACxD,IAAA,OAAO,UAAA,CAAW,MAAA,KAAW,CAAA,GAAI,CAAA,GAAI,MAAM,UAAA,CAAW,MAAA;AAAA,EACxD;AAEA,EAAA,IAAI,YAAY,UAAA,EAAY;AAC1B,IAAA,IAAI,WAAA,GAAc,CAAA;AAClB,IAAA,IAAI,YAAA,GAAe,CAAA;AAEnB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,UAAA,CAAW,QAAQ,CAAA,EAAA,EAAK;AAC1C,MAAA,WAAA,IAAA,CAAgB,WAAW,CAAC,CAAA,IAAK,CAAA,KAAM,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA,CAAA;AACtD,MAAA,YAAA,IAAgB,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA;AAAA,IACjC;AAEA,IAAA,OAAO,YAAA,KAAiB,CAAA,GAAI,CAAA,GAAI,WAAA,GAAc,YAAA;AAAA,EAChD;AAEA,EAAA,MAAM,IAAI,qBAAA;AAAA,IACR,8BAA8B,OAAO,CAAA,iEAAA,CAAA;AAAA,IACrC,SAAA;AAAA,IACA;AAAA,GACF;AACF;AA0CO,SAAS,MAAA,CACd,KAAA,EACA,KAAA,EACA,OAAA,GAA4D,QAAA,EACzC;AAEnB,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAGpD,EAAA,IAAI,MAAM,IAAA,KAAS,CAAA,SAAU,OAAA,KAAY,IAAA,GAAO,EAAC,GAAI,CAAA;AAGrD,EAAA,IAAI,YAAY,QAAA,EAAU;AACxB,IAAA,IAAI,KAAA,CAAM,KAAA,KAAU,QAAA,IAAY,KAAA,CAAM,UAAU,QAAA,EAAU;AACxD,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR;AAAA,OACF;AAAA,IACF;AACA,IAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,IAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,IAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,IAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,IAAA,IAAI,EAAA,GAAK,CAAA;AACT,IAAA,IAAI,EAAA,GAAK,CAAA;AAET,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,MAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,MAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,MAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AACrC,MAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AAErC,MAAA,IAAI,YAAY,CAAA,EAAG;AACjB,QAAA,IAAI,YAAY,CAAA,EAAG;AACjB,UAAA,EAAA,EAAA;AAAA,QACF,CAAA,MAAO;AACL,UAAA,EAAA,EAAA;AAAA,QACF;AAAA,MACF;AAAA,IACF;AAEA,IAAA,OAAO,EAAA,GAAK,EAAA,KAAO,CAAA,GAAI,CAAA,GAAI,MAAM,EAAA,GAAK,EAAA,CAAA;AAAA,EACxC;AAEA,EAAA,MAAM,EAAE,SAAS,KAAA,EAAO,OAAA,EAAS,SAAQ,GAAI,eAAA,CAAgB,OAAO,KAAK,CAAA;AAEzE,EAAA,MAAM,UAAoB,EAAC;AAC3B,EAAA,MAAM,WAAqB,EAAC;AAC5B,EAAA,KAAA,MAAW,OAAO,OAAA,EAAS;AACzB,IAAA,MAAM,UAAA,GAAa,KAAA,CAAM,GAAA,CAAI,GAAG,CAAA;AAChC,IAAA,MAAM,EAAA,GAAK,YAAY,EAAA,IAAM,CAAA;AAC7B,IAAA,MAAM,EAAA,GAAK,YAAY,EAAA,IAAM,CAAA;AAC7B,IAAA,MAAM,OAAA,GAAU,YAAY,OAAA,IAAW,CAAA;AACvC,IAAA,OAAA,CAAQ,KAAK,EAAA,GAAK,EAAA,KAAO,IAAI,CAAA,GAAI,EAAA,IAAM,KAAK,EAAA,CAAG,CAAA;AAC/C,IAAA,QAAA,CAAS,KAAK,OAAO,CAAA;AAAA,EACvB;AAEA,EAAA,IAAI,YAAY,IAAA,EAAM;AACpB,IAAA,OAAO,OAAA;AAAA,EACT;AAEA,EAAA,IAAI,YAAY,OAAA,EAAS;AACvB,IAAA,OAAO,OAAA,GAAU,OAAA,KAAY,CAAA,GAAI,CAAA,GAAI,WAAW,OAAA,GAAU,OAAA,CAAA;AAAA,EAC5D;AAEA,EAAA,IAAI,YAAY,OAAA,EAAS;AACvB,IAAA,MAAM,GAAA,GAAM,QAAQ,MAAA,CAAO,CAAC,KAAK,GAAA,KAAQ,GAAA,GAAM,KAAK,CAAC,CAAA;AACrD,IAAA,OAAO,OAAA,CAAQ,MAAA,KAAW,CAAA,GAAI,CAAA,GAAI,MAAM,OAAA,CAAQ,MAAA;AAAA,EAClD;AAEA,EAAA,IAAI,YAAY,UAAA,EAAY;AAC1B,IAAA,IAAI,WAAA,GAAc,CAAA;AAClB,IAAA,IAAI,YAAA,GAAe,CAAA;AAEnB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,OAAA,CAAQ,QAAQ,CAAA,EAAA,EAAK;AACvC,MAAA,WAAA,IAAA,CAAgB,QAAQ,CAAC,CAAA,IAAK,CAAA,KAAM,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA,CAAA;AACnD,MAAA,YAAA,IAAgB,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA;AAAA,IACjC;AAEA,IAAA,OAAO,YAAA,KAAiB,CAAA,GAAI,CAAA,GAAI,WAAA,GAAc,YAAA;AAAA,EAChD;AAEA,EAAA,MAAM,IAAI,qBAAA;AAAA,IACR,8BAA8B,OAAO,CAAA,iEAAA,CAAA;AAAA,IACrC,SAAA;AAAA,IACA;AAAA,GACF;AACF;AA0CO,SAAS,OAAA,CACd,KAAA,EACA,KAAA,EACA,OAAA,GAA4D,QAAA,EACzC;AAInB,EAAA,IAAI,OAAA,KAAY,QAAA,IAAY,OAAA,KAAY,OAAA,EAAS;AAC/C,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,KAAA,EAAO,KAAA,EAAO,OAAO,CAAA;AACzC,IAAA,MAAM,CAAA,GAAI,MAAA,CAAO,KAAA,EAAO,KAAA,EAAO,OAAO,CAAA;AACtC,IAAA,OAAO,IAAI,CAAA,KAAM,CAAA,GAAI,IAAK,CAAA,GAAI,CAAA,GAAI,KAAM,CAAA,GAAI,CAAA,CAAA;AAAA,EAC9C;AAGA,EAAA,MAAM,IAAA,GAAO,SAAA,CAAU,KAAA,EAAO,KAAA,EAAO,IAAI,CAAA;AACzC,EAAA,MAAM,GAAA,GAAM,MAAA,CAAO,KAAA,EAAO,KAAA,EAAO,IAAI,CAAA;AAGrC,EAAA,MAAM,WAAqB,EAAC;AAC5B,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,QAAQ,CAAA,EAAA,EAAK;AACpC,IAAA,MAAM,CAAA,GAAI,IAAA,CAAK,CAAC,CAAA,IAAK,CAAA;AACrB,IAAA,MAAM,CAAA,GAAI,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA;AACpB,IAAA,QAAA,CAAS,IAAA,CAAK,IAAI,CAAA,KAAM,CAAA,GAAI,IAAK,CAAA,GAAI,CAAA,GAAI,CAAA,IAAM,CAAA,GAAI,CAAA,CAAE,CAAA;AAAA,EACvD;AAEA,EAAA,IAAI,YAAY,IAAA,EAAM;AACpB,IAAA,OAAO,QAAA;AAAA,EACT;AAEA,EAAA,IAAI,QAAA,CAAS,MAAA,KAAW,CAAA,EAAG,OAAO,CAAA;AAElC,EAAA,IAAI,YAAY,OAAA,EAAS;AACvB,IAAA,MAAM,GAAA,GAAM,SAAS,MAAA,CAAO,CAAC,KAAK,GAAA,KAAQ,GAAA,GAAM,KAAK,CAAC,CAAA;AACtD,IAAA,OAAO,MAAM,QAAA,CAAS,MAAA;AAAA,EACxB;AAEA,EAAA,IAAI,YAAY,UAAA,EAAY;AAC1B,IAAA,MAAM,EAAE,OAAA,EAAS,KAAA,EAAM,GAAI,eAAA,CAAgB,OAAO,KAAK,CAAA;AACvD,IAAA,IAAI,WAAA,GAAc,CAAA;AAClB,IAAA,IAAI,YAAA,GAAe,CAAA;AACnB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,CAAS,QAAQ,CAAA,EAAA,EAAK;AACxC,MAAA,MAAM,GAAA,GAAM,QAAQ,CAAC,CAAA;AACrB,MAAA,MAAM,OAAA,GAAU,QAAQ,MAAA,GAAa,KAAA,CAAM,IAAI,GAAG,CAAA,EAAG,WAAW,CAAA,GAAK,CAAA;AACrE,MAAA,WAAA,IAAA,CAAgB,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA,IAAK,OAAA;AACpC,MAAA,YAAA,IAAgB,OAAA;AAAA,IAClB;AACA,IAAA,OAAO,YAAA,KAAiB,CAAA,GAAI,CAAA,GAAI,WAAA,GAAc,YAAA;AAAA,EAChD;AAEA,EAAA,MAAM,IAAI,qBAAA;AAAA,IACR,8BAA8B,OAAO,CAAA,iEAAA,CAAA;AAAA,IACrC,SAAA;AAAA,IACA;AAAA,GACF;AACF;AA8CO,SAAS,UAAA,CACd,KAAA,EACA,KAAA,EACA,IAAA,EACA,UAA4D,QAAA,EAC5D;AACA,EAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,IAAI,CAAA,IAAK,QAAQ,CAAA,EAAG;AACvC,IAAA,MAAM,IAAI,qBAAA,CAAsB,uCAAA,EAAyC,MAAA,EAAQ,IAAI,CAAA;AAAA,EACvF;AAEA,EAAA,MAAM,SAAS,IAAA,GAAO,IAAA;AAKtB,EAAA,IAAI,OAAA,KAAY,QAAA,IAAY,OAAA,KAAY,OAAA,EAAS;AAC/C,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,KAAA,EAAO,KAAA,EAAO,OAAO,CAAA;AACzC,IAAA,MAAM,CAAA,GAAI,MAAA,CAAO,KAAA,EAAO,KAAA,EAAO,OAAO,CAAA;AACtC,IAAA,OAAO,CAAA,GAAI,MAAM,CAAA,GAAI,CAAA,GAAA,CAAM,IAAI,MAAA,IAAU,CAAA,GAAI,CAAA,IAAM,MAAA,GAAS,CAAA,GAAI,CAAA,CAAA;AAAA,EAClE;AAGA,EAAA,MAAM,IAAA,GAAO,SAAA,CAAU,KAAA,EAAO,KAAA,EAAO,IAAI,CAAA;AACzC,EAAA,MAAM,GAAA,GAAM,MAAA,CAAO,KAAA,EAAO,KAAA,EAAO,IAAI,CAAA;AAGrC,EAAA,MAAM,cAAwB,EAAC;AAC/B,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,QAAQ,CAAA,EAAA,EAAK;AACpC,IAAA,MAAM,CAAA,GAAI,IAAA,CAAK,CAAC,CAAA,IAAK,CAAA;AACrB,IAAA,MAAM,CAAA,GAAI,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA;AACpB,IAAA,WAAA,CAAY,IAAA,CAAK,CAAA,GAAI,CAAA,KAAM,CAAA,GAAI,CAAA,GAAA,CAAM,CAAA,GAAI,MAAA,IAAU,CAAA,GAAI,CAAA,IAAM,MAAA,GAAS,CAAA,GAAI,CAAA,CAAE,CAAA;AAAA,EAC9E;AAEA,EAAA,IAAI,YAAY,IAAA,EAAM;AACpB,IAAA,OAAO,WAAA;AAAA,EACT;AAEA,EAAA,IAAI,WAAA,CAAY,MAAA,KAAW,CAAA,EAAG,OAAO,CAAA;AAErC,EAAA,IAAI,YAAY,OAAA,EAAS;AACvB,IAAA,MAAM,GAAA,GAAM,YAAY,MAAA,CAAO,CAAC,KAAK,GAAA,KAAQ,GAAA,GAAM,KAAK,CAAC,CAAA;AACzD,IAAA,OAAO,MAAM,WAAA,CAAY,MAAA;AAAA,EAC3B;AAEA,EAAA,IAAI,YAAY,UAAA,EAAY;AAC1B,IAAA,MAAM,EAAE,OAAA,EAAS,KAAA,EAAM,GAAI,eAAA,CAAgB,OAAO,KAAK,CAAA;AACvD,IAAA,IAAI,WAAA,GAAc,CAAA;AAClB,IAAA,IAAI,YAAA,GAAe,CAAA;AACnB,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,WAAA,CAAY,QAAQ,CAAA,EAAA,EAAK;AAC3C,MAAA,MAAM,GAAA,GAAM,QAAQ,CAAC,CAAA;AACrB,MAAA,MAAM,OAAA,GAAU,QAAQ,MAAA,GAAa,KAAA,CAAM,IAAI,GAAG,CAAA,EAAG,WAAW,CAAA,GAAK,CAAA;AACrE,MAAA,WAAA,IAAA,CAAgB,WAAA,CAAY,CAAC,CAAA,IAAK,CAAA,IAAK,OAAA;AACvC,MAAA,YAAA,IAAgB,OAAA;AAAA,IAClB;AACA,IAAA,OAAO,YAAA,KAAiB,CAAA,GAAI,CAAA,GAAI,WAAA,GAAc,YAAA;AAAA,EAChD;AAEA,EAAA,MAAM,IAAI,qBAAA;AAAA,IACR,8BAA8B,OAAO,CAAA,iEAAA,CAAA;AAAA,IACrC,SAAA;AAAA,IACA;AAAA,GACF;AACF;AAiCO,SAAS,eAAA,CAAgB,OAAe,KAAA,EAAuB;AAEpE,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,0BAAA,CAA2B,OAAO,KAAK,CAAA;AAEvC,EAAA,IAAI,KAAA,CAAM,SAAS,CAAA,EAAG;AACpB,IAAA,OAAO,MAAA,CAAO,EAAE,CAAA,CAAE,QAAQ,CAAC,CAAA,EAAG,CAAC,CAAC,CAAA;AAAA,EAClC;AAEA,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAG5C,EAAA,MAAM,QAAA,uBAAe,GAAA,EAA8B;AACnD,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,QAAA,CAAS,IAAI,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,CAAA,EAAG,OAAO,CAAC,CAAA;AAC/D,IAAA,QAAA,CAAS,IAAI,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,CAAA,EAAG,OAAO,CAAC,CAAA;AAAA,EACjE;AAGA,EAAA,MAAM,MAAA,GAAS,MAAM,IAAA,CAAK,QAAQ,EAAE,IAAA,CAAK,CAAC,GAAG,CAAA,KAAM;AACjD,IAAA,IAAI,OAAO,CAAA,KAAM,QAAA,IAAY,OAAO,CAAA,KAAM,QAAA,SAAiB,CAAA,GAAI,CAAA;AAC/D,IAAA,IAAI,OAAO,MAAM,QAAA,IAAY,OAAO,MAAM,QAAA,EAAU,OAAO,CAAA,CAAE,aAAA,CAAc,CAAC,CAAA;AAC5E,IAAA,IAAI,OAAO,CAAA,KAAM,QAAA,IAAY,OAAO,CAAA,KAAM,QAAA,EAAU,OAAO,CAAA,KAAM,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,CAAA,GAAI,EAAA,GAAK,CAAA;AACtF,IAAA,OAAO,OAAO,CAAC,CAAA,CAAE,aAAA,CAAc,MAAA,CAAO,CAAC,CAAC,CAAA;AAAA,EAC1C,CAAC,CAAA;AAGD,EAAA,MAAM,YAAA,uBAAmB,GAAA,EAAsC;AAC/D,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,MAAA,CAAO,QAAQ,CAAA,EAAA,EAAK;AACtC,IAAA,MAAM,KAAA,GAAQ,OAAO,CAAC,CAAA;AACtB,IAAA,IAAI,UAAU,MAAA,EAAW;AACzB,IAAA,YAAA,CAAa,GAAA,CAAI,OAAO,CAAC,CAAA;AAAA,EAC3B;AAGA,EAAA,MAAM,WAAW,MAAA,CAAO,MAAA;AACxB,EAAA,MAAM,MAAA,GAAS,KAAA,CAAM,IAAA,CAAK,EAAE,QAAQ,QAAA,EAAS,EAAG,MAAM,IAAI,KAAA,CAAc,QAAQ,CAAA,CAAE,IAAA,CAAK,CAAC,CAAC,CAAA;AAGzF,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,SAAA,GAAY,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AACnE,IAAA,MAAM,SAAA,GAAY,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AAGnE,IAAA,MAAM,CAAA,GAAI,YAAA,CAAa,GAAA,CAAI,SAAS,CAAA;AACpC,IAAA,MAAM,CAAA,GAAI,YAAA,CAAa,GAAA,CAAI,SAAS,CAAA;AACpC,IAAA,IAAI,CAAA,KAAM,MAAA,IAAa,CAAA,KAAM,MAAA,EAAW;AAGxC,IAAA,MAAM,GAAA,GAAM,OAAO,CAAC,CAAA;AACpB,IAAA,IAAI,KAAK,GAAA,CAAI,CAAC,KAAK,GAAA,CAAI,CAAC,KAAK,CAAA,IAAK,CAAA;AAAA,EACpC;AAGA,EAAA,OAAO,OAAO,MAAM,CAAA;AACtB;AAoCO,SAAS,oBAAA,CAAqB,OAAe,KAAA,EAAuB;AAEzE,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAGpD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,mCAAA;AAE7B,EAAA,kBAAA,CAAmB,OAAO,KAAK,CAAA;AAG/B,EAAA,MAAM,EAAE,OAAA,EAAS,KAAA,EAAM,GAAI,eAAA,CAAgB,OAAO,KAAK,CAAA;AACvD,EAAA,MAAM,KAAA,GAAQ,SAAA,CAAU,KAAA,EAAO,KAAA,EAAO,IAAI,CAAA;AAC1C,EAAA,MAAM,IAAA,GAAO,MAAA,CAAO,KAAA,EAAO,KAAA,EAAO,IAAI,CAAA;AACtC,EAAA,MAAM,GAAA,GAAM,OAAA,CAAQ,KAAA,EAAO,KAAA,EAAO,IAAI,CAAA;AACtC,EAAA,MAAM,GAAA,GAAM,QAAA,CAAS,KAAA,EAAO,KAAK,CAAA;AAGjC,EAAA,MAAM,WAAA,GAAc,IAAA,CAAK,GAAA,CAAI,GAAG,QAAQ,GAAA,CAAI,CAAC,CAAA,KAAM,MAAA,CAAO,CAAC,CAAA,CAAE,MAAM,CAAA,EAAG,QAAQ,MAAM,CAAA;AACpF,EAAA,MAAM,QAAA,GAAW,IAAA,CAAK,GAAA,CAAI,EAAA,EAAI,cAAc,CAAC,CAAA;AAG7C,EAAA,IAAI,MAAA,GAAS,0BAAA;AACb,EAAA,MAAA,IACE,OAAA,CAAQ,MAAA,CAAO,QAAQ,CAAA,GACvB,YAAY,MAAA,CAAO,EAAE,CAAA,GACrB,QAAA,CAAS,OAAO,EAAE,CAAA,GAClB,UAAA,CAAW,MAAA,CAAO,EAAE,CAAA,GACpB,WAAA;AACF,EAAA,MAAA,IAAU,GAAG,GAAA,CAAI,MAAA,CAAO,QAAA,GAAW,EAAA,GAAK,CAAC,CAAC;AAAA,CAAA;AAE1C,EAAA,IAAI,YAAA,GAAe,CAAA;AACnB,EAAA,IAAI,YAAA,GAAe,CAAA;AACnB,EAAA,IAAI,WAAA,GAAc,CAAA;AAClB,EAAA,IAAI,UAAA,GAAa,CAAA;AACjB,EAAA,IAAI,SAAA,GAAY,CAAA;AAChB,EAAA,IAAI,QAAA,GAAW,CAAA;AACf,EAAA,IAAI,OAAA,GAAU,CAAA;AAEd,EAAA,KAAA,MAAW,CAAC,CAAA,EAAG,GAAG,CAAA,IAAK,OAAA,CAAQ,SAAQ,EAAG;AACxC,IAAA,MAAM,CAAA,GAAI,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACtB,IAAA,MAAM,CAAA,GAAI,IAAA,CAAK,CAAC,CAAA,IAAK,CAAA;AACrB,IAAA,MAAM,EAAA,GAAK,GAAA,CAAI,CAAC,CAAA,IAAK,CAAA;AACrB,IAAA,MAAM,CAAA,GAAI,KAAA,CAAM,GAAA,CAAI,GAAG,GAAG,OAAA,IAAW,CAAA;AAErC,IAAA,YAAA,IAAgB,CAAA;AAChB,IAAA,YAAA,IAAgB,CAAA,GAAI,CAAA;AACpB,IAAA,WAAA,IAAe,CAAA,GAAI,CAAA;AACnB,IAAA,UAAA,IAAc,EAAA,GAAK,CAAA;AACnB,IAAA,SAAA,IAAa,CAAA;AACb,IAAA,QAAA,IAAY,CAAA;AACZ,IAAA,OAAA,IAAW,EAAA;AAEX,IAAA,MAAA,IACE,MAAA,CAAO,GAAG,CAAA,CAAE,MAAA,CAAO,QAAQ,CAAA,GAC3B,CAAA,CAAE,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GACtB,CAAA,CAAE,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GACtB,EAAA,CAAG,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GACvB,MAAA,CAAO,CAAC,CAAA,GACR,IAAA;AAAA,EACJ;AAEA,EAAA,MAAA,IAAU,IAAA;AAIV,EAAA,MAAM,WAAW,OAAA,CAAQ,MAAA;AACzB,EAAA,IAAI,WAAW,CAAA,EAAG;AAChB,IAAA,SAAA,IAAa,QAAA;AACb,IAAA,QAAA,IAAY,QAAA;AACZ,IAAA,OAAA,IAAW,QAAA;AAAA,EACb,CAAA,MAAO;AACL,IAAA,SAAA,GAAY,CAAA;AACZ,IAAA,QAAA,GAAW,CAAA;AACX,IAAA,OAAA,GAAU,CAAA;AAAA,EACZ;AAEA,EAAA,YAAA,GAAe,YAAA,KAAiB,CAAA,GAAI,CAAA,GAAI,YAAA,GAAe,YAAA;AACvD,EAAA,WAAA,GAAc,YAAA,KAAiB,CAAA,GAAI,CAAA,GAAI,WAAA,GAAc,YAAA;AACrD,EAAA,UAAA,GAAa,YAAA,KAAiB,CAAA,GAAI,CAAA,GAAI,UAAA,GAAa,YAAA;AAEnD,EAAA,MAAA,IACE,UAAA,CAAW,OAAO,QAAQ,CAAA,GAC1B,GAAG,MAAA,CAAO,EAAE,IACZ,EAAA,CAAG,MAAA,CAAO,EAAE,CAAA,GACZ,GAAA,CAAI,QAAQ,CAAC,CAAA,CAAE,OAAO,EAAE,CAAA,GACxB,MAAA,CAAO,YAAY,CAAA,GACnB,IAAA;AACF,EAAA,MAAA,IACE,WAAA,CAAY,MAAA,CAAO,QAAQ,CAAA,GAC3B,SAAA,CAAU,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GAC9B,QAAA,CAAS,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GAC7B,OAAA,CAAQ,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GAC5B,MAAA,CAAO,YAAY,CAAA,GACnB,IAAA;AACF,EAAA,MAAA,IACE,cAAA,CAAe,MAAA,CAAO,QAAQ,CAAA,GAC9B,YAAA,CAAa,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GACjC,WAAA,CAAY,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GAChC,UAAA,CAAW,OAAA,CAAQ,CAAC,CAAA,CAAE,MAAA,CAAO,EAAE,CAAA,GAC/B,MAAA,CAAO,YAAY,CAAA;AAErB,EAAA,OAAO,MAAA;AACT;AA6BO,SAAS,QAAA,CAAS,OAAe,MAAA,EAA0C;AAChF,EAAA,qBAAA,CAAsB,KAAA,EAAO,MAAA,EAAQ,OAAA,EAAS,QAAQ,CAAA;AAEtD,EAAA,MAAM,IAAI,KAAA,CAAM,IAAA;AAEhB,EAAA,IAAI,CAAA,KAAM,CAAA,EAAG,OAAO,CAAC,OAAO,EAAE,CAAA,EAAG,MAAA,CAAO,EAAE,CAAA,EAAG,MAAA,CAAO,EAAE,CAAC,CAAA;AAEvD,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,UAAA,GAAa,oBAAoB,MAAM,CAAA;AAC7C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,WAAA,GAAc,oBAAoB,MAAM,CAAA;AAG9C,EAAA,MAAM,QAAiD,EAAC;AACxD,EAAA,IAAI,IAAA,GAAO,CAAA;AACX,EAAA,IAAI,IAAA,GAAO,CAAA;AACX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,KAAA,GAAQ,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAChE,IAAA,iBAAA,CAAkB,KAAA,EAAO,SAAS,CAAC,CAAA;AACnC,IAAA,MAAM,KAAA,GAAQ,gBAAA,CAAiB,UAAA,EAAY,WAAA,EAAa,GAAG,QAAQ,CAAA;AACnE,IAAA,KAAA,CAAM,IAAA,CAAK,EAAE,KAAA,EAAO,KAAA,EAAO,CAAA;AAC3B,IAAA,IAAI,UAAU,CAAA,EAAG,IAAA,EAAA;AAAA,SACZ,IAAA,EAAA;AAAA,EACP;AAGA,EAAA,KAAA,CAAM,KAAK,CAAC,CAAA,EAAG,MAAM,CAAA,CAAE,KAAA,GAAQ,EAAE,KAAK,CAAA;AAEtC,EAAA,IAAI,SAAS,CAAA,IAAK,IAAA,KAAS,CAAA,EAAG,OAAO,CAAC,MAAA,CAAO,EAAE,CAAA,EAAG,OAAO,EAAE,GAAG,MAAA,CAAO,EAAE,CAAC,CAAA;AAExE,EAAA,MAAM,GAAA,GAAM,CAAC,CAAC,CAAA;AACd,EAAA,MAAM,GAAA,GAAM,CAAC,CAAC,CAAA;AACd,EAAA,MAAM,UAAA,GAAa,CAAC,QAAQ,CAAA;AAE5B,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,IAAI,GAAA,GAAM,CAAA;AACV,EAAA,OAAO,GAAA,GAAM,MAAM,MAAA,EAAQ;AACzB,IAAA,MAAM,SAAA,GAAY,KAAA,CAAM,GAAG,CAAA,EAAG,KAAA,IAAS,CAAA;AAGvC,IAAA,OAAO,GAAA,GAAM,MAAM,MAAA,IAAA,CAAW,KAAA,CAAM,GAAG,CAAA,EAAG,KAAA,IAAS,OAAO,SAAA,EAAW;AACnE,MAAA,MAAM,KAAA,GAAQ,KAAA,CAAM,GAAG,CAAA,EAAG,KAAA,IAAS,CAAA;AACnC,MAAA,IAAI,UAAU,CAAA,EAAG,EAAA,EAAA;AAAA,WACZ,EAAA,EAAA;AACL,MAAA,GAAA,EAAA;AAAA,IACF;AAEA,IAAA,GAAA,CAAI,IAAA,CAAK,KAAK,IAAI,CAAA;AAClB,IAAA,GAAA,CAAI,IAAA,CAAK,KAAK,IAAI,CAAA;AAClB,IAAA,UAAA,CAAW,KAAK,SAAS,CAAA;AAAA,EAC3B;AAEA,EAAA,OAAO,CAAC,OAAO,GAAG,CAAA,EAAG,OAAO,GAAG,CAAA,EAAG,MAAA,CAAO,UAAU,CAAC,CAAA;AACtD;AAkCO,SAAS,WAAA,CAAY,OAAe,MAAA,EAAwB;AACjE,EAAA,MAAM,MAAA,GAAS,QAAA,CAAS,KAAA,EAAO,MAAM,CAAA;AACrC,EAAA,MAAM,IAAA,GAAO,OAAO,CAAC,CAAA;AACrB,EAAA,MAAM,IAAA,GAAO,OAAO,CAAC,CAAA;AACrB,EAAA,IAAI,CAAC,IAAA,IAAQ,CAAC,IAAA,IAAQ,IAAA,CAAK,SAAS,CAAA,IAAK,IAAA,CAAK,IAAA,KAAS,CAAA,EAAG,OAAO,GAAA;AAEjE,EAAA,MAAM,OAAA,GAAU,oBAAoB,IAAI,CAAA;AACxC,EAAA,MAAM,OAAA,GAAU,oBAAoB,IAAI,CAAA;AACxC,EAAA,MAAM,SAAA,GAAY,oBAAoB,IAAI,CAAA;AAC1C,EAAA,MAAM,SAAA,GAAY,oBAAoB,IAAI,CAAA;AAE1C,EAAA,IAAI,GAAA,GAAM,CAAA;AACV,EAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,EAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,MAAM,CAAA,EAAA,EAAK;AAClC,IAAA,MAAM,CAAA,GAAI,gBAAA,CAAiB,OAAA,EAAS,SAAA,EAAW,GAAG,KAAK,CAAA;AACvD,IAAA,MAAM,CAAA,GAAI,gBAAA,CAAiB,OAAA,EAAS,SAAA,EAAW,GAAG,KAAK,CAAA;AACvD,IAAA,GAAA,IAAA,CAAQ,CAAA,GAAI,KAAA,KAAA,CAAW,CAAA,GAAI,KAAA,IAAS,CAAA,CAAA;AACpC,IAAA,KAAA,GAAQ,CAAA;AACR,IAAA,KAAA,GAAQ,CAAA;AAAA,EACV;AAEA,EAAA,OAAO,GAAA;AACT;AAqCO,SAAS,oBAAA,CAAqB,OAAe,MAAA,EAA0C;AAC5F,EAAA,qBAAA,CAAsB,KAAA,EAAO,MAAA,EAAQ,OAAA,EAAS,QAAQ,CAAA;AAEtD,EAAA,MAAM,IAAI,KAAA,CAAM,IAAA;AAEhB,EAAA,IAAI,CAAA,KAAM,CAAA,EAAG,OAAO,CAAC,OAAO,EAAE,CAAA,EAAG,MAAA,CAAO,EAAE,CAAA,EAAG,MAAA,CAAO,EAAE,CAAC,CAAA;AAEvD,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,UAAA,GAAa,oBAAoB,MAAM,CAAA;AAC7C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,WAAA,GAAc,oBAAoB,MAAM,CAAA;AAG9C,EAAA,MAAM,QAAiD,EAAC;AACxD,EAAA,IAAI,IAAA,GAAO,CAAA;AACX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,KAAA,GAAQ,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAChE,IAAA,iBAAA,CAAkB,KAAA,EAAO,SAAS,CAAC,CAAA;AACnC,IAAA,MAAM,KAAA,GAAQ,gBAAA,CAAiB,UAAA,EAAY,WAAA,EAAa,GAAG,QAAQ,CAAA;AACnE,IAAA,KAAA,CAAM,IAAA,CAAK,EAAE,KAAA,EAAO,KAAA,EAAO,CAAA;AAC3B,IAAA,IAAI,UAAU,CAAA,EAAG,IAAA,EAAA;AAAA,EACnB;AAGA,EAAA,KAAA,CAAM,KAAK,CAAC,CAAA,EAAG,MAAM,CAAA,CAAE,KAAA,GAAQ,EAAE,KAAK,CAAA;AAEtC,EAAA,IAAI,IAAA,KAAS,CAAA,EAAG,OAAO,CAAC,OAAO,EAAE,CAAA,EAAG,MAAA,CAAO,EAAE,CAAA,EAAG,MAAA,CAAO,EAAE,CAAC,CAAA;AAE1D,EAAA,MAAM,IAAA,GAAO,CAAC,CAAC,CAAA;AACf,EAAA,MAAM,GAAA,GAAM,CAAC,CAAC,CAAA;AACd,EAAA,MAAM,UAAA,GAAa,CAAC,QAAQ,CAAA;AAE5B,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,IAAI,GAAA,GAAM,CAAA;AACV,EAAA,OAAO,GAAA,GAAM,MAAM,MAAA,EAAQ;AACzB,IAAA,MAAM,SAAA,GAAY,KAAA,CAAM,GAAG,CAAA,EAAG,KAAA,IAAS,CAAA;AACvC,IAAA,OAAO,GAAA,GAAM,MAAM,MAAA,IAAA,CAAW,KAAA,CAAM,GAAG,CAAA,EAAG,KAAA,IAAS,OAAO,SAAA,EAAW;AACnE,MAAA,MAAM,KAAA,GAAQ,KAAA,CAAM,GAAG,CAAA,EAAG,KAAA,IAAS,CAAA;AACnC,MAAA,IAAI,UAAU,CAAA,EAAG,EAAA,EAAA;AAAA,WACZ,EAAA,EAAA;AACL,MAAA,GAAA,EAAA;AAAA,IACF;AAEA,IAAA,MAAM,eAAe,EAAA,GAAK,EAAA,KAAO,CAAA,GAAI,CAAA,GAAI,MAAM,EAAA,GAAK,EAAA,CAAA;AACpD,IAAA,MAAM,YAAY,EAAA,GAAK,IAAA;AACvB,IAAA,IAAA,CAAK,KAAK,YAAY,CAAA;AACtB,IAAA,GAAA,CAAI,KAAK,SAAS,CAAA;AAClB,IAAA,UAAA,CAAW,KAAK,SAAS,CAAA;AAAA,EAC3B;AAEA,EAAA,OAAO,CAAC,OAAO,IAAI,CAAA,EAAG,OAAO,GAAG,CAAA,EAAG,MAAA,CAAO,UAAU,CAAC,CAAA;AACvD;AAmCO,SAAS,qBAAA,CAAsB,OAAe,MAAA,EAAwB;AAC3E,EAAA,MAAM,MAAA,GAAS,oBAAA,CAAqB,KAAA,EAAO,MAAM,CAAA;AACjD,EAAA,MAAM,KAAA,GAAQ,OAAO,CAAC,CAAA;AACtB,EAAA,MAAM,IAAA,GAAO,OAAO,CAAC,CAAA;AACrB,EAAA,IAAI,CAAC,KAAA,IAAS,CAAC,IAAA,IAAQ,KAAA,CAAM,SAAS,CAAA,IAAK,IAAA,CAAK,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAEnE,EAAA,MAAM,QAAA,GAAW,oBAAoB,KAAK,CAAA;AAC1C,EAAA,MAAM,OAAA,GAAU,oBAAoB,IAAI,CAAA;AACxC,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,SAAA,GAAY,oBAAoB,IAAI,CAAA;AAE1C,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,IAAI,UAAA,GAAa,gBAAA,CAAiB,OAAA,EAAS,SAAA,EAAW,GAAG,QAAQ,CAAA;AACjE,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,IAAA,CAAK,MAAM,CAAA,EAAA,EAAK;AAClC,IAAA,MAAMA,OAAAA,GAAS,gBAAA,CAAiB,OAAA,EAAS,SAAA,EAAW,GAAG,QAAQ,CAAA;AAC/D,IAAA,MAAMC,UAAAA,GAAY,gBAAA,CAAiB,QAAA,EAAU,UAAA,EAAY,GAAG,WAAW,CAAA;AACvE,IAAA,MAAM,cAAcD,OAAAA,GAAS,UAAA;AAC7B,IAAA,IAAI,cAAc,CAAA,EAAG;AACnB,MAAA,EAAA,IAAM,WAAA,GAAcC,UAAAA;AAAA,IACtB;AACA,IAAA,UAAA,GAAaD,OAAAA;AAAA,EACf;AAEA,EAAA,OAAO,EAAA;AACT;AAyBO,SAAS,OAAA,CAAQ,OAAe,KAAA,EAAuB;AAC5D,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAEpD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,MAAM,GAAA,GAAM,KAAA;AACZ,EAAA,IAAI,IAAA,GAAO,CAAA;AAEX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AACrC,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,IAAI,OAAA,GAAU,CAAA,IAAK,OAAA,GAAU,CAAA,EAAG;AAC9B,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,CAAA,wDAAA,EAA2D,MAAA,CAAO,OAAO,CAAC,aAAa,CAAC,CAAA,CAAA;AAAA,QACxF,OAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AACA,IAAA,MAAM,OAAA,GAAU,KAAK,GAAA,CAAI,GAAA,EAAK,KAAK,GAAA,CAAI,CAAA,GAAI,GAAA,EAAK,OAAO,CAAC,CAAA;AACxD,IAAA,IAAA,IAAQ,OAAA,GAAU,IAAA,CAAK,GAAA,CAAI,OAAO,CAAA,GAAA,CAAK,IAAI,OAAA,IAAW,IAAA,CAAK,GAAA,CAAI,CAAA,GAAI,OAAO,CAAA;AAAA,EAC5E;AAEA,EAAA,OAAO,OAAO,KAAA,CAAM,IAAA;AACtB;AAkCO,SAAS,WAAA,CAAY,OAAe,KAAA,EAAuB;AAChE,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAEpD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,0BAAA,CAA2B,OAAO,KAAK,CAAA;AACvC,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,MAAA,GAAS,CAAA;AACb,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AACjE,IAAA,MAAM,OAAA,GAAU,mBAAA,CAAoB,KAAA,EAAO,UAAA,EAAY,GAAG,OAAO,CAAA;AACjE,IAAA,IAAI,YAAY,OAAA,EAAS;AACvB,MAAA,MAAA,EAAA;AAAA,IACF;AAAA,EACF;AAEA,EAAA,OAAO,SAAS,KAAA,CAAM,IAAA;AACxB;AAuCO,SAAS,YAAA,CAAa,OAAe,KAAA,EAAuB;AACjE,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAEpD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,EAAA,GAAK,CAAA,EACP,EAAA,GAAK,CAAA,EACL,EAAA,GAAK,CAAA;AAEP,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AACrC,IAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AAErC,IAAA,IAAI,OAAA,KAAY,CAAA,IAAK,OAAA,KAAY,CAAA,EAAG,EAAA,EAAA;AAAA,SAAA,IAC3B,OAAA,KAAY,CAAA,IAAK,OAAA,KAAY,CAAA,EAAG,EAAA,EAAA;AAAA,SAAA,IAChC,OAAA,KAAY,CAAA,IAAK,OAAA,KAAY,CAAA,EAAG,EAAA,EAAA;AAAA,EAC3C;AAEA,EAAA,OAAO,KAAK,EAAA,GAAK,EAAA,KAAO,IAAI,CAAA,GAAI,EAAA,IAAM,KAAK,EAAA,GAAK,EAAA,CAAA;AAClD;AAmCO,SAAS,gBAAA,CAAiB,OAAe,KAAA,EAAuB;AACrE,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAEpD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,KAAK,CAAA,EACP,EAAA,GAAK,CAAA,EACL,EAAA,GAAK,GACL,EAAA,GAAK,CAAA;AAEP,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,MAAM,OAAA,GAAU,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAClE,IAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AACrC,IAAA,iBAAA,CAAkB,OAAA,EAAS,SAAS,CAAC,CAAA;AAErC,IAAA,IAAI,OAAA,KAAY,CAAA,IAAK,OAAA,KAAY,CAAA,EAAG,EAAA,EAAA;AAAA,SAAA,IAC3B,OAAA,KAAY,CAAA,IAAK,OAAA,KAAY,CAAA,EAAG,EAAA,EAAA;AAAA,SAAA,IAChC,OAAA,KAAY,CAAA,IAAK,OAAA,KAAY,CAAA,EAAG,EAAA,EAAA;AAAA,SAAA,IAChC,OAAA,KAAY,CAAA,IAAK,OAAA,KAAY,CAAA,EAAG,EAAA,EAAA;AAAA,EAC3C;AAEA,EAAA,MAAM,SAAA,GAAY,EAAA,GAAK,EAAA,GAAK,EAAA,GAAK,EAAA;AACjC,EAAA,MAAM,WAAA,GAAc,IAAA,CAAK,IAAA,CAAA,CAAM,EAAA,GAAK,EAAA,KAAO,KAAK,EAAA,CAAA,IAAO,EAAA,GAAK,EAAA,CAAA,IAAO,EAAA,GAAK,EAAA,CAAG,CAAA;AAE3E,EAAA,OAAO,WAAA,KAAgB,CAAA,GAAI,CAAA,GAAI,SAAA,GAAY,WAAA;AAC7C;AAqCO,SAAS,eAAA,CAAgB,OAAe,KAAA,EAAuB;AACpE,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAEpD,EAAA,MAAM,IAAI,KAAA,CAAM,IAAA;AAEhB,EAAA,IAAI,CAAA,KAAM,GAAG,OAAO,CAAA;AAEpB,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,SAAA,GAAY,oBAAoB,KAAK,CAAA;AAC3C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAG5C,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,MAAM,SAAA,uBAAgB,GAAA,EAAoB;AAC1C,EAAA,MAAM,SAAA,uBAAgB,GAAA,EAAoB;AAE1C,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,CAAA,GAAI,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC5D,IAAA,MAAM,CAAA,GAAI,gBAAA,CAAiB,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC5D,IAAA,IAAI,MAAM,CAAA,EAAG;AACX,MAAA,EAAA,EAAA;AAAA,IACF;AACA,IAAA,SAAA,CAAU,IAAI,CAAA,EAAA,CAAI,SAAA,CAAU,IAAI,CAAC,CAAA,IAAK,KAAK,CAAC,CAAA;AAC5C,IAAA,SAAA,CAAU,IAAI,CAAA,EAAA,CAAI,SAAA,CAAU,IAAI,CAAC,CAAA,IAAK,KAAK,CAAC,CAAA;AAAA,EAC9C;AAEA,EAAA,EAAA,IAAM,CAAA;AAEN,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,MAAM,UAAA,mBAAa,IAAI,GAAA,CAAI,CAAC,GAAG,SAAA,CAAU,IAAA,EAAK,EAAG,GAAG,SAAA,CAAU,IAAA,EAAM,CAAC,CAAA;AACrE,EAAA,KAAA,MAAW,KAAK,UAAA,EAAY;AAC1B,IAAA,MAAM,QAAA,GAAA,CAAY,SAAA,CAAU,GAAA,CAAI,CAAC,KAAK,CAAA,IAAK,CAAA;AAC3C,IAAA,MAAM,QAAA,GAAA,CAAY,SAAA,CAAU,GAAA,CAAI,CAAC,KAAK,CAAA,IAAK,CAAA;AAC3C,IAAA,EAAA,IAAM,QAAA,GAAW,QAAA;AAAA,EACnB;AAEA,EAAA,MAAM,QAAQ,CAAA,GAAI,EAAA;AAClB,EAAA,IAAI,KAAA,KAAU,CAAA,EAAG,OAAO,EAAA,KAAO,IAAI,CAAA,GAAI,CAAA;AACvC,EAAA,OAAA,CAAQ,KAAK,EAAA,IAAM,KAAA;AACrB;AAkCO,SAAS,qBAAA,CAAsB,OAAe,KAAA,EAAuB;AAC1E,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AAEpD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,EAAE,OAAA,EAAS,KAAA,EAAM,GAAI,eAAA,CAAgB,OAAO,KAAK,CAAA;AACvD,EAAA,IAAI,SAAA,GAAY,CAAA;AAChB,EAAA,IAAI,UAAA,GAAa,CAAA;AAEjB,EAAA,KAAA,MAAW,OAAO,OAAA,EAAS;AACzB,IAAA,MAAM,UAAA,GAAa,KAAA,CAAM,GAAA,CAAI,GAAG,CAAA;AAChC,IAAA,MAAM,OAAA,GAAU,YAAY,OAAA,IAAW,CAAA;AACvC,IAAA,IAAI,YAAY,CAAA,EAAG;AACnB,IAAA,MAAM,EAAA,GAAK,YAAY,EAAA,IAAM,CAAA;AAC7B,IAAA,MAAM,EAAA,GAAK,YAAY,EAAA,IAAM,CAAA;AAC7B,IAAA,MAAMA,UAAS,EAAA,GAAK,EAAA,GAAK,CAAA,GAAI,EAAA,IAAM,KAAK,EAAA,CAAA,GAAM,CAAA;AAC9C,IAAA,SAAA,IAAaA,OAAAA;AACb,IAAA,UAAA,EAAA;AAAA,EACF;AAEA,EAAA,OAAO,UAAA,KAAe,CAAA,GAAI,CAAA,GAAI,SAAA,GAAY,UAAA;AAC5C;;;AC1hDA,SAAS,mBAAmB,CAAA,EAAiD;AAC3E,EAAA,OAAO,CAAA,YAAa,iBAAiB,CAAA,YAAa,cAAA;AACpD;AAEA,SAAS,SAAA,CAAa,GAAA,EAAmB,KAAA,EAAe,IAAA,EAAiB;AACvE,EAAA,MAAM,CAAA,GAAI,IAAI,KAAK,CAAA;AACnB,EAAA,IAAI,MAAM,MAAA,EAAW;AACnB,IAAA,MAAM,IAAI,mBAAA,CAAoB,CAAA,EAAG,IAAI,CAAA,qBAAA,EAAwB,KAAK,CAAA,CAAE,CAAA;AAAA,EACtE;AACA,EAAA,OAAO,CAAA;AACT;AAEA,SAAS,oBAAA,CAAqB,GAAW,IAAA,EAAiC;AACxE,EAAA,IAAI,CAAA,CAAE,UAAU,QAAA,EAAU;AACxB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,+CAAA,CAAiD,CAAA;AAAA,EAC/E;AACA,EAAA,IAAI,CAAA,CAAE,UAAU,OAAA,EAAS;AACvB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,8CAAA,CAAgD,CAAA;AAAA,EAC9E;AACA,EAAA,MAAM,OAAO,CAAA,CAAE,IAAA;AACf,EAAA,IAAI,CAAC,YAAA,CAAa,IAAI,KAAK,CAAC,mBAAA,CAAoB,IAAI,CAAA,EAAG;AACrD,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,yBAAA,CAA2B,CAAA;AAAA,EACzD;AACA,EAAA,OAAO,IAAA;AACT;AAEA,SAAS,mBAAmB,CAAA,EAAW;AACrC,EAAA,MAAM,IAAA,GAAO,oBAAA,CAAqB,CAAA,EAAG,GAAG,CAAA;AAExC,EAAA,IAAI,CAAA,CAAE,IAAA,KAAS,CAAA,IAAK,CAAA,CAAE,OAAO,CAAA,EAAG;AAC9B,IAAA,MAAM,IAAI,WAAW,6BAA6B,CAAA;AAAA,EACpD;AAEA,EAAA,MAAM,QAAA,GAAW,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC/B,EAAA,MAAM,SAAA,GAAY,EAAE,IAAA,KAAS,CAAA,GAAI,IAAK,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AACpD,EAAA,IAAI,cAAc,CAAA,EAAG;AACnB,IAAA,MAAM,IAAI,WAAW,kCAAkC,CAAA;AAAA,EACzD;AAEA,EAAA,MAAM,cAAA,GAAiB,qBAAA,CAAsB,CAAA,CAAE,KAAK,CAAA;AACpD,EAAA,MAAM,QAAA,GAAW,eAAe,CAAC,CAAA;AACjC,EAAA,IAAI,QAAA,KAAa,MAAA,EAAW,MAAM,IAAI,WAAW,yBAAyB,CAAA;AAE1E,EAAA,MAAM,YAAA,GAAe,CAAA,CAAE,OAAA,CAAQ,CAAC,KAAK,QAAA,IAAY,SAAA;AAEjD,EAAA,MAAM,aAAA,GAAgB,CAAA,CAAE,IAAA,KAAS,CAAA,GAAI,CAAA,GAAK,CAAA,CAAE,OAAA,CAAQ,CAAC,CAAA,IAAK,cAAA,CAAe,CAAC,CAAA,IAAK,CAAA;AAE/E,EAAA,IAAI,CAAA,CAAE,IAAA,KAAS,CAAA,IAAK,aAAA,KAAkB,CAAA,EAAG;AACvC,IAAA,MAAM,IAAI,WAAW,6CAA6C,CAAA;AAAA,EACpE;AAEA,EAAA,OAAO;AAAA,IACL,IAAA;AAAA,IACA,QAAA;AAAA,IACA,SAAA;AAAA,IACA,YAAA;AAAA,IACA,aAAA;AAAA,IACA,QAAQ,CAAA,CAAE;AAAA,GACZ;AACF;AAOA,SAAS,cAAA,CACP,MAAA,EACA,SAAA,EACA,KAAA,EACA,IAAA,EACoC;AACpC,EAAA,MAAM,OAAO,MAAA,CAAO,IAAA;AACpB,EAAA,MAAM,IAAA,GAAO,UAAU,KAAK,CAAA;AAE5B,EAAA,IAAI,kBAAA,CAAmB,IAAI,CAAA,EAAG;AAC5B,IAAA,MAAM,CAAA,GAAI,KAAK,IAAI,CAAA;AACnB,IAAA,IAAI,MAAM,MAAA,EAAW;AACnB,MAAA,MAAM,IAAI,mBAAA,CAAoB,CAAA,EAAG,IAAI,CAAA,gCAAA,EAAmC,KAAK,CAAA,CAAE,CAAA;AAAA,IACjF;AACA,IAAA,OAAO,CAAA;AAAA,EACT;AAEA,EAAA,IAAI,YAAA,CAAa,IAAI,CAAA,IAAK,mBAAA,CAAoB,IAAI,CAAA,EAAG;AACnD,IAAA,MAAM,CAAA,GAAI,iBAAA,CAAkB,IAAA,EAAM,IAAI,CAAA;AACtC,IAAA,kBAAA,CAAmB,CAAA,EAAG,IAAA,EAAM,CAAA,MAAA,EAAS,KAAK,CAAA,CAAE,CAAA;AAE5C,IAAA,IAAA,CAAK,MAAA,CAAO,KAAA,KAAU,SAAA,IAAa,MAAA,CAAO,KAAA,KAAU,cAAc,CAAC,MAAA,CAAO,SAAA,CAAU,CAAC,CAAA,EAAG;AACtF,MAAA,MAAM,IAAI,mBAAA;AAAA,QACR,GAAG,IAAI,CAAA,iDAAA,EAAoD,OAAO,CAAC,CAAC,aAAa,KAAK,CAAA;AAAA,OACxF;AAAA,IACF;AAEA,IAAA,IAAI,MAAA,CAAO,KAAA,KAAU,OAAA,KAAY,CAAC,MAAA,CAAO,SAAA,CAAU,CAAC,CAAA,IAAK,CAAC,MAAA,CAAO,aAAA,CAAc,CAAC,CAAA,CAAA,EAAI;AAClF,MAAA,MAAM,IAAI,mBAAA;AAAA,QACR,CAAA,EAAG,IAAI,CAAA,gFAAA,EAAmF,KAAK,CAAA;AAAA,OACjG;AAAA,IACF;AAEA,IAAA,OAAO,CAAA;AAAA,EACT;AAEA,EAAA,IAAI,KAAA,CAAM,OAAA,CAAQ,IAAI,CAAA,EAAG;AACvB,IAAA,MAAM,CAAA,GAAI,KAAK,IAAI,CAAA;AACnB,IAAA,IAAI,CAAA,KAAM,MAAA,IAAa,CAAA,KAAM,IAAA,EAAM;AACjC,MAAA,MAAM,IAAI,mBAAA,CAAoB,CAAA,EAAG,IAAI,CAAA,gCAAA,EAAmC,KAAK,CAAA,CAAE,CAAA;AAAA,IACjF;AAEA,IAAA,IAAI,OAAO,CAAA,KAAM,QAAA,EAAU,OAAO,CAAA;AAClC,IAAA,IAAI,OAAO,CAAA,KAAM,SAAA,EAAW,OAAO,CAAA;AAEnC,IAAA,IAAI,OAAO,MAAM,QAAA,EAAU;AACzB,MAAA,kBAAA,CAAmB,CAAA,EAAG,IAAA,EAAM,CAAA,MAAA,EAAS,KAAK,CAAA,CAAE,CAAA;AAE5C,MAAA,IAAA,CAAK,MAAA,CAAO,KAAA,KAAU,SAAA,IAAa,MAAA,CAAO,KAAA,KAAU,cAAc,CAAC,MAAA,CAAO,SAAA,CAAU,CAAC,CAAA,EAAG;AACtF,QAAA,MAAM,IAAI,mBAAA;AAAA,UACR,GAAG,IAAI,CAAA,iDAAA,EAAoD,OAAO,CAAC,CAAC,aAAa,KAAK,CAAA;AAAA,SACxF;AAAA,MACF;AAEA,MAAA,IAAI,MAAA,CAAO,KAAA,KAAU,OAAA,KAAY,CAAC,MAAA,CAAO,SAAA,CAAU,CAAC,CAAA,IAAK,CAAC,MAAA,CAAO,aAAA,CAAc,CAAC,CAAA,CAAA,EAAI;AAClF,QAAA,MAAM,IAAI,mBAAA;AAAA,UACR,CAAA,EAAG,IAAI,CAAA,gFAAA,EAAmF,KAAK,CAAA;AAAA,SACjG;AAAA,MACF;AAEA,MAAA,OAAO,CAAA;AAAA,IACT;AAEA,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,gEAAA,CAAkE,CAAA;AAAA,EAChG;AAEA,EAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,2CAAA,CAA6C,CAAA;AAC3E;AAEA,SAAS,YAAA,CAAa,QAAgB,IAAA,EAA6B;AACjE,EAAA,IAAI,MAAA,CAAO,UAAU,QAAA,EAAU;AAC7B,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,oDAAA,CAAsD,CAAA;AAAA,EACpF;AAEA,EAAA,MAAM,IAAI,MAAA,CAAO,IAAA;AACjB,EAAA,MAAM,KAAA,GAAQ,IAAI,UAAA,CAAW,CAAC,CAAA;AAC9B,EAAA,MAAM,SAAA,GAAY,oBAAoB,MAAM,CAAA;AAE5C,EAAA,MAAM,GAAA,uBAAU,GAAA,EAAgD;AAChE,EAAA,IAAI,IAAA,GAAO,CAAA;AAEX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,GAAA,GAAM,cAAA,CAAe,MAAA,EAAQ,SAAA,EAAW,GAAG,IAAI,CAAA;AAErD,IAAA,MAAM,QAAA,GAAW,GAAA,CAAI,GAAA,CAAI,GAAG,CAAA;AAC5B,IAAA,IAAI,aAAa,MAAA,EAAW;AAC1B,MAAA,GAAA,CAAI,GAAA,CAAI,KAAK,IAAI,CAAA;AACjB,MAAA,KAAA,CAAM,CAAC,CAAA,GAAI,IAAA;AACX,MAAA,IAAA,EAAA;AAAA,IACF,CAAA,MAAO;AACL,MAAA,KAAA,CAAM,CAAC,CAAA,GAAI,QAAA;AAAA,IACb;AAAA,EACF;AAEA,EAAA,OAAO,EAAE,KAAA,EAAO,SAAA,EAAW,IAAA,EAAK;AAClC;AAEA,SAAS,MAAM,CAAA,EAAmB;AAChC,EAAA,OAAO,CAAA,IAAK,CAAA,GAAI,CAAA,GAAK,CAAA,IAAK,IAAI,CAAA,CAAA,GAAM,CAAA;AACtC;AAYA,SAAS,qBAAA,CAAsB,YAAoB,UAAA,EAAsC;AACvF,EAAA,cAAA,CAAe,UAAA,EAAY,UAAA,EAAY,YAAA,EAAc,YAAY,CAAA;AAEjE,EAAA,MAAM,IAAI,UAAA,CAAW,IAAA;AACrB,EAAA,MAAM,IAAA,GAAO,YAAA,CAAa,UAAA,EAAY,YAAY,CAAA;AAClD,EAAA,MAAM,IAAA,GAAO,YAAA,CAAa,UAAA,EAAY,YAAY,CAAA;AAElD,EAAA,MAAM,YAAY,IAAA,CAAK,KAAA;AACvB,EAAA,MAAM,YAAY,IAAA,CAAK,KAAA;AACvB,EAAA,MAAM,QAAQ,IAAA,CAAK,SAAA;AACnB,EAAA,MAAM,QAAQ,IAAA,CAAK,SAAA;AAEnB,EAAA,MAAM,SAAA,GAAY,IAAI,UAAA,CAAW,KAAK,CAAA;AACtC,EAAA,MAAM,SAAA,GAAY,IAAI,UAAA,CAAW,KAAK,CAAA;AAEtC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC7C,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC7C,IAAA,SAAA,CAAU,CAAC,CAAA,GAAA,CAAK,SAAA,CAAU,CAAC,KAAK,CAAA,IAAK,CAAA;AACrC,IAAA,SAAA,CAAU,CAAC,CAAA,GAAA,CAAK,SAAA,CAAU,CAAC,KAAK,CAAA,IAAK,CAAA;AAAA,EACvC;AAEA,EAAA,MAAM,YAAY,KAAA,GAAQ,KAAA;AAC1B,EAAA,MAAM,aAAA,GAAgB,GAAA;AAEtB,EAAA,IAAI,SAAA,GAAY,CAAA,IAAK,SAAA,IAAa,aAAA,EAAe;AAC/C,IAAA,MAAME,YAAAA,GAAc,IAAI,UAAA,CAAW,SAAS,CAAA;AAC5C,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,MAAM,CAAA,GAAI,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC7C,MAAA,MAAM,CAAA,GAAI,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC7C,MAAA,MAAM,GAAA,GAAM,IAAI,KAAA,GAAQ,CAAA;AACxB,MAAAA,aAAY,GAAG,CAAA,GAAA,CAAKA,YAAAA,CAAY,GAAG,KAAK,CAAA,IAAK,CAAA;AAAA,IAC/C;AACA,IAAA,OAAO;AAAA,MACL,gBAAA,EAAkBA,YAAAA;AAAA,MAClB,iBAAA,EAAmB,IAAA;AAAA,MACnB,SAAA;AAAA,MACA,SAAA;AAAA,MACA,KAAA;AAAA,MACA,KAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,WAAA,uBAAkB,GAAA,EAAoB;AAC5C,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC7C,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC7C,IAAA,MAAM,GAAA,GAAM,IAAI,KAAA,GAAQ,CAAA;AACxB,IAAA,WAAA,CAAY,IAAI,GAAA,EAAA,CAAM,WAAA,CAAY,IAAI,GAAG,CAAA,IAAK,KAAK,CAAC,CAAA;AAAA,EACtD;AAEA,EAAA,OAAO;AAAA,IACL,gBAAA,EAAkB,IAAA;AAAA,IAClB,iBAAA,EAAmB,WAAA;AAAA,IACnB,SAAA;AAAA,IACA,SAAA;AAAA,IACA,KAAA;AAAA,IACA,KAAA;AAAA,IACA;AAAA,GACF;AACF;AAEA,SAAS,qBAAA,CAAsB,QAAoB,CAAA,EAAmB;AACpE,EAAA,IAAI,CAAA,KAAM,GAAG,OAAO,CAAA;AACpB,EAAA,IAAI,CAAA,GAAI,CAAA;AAER,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,MAAA,CAAO,QAAQ,CAAA,EAAA,EAAK;AACtC,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,MAAA,EAAQ,CAAA,EAAG,QAAQ,CAAA;AACvC,IAAA,IAAI,IAAI,CAAA,EAAG;AACT,MAAA,MAAM,IAAI,CAAA,GAAI,CAAA;AACd,MAAA,CAAA,IAAK,CAAA,GAAI,IAAA,CAAK,GAAA,CAAI,CAAC,CAAA;AAAA,IACrB;AAAA,EACF;AACA,EAAA,OAAO,CAAA;AACT;AAEA,SAAS,iCAAiC,KAAA,EAAiC;AACzE,EAAA,MAAM,EAAE,gBAAA,EAAkB,iBAAA,EAAmB,WAAW,SAAA,EAAW,KAAA,EAAO,GAAE,GAAI,KAAA;AAChF,EAAA,IAAI,CAAA,KAAM,GAAG,OAAO,CAAA;AAEpB,EAAA,IAAI,EAAA,GAAK,CAAA;AAET,EAAA,IAAI,gBAAA,EAAkB;AACpB,IAAA,KAAA,IAAS,GAAA,GAAM,CAAA,EAAG,GAAA,GAAM,gBAAA,CAAiB,QAAQ,GAAA,EAAA,EAAO;AACtD,MAAA,MAAM,GAAA,GAAM,SAAA,CAAU,gBAAA,EAAkB,GAAA,EAAK,kBAAkB,CAAA;AAC/D,MAAA,IAAI,OAAO,CAAA,EAAG;AAEd,MAAA,MAAM,CAAA,GAAI,IAAA,CAAK,KAAA,CAAM,GAAA,GAAM,KAAK,CAAA;AAChC,MAAA,MAAM,CAAA,GAAI,MAAM,CAAA,GAAI,KAAA;AAEpB,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,IAAI,EAAA,GAAK,CAAA,IAAK,EAAA,GAAK,CAAA,EAAG;AACpB,QAAA,EAAA,IAAO,MAAM,CAAA,GAAK,IAAA,CAAK,IAAK,CAAA,GAAI,GAAA,IAAQ,KAAK,EAAA,CAAG,CAAA;AAAA,MAClD;AAAA,IACF;AACA,IAAA,OAAO,EAAA;AAAA,EACT;AAEA,EAAA,IAAI,iBAAA,EAAmB;AACrB,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,GAAG,CAAA,IAAK,iBAAA,EAAmB;AAC1C,MAAA,IAAI,OAAO,CAAA,EAAG;AAEd,MAAA,MAAM,CAAA,GAAI,IAAA,CAAK,KAAA,CAAM,GAAA,GAAM,KAAK,CAAA;AAChC,MAAA,MAAM,CAAA,GAAI,MAAM,CAAA,GAAI,KAAA;AAEpB,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,IAAI,EAAA,GAAK,CAAA,IAAK,EAAA,GAAK,CAAA,EAAG;AACpB,QAAA,EAAA,IAAO,MAAM,CAAA,GAAK,IAAA,CAAK,IAAK,CAAA,GAAI,GAAA,IAAQ,KAAK,EAAA,CAAG,CAAA;AAAA,MAClD;AAAA,IACF;AACA,IAAA,OAAO,EAAA;AAAA,EACT;AAEA,EAAA,OAAO,CAAA;AACT;AAEA,SAAS,mBAAmB,CAAA,EAAyB;AACnD,EAAA,MAAM,GAAA,GAAM,IAAI,YAAA,CAAa,CAAA,GAAI,CAAC,CAAA;AAClC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,IAAK,CAAA,EAAG,CAAA,EAAA,EAAK;AAC3B,IAAA,GAAA,CAAI,CAAC,KAAK,GAAA,CAAI,CAAA,GAAI,CAAC,CAAA,IAAK,CAAA,IAAK,IAAA,CAAK,GAAA,CAAI,CAAC,CAAA;AAAA,EACzC;AACA,EAAA,OAAO,GAAA;AACT;AAEA,SAAS,cAAA,CAAe,CAAA,EAAW,CAAA,EAAW,aAAA,EAAqC;AACjF,EAAA,IAAI,CAAA,GAAI,CAAA,IAAK,CAAA,GAAI,CAAA,SAAU,MAAA,CAAO,iBAAA;AAClC,EAAA,MAAM,CAAA,GAAI,SAAA,CAAU,aAAA,EAAe,CAAA,EAAG,eAAe,CAAA;AACrD,EAAA,MAAM,CAAA,GAAI,SAAA,CAAU,aAAA,EAAe,CAAA,EAAG,eAAe,CAAA;AACrD,EAAA,MAAM,CAAA,GAAI,SAAA,CAAU,aAAA,EAAe,CAAA,GAAI,GAAG,eAAe,CAAA;AACzD,EAAA,OAAO,IAAI,CAAA,GAAI,CAAA;AACjB;AAEA,IAAM,wBAAA,GAA2B,IAAA;AAEjC,SAAS,0BAA0B,KAAA,EAAiC;AAClE,EAAA,MAAM,EAAE,SAAA,EAAW,SAAA,EAAW,CAAA,EAAE,GAAI,KAAA;AACpC,EAAA,IAAI,CAAA,IAAK,GAAG,OAAO,CAAA;AAEnB,EAAA,MAAM,OAAA,GAAU,KAAA,CAAM,IAAA,CAAK,SAAS,CAAA;AACpC,EAAA,MAAM,OAAA,GAAU,KAAA,CAAM,IAAA,CAAK,SAAS,CAAA;AACpC,EAAA,MAAM,aAAA,GAAgB,mBAAmB,CAAC,CAAA;AAE1C,EAAA,IAAI,GAAA,GAAM,CAAA;AACV,EAAA,IAAI,IAAA,GAAO,CAAA;AAEX,EAAA,KAAA,MAAW,KAAK,OAAA,EAAS;AACvB,IAAA,IAAI,KAAK,CAAA,EAAG;AACZ,IAAA,KAAA,MAAW,KAAK,OAAA,EAAS;AACvB,MAAA,IAAI,KAAK,CAAA,EAAG;AAEZ,MAAA,MAAM,SAAS,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,CAAA,GAAI,IAAI,CAAC,CAAA;AACpC,MAAA,MAAM,MAAA,GAAS,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,CAAC,CAAA;AAC5B,MAAA,IAAI,SAAS,MAAA,EAAQ;AAErB,MAAA,MAAM,cAAA,GAAiB,cAAA,CAAe,CAAA,EAAG,CAAA,EAAG,aAAa,CAAA;AAEzD,MAAA,KAAA,IAAS,GAAA,GAAM,MAAA,EAAQ,GAAA,IAAO,MAAA,EAAQ,GAAA,EAAA,EAAO;AAC3C,QAAA,MAAM,cAAA,GACJ,cAAA,CAAe,CAAA,EAAG,GAAA,EAAK,aAAa,CAAA,GACpC,cAAA,CAAe,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,GAAA,EAAK,aAAa,CAAA,GAC5C,cAAA;AAEF,QAAA,IAAI,iBAAiB,wBAAA,EAA0B;AAE/C,QAAA,MAAM,WAAA,GAAc,IAAA,CAAK,GAAA,CAAI,cAAc,CAAA;AAC3C,QAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,WAAW,CAAA,IAAK,gBAAgB,CAAA,EAAG;AAExD,QAAA,MAAM,MAAA,GAAU,MAAM,CAAA,GAAK,IAAA,CAAK,IAAK,CAAA,GAAI,GAAA,IAAQ,IAAI,CAAA,CAAE,CAAA;AACvD,QAAA,MAAM,CAAA,GAAI,cAAc,MAAA,GAAS,IAAA;AACjC,QAAA,MAAM,IAAI,GAAA,GAAM,CAAA;AAChB,QAAA,IAAA,GAAO,IAAI,GAAA,GAAM,CAAA;AACjB,QAAA,GAAA,GAAM,CAAA;AAAA,MACR;AAAA,IACF;AAAA,EACF;AAEA,EAAA,OAAO,GAAA;AACT;AAIA,SAAS,cAAA,CAAe,KAAA,EAAe,KAAA,EAAe,MAAA,EAA+B;AACnF,EAAA,IAAI,WAAW,KAAA,EAAO,OAAO,IAAA,CAAK,GAAA,CAAI,OAAO,KAAK,CAAA;AAClD,EAAA,IAAI,WAAW,KAAA,EAAO,OAAO,IAAA,CAAK,GAAA,CAAI,OAAO,KAAK,CAAA;AAClD,EAAA,IAAI,WAAW,WAAA,EAAa,OAAO,IAAA,CAAK,IAAA,CAAK,QAAQ,KAAK,CAAA;AAC1D,EAAA,OAAA,CAAQ,QAAQ,KAAA,IAAS,CAAA;AAC3B;AAEA,SAAS,kBACP,IAAA,EACA,MAAA,EACA,cACA,aAAA,EACA,SAAA,EACA,GACA,CAAA,EACQ;AACR,EAAA,IAAI,GAAA,GAAM,CAAA;AACV,EAAA,MAAM,KAAA,GAAQ,SAAS,CAAA,GAAI,YAAA;AAC3B,EAAA,MAAM,KAAA,GAAQ,SAAS,CAAA,GAAI,YAAA;AAE3B,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,IAAA,MAAM,EAAA,GAAK,iBAAA,CAAkB,IAAA,EAAM,KAAA,GAAQ,IAAI,aAAa,CAAA;AAC5D,IAAA,MAAM,EAAA,GAAK,iBAAA,CAAkB,IAAA,EAAM,KAAA,GAAQ,IAAI,aAAa,CAAA;AAC5D,IAAA,kBAAA,CAAmB,IAAI,GAAA,EAAK,CAAA,OAAA,EAAU,CAAC,CAAA,UAAA,EAAa,CAAC,CAAA,CAAE,CAAA;AACvD,IAAA,kBAAA,CAAmB,IAAI,GAAA,EAAK,CAAA,OAAA,EAAU,CAAC,CAAA,UAAA,EAAa,CAAC,CAAA,CAAE,CAAA;AACvD,IAAA,MAAM,IAAI,EAAA,GAAK,EAAA;AACf,IAAA,GAAA,IAAO,CAAA,GAAI,CAAA;AAAA,EACb;AACA,EAAA,OAAO,IAAA,CAAK,KAAK,GAAG,CAAA;AACtB;AAIA,SAAS,8BAAA,CAA+B,GAAW,CAAA,EAAW;AAC5D,EAAA,MAAM,IAAA,GAAO,oBAAA,CAAqB,CAAA,EAAG,GAAG,CAAA;AAExC,EAAA,IAAI,CAAA,CAAE,SAAS,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,WAAW,gDAAgD,CAAA;AAAA,EACvE;AACA,EAAA,MAAM,IAAA,GAAO,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC3B,EAAA,MAAM,IAAA,GAAO,CAAA,CAAE,KAAA,CAAM,CAAC,CAAA,IAAK,CAAA;AAC3B,EAAA,IAAI,IAAA,KAAS,CAAA,IAAK,IAAA,KAAS,CAAA,EAAG;AAC5B,IAAA,MAAM,IAAI,UAAA;AAAA,MACR;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,cAAA,GAAiB,qBAAA,CAAsB,CAAA,CAAE,KAAK,CAAA;AACpD,EAAA,MAAM,YAAY,CAAA,CAAE,OAAA,CAAQ,CAAC,CAAA,IAAK,cAAA,CAAe,CAAC,CAAA,IAAK,CAAA;AACvD,EAAA,MAAM,YAAY,CAAA,CAAE,OAAA,CAAQ,CAAC,CAAA,IAAK,cAAA,CAAe,CAAC,CAAA,IAAK,CAAA;AAEvD,EAAA,IAAI,SAAA,KAAc,CAAA,IAAK,SAAA,KAAc,CAAA,EAAG;AACtC,IAAA,MAAM,IAAI,WAAW,8DAA8D,CAAA;AAAA,EACrF;AAEA,EAAA,MAAM,OAAO,CAAA,CAAE,MAAA;AAEf,EAAA,OAAO,CAAC,GAAW,CAAA,KAAc;AAC/B,IAAA,MAAM,IAAI,iBAAA,CAAkB,IAAA,EAAM,OAAO,CAAA,GAAI,SAAA,GAAY,IAAI,SAAS,CAAA;AACtE,IAAA,kBAAA,CAAmB,GAAG,GAAA,EAAK,CAAA,SAAA,EAAY,CAAC,CAAA,CAAA,EAAI,CAAC,CAAA,CAAA,CAAG,CAAA;AAChD,IAAA,IAAI,IAAI,CAAA,EAAG;AACT,MAAA,MAAM,IAAI,mBAAA;AAAA,QACR,qDAAqD,MAAA,CAAO,CAAC,CAAC,CAAA,KAAA,EAAQ,CAAC,IAAI,CAAC,CAAA,CAAA;AAAA,OAC9E;AAAA,IACF;AACA,IAAA,OAAO,CAAA;AAAA,EACT,CAAA;AACF;AAEA,SAAS,wBAAA,CAAyB,QAAgB,QAAA,EAAiC;AACjF,EAAA,IAAI,MAAA,CAAO,SAAS,QAAA,EAAU;AAC5B,IAAA,MAAM,IAAI,WAAW,4CAA4C,CAAA;AAAA,EACnE;AAEA,EAAA,MAAM,GAAA,GAAM,YAAA,CAAa,MAAA,EAAQ,QAAQ,CAAA;AACzC,EAAA,MAAM,IAAI,GAAA,CAAI,SAAA;AAEd,EAAA,IAAI,CAAA,GAAI,CAAA,IAAK,CAAA,GAAI,QAAA,GAAW,CAAA,EAAG;AAC7B,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,sDAAA;AAAA,MACA,YAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,OAAO,GAAA;AACT;AAEA,SAAS,sBAAA,CAAuB,CAAA,EAAW,CAAA,EAAW,IAAA,EAAsC;AAC1F,EAAA,IAAI,KAAA,GAAA,CAAS,QAAQ,CAAA,MAAO,CAAA;AAC5B,EAAA,MAAM,UAAU,IAAA,KAAS,MAAA;AAEzB,EAAA,MAAM,UAAU,MAAM;AACpB,IAAA,IAAI,CAAC,OAAA,EAAS,OAAQ,IAAA,CAAK,MAAA,KAAW,UAAA,KAAmB,CAAA;AACzD,IAAA,KAAA,GAAS,OAAA,GAAU,QAAQ,UAAA,KAAgB,CAAA;AAC3C,IAAA,OAAO,KAAA;AAAA,EACT,CAAA;AAEA,EAAA,MAAM,OAAA,GAAU,CAAC,YAAA,KAAyB,OAAA,EAAQ,GAAI,YAAA;AAEtD,EAAA,MAAM,GAAA,GAAM,IAAI,UAAA,CAAW,CAAC,CAAA;AAC5B,EAAA,KAAA,IAAS,IAAI,CAAA,EAAG,CAAA,GAAI,GAAG,CAAA,EAAA,EAAK,GAAA,CAAI,CAAC,CAAA,GAAI,CAAA;AAErC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,CAAA,GAAI,OAAA,CAAQ,CAAA,GAAI,CAAC,CAAA;AACvB,IAAA,IAAI,CAAA,GAAI,CAAA,EAAG,GAAA,CAAI,CAAC,CAAA,GAAI,CAAA;AAAA,EACtB;AACA,EAAA,OAAO,GAAA;AACT;AAEA,SAAS,eAAe,KAAA,EAGtB;AACA,EAAA,MAAM,GAAA,GAAM,IAAI,UAAA,CAAW,KAAA,CAAM,MAAM,CAAA;AACvC,EAAA,MAAM,GAAA,uBAAU,GAAA,EAAoB;AACpC,EAAA,IAAI,IAAA,GAAO,CAAA;AAEX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,QAAQ,CAAA,EAAA,EAAK;AACrC,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACrC,IAAA,MAAM,QAAA,GAAW,GAAA,CAAI,GAAA,CAAI,CAAC,CAAA;AAC1B,IAAA,IAAI,aAAa,MAAA,EAAW;AAC1B,MAAA,GAAA,CAAI,GAAA,CAAI,GAAG,IAAI,CAAA;AACf,MAAA,GAAA,CAAI,CAAC,CAAA,GAAI,IAAA;AACT,MAAA,IAAA,EAAA;AAAA,IACF,CAAA,MAAO;AACL,MAAA,GAAA,CAAI,CAAC,CAAA,GAAI,QAAA;AAAA,IACX;AAAA,EACF;AAEA,EAAA,OAAO,EAAE,KAAA,EAAO,GAAA,EAAK,SAAA,EAAW,IAAA,EAAK;AACvC;AAEA,SAAS,uBAAA,CAAwB,CAAA,EAAW,MAAA,EAAgB,OAAA,EAAoC;AAC9F,EAAA,MAAM,EAAE,MAAM,QAAA,EAAU,SAAA,EAAW,cAAc,aAAA,EAAe,MAAA,EAAO,GAAI,kBAAA,CAAmB,CAAC,CAAA;AAE/F,EAAA,MAAM,CAAA,GAAI,OAAA,GAAU,OAAA,CAAQ,MAAA,GAAS,QAAA;AACrC,EAAA,IAAI,QAAA,GAAW,CAAA,IAAK,CAAA,GAAI,CAAA,EAAG;AACzB,IAAA,MAAM,IAAI,qBAAA,CAAsB,wCAAA,EAA0C,WAAA,EAAa,CAAC,CAAA;AAAA,EAC1F;AAEA,EAAA,MAAM,MAAA,GAAS,wBAAA,CAAyB,MAAA,EAAQ,QAAQ,CAAA;AAExD,EAAA,MAAM,cAAA,GAAiB,IAAI,UAAA,CAAW,CAAC,CAAA;AACvC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,MAAM,OAAA,GAAU,SAAA,CAAU,OAAA,EAAS,CAAA,EAAG,SAAS,CAAA,GAAI,CAAA;AACzD,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,MAAA,CAAO,KAAA,EAAO,KAAK,cAAc,CAAA;AACrD,IAAA,cAAA,CAAe,CAAC,CAAA,GAAI,CAAA;AAAA,EACtB;AAEA,EAAA,MAAM,EAAA,GAAK,eAAe,cAAc,CAAA;AACxC,EAAA,MAAM,QAAQ,EAAA,CAAG,KAAA;AACjB,EAAA,MAAM,IAAI,EAAA,CAAG,SAAA;AAEb,EAAA,IAAI,CAAA,GAAI,CAAA,IAAK,CAAA,GAAI,CAAA,GAAI,CAAA,EAAG;AACtB,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,sDAAA;AAAA,MACA,YAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,YAAA,GAAe,IAAI,UAAA,CAAW,CAAC,CAAA;AACrC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACtC,IAAA,YAAA,CAAa,EAAE,CAAA,GAAA,CAAK,YAAA,CAAa,EAAE,KAAK,CAAA,IAAK,CAAA;AAAA,EAC/C;AAEA,EAAA,MAAM,cAAA,GAAiB,IAAI,YAAA,CAAa,CAAC,CAAA;AAEzC,EAAA,IAAI,GAAA,GAAM,CAAA;AACV,EAAA,IAAI,IAAA,GAAO,CAAA;AAEX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACtC,IAAA,MAAM,OAAA,GAAU,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AAE1D,IAAA,IAAI,WAAW,CAAA,EAAG;AAElB,IAAA,cAAA,CAAe,KAAK,CAAC,CAAA;AAErB,IAAA,MAAM,OAAO,OAAA,GAAU,SAAA,CAAU,OAAA,EAAS,CAAA,EAAG,SAAS,CAAA,GAAI,CAAA;AAE1D,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,IAAI,MAAM,CAAA,EAAG;AACb,MAAA,MAAM,OAAO,OAAA,GAAU,SAAA,CAAU,OAAA,EAAS,CAAA,EAAG,SAAS,CAAA,GAAI,CAAA;AAC1D,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AAEtC,MAAA,MAAM,CAAA,GAAI,kBAAkB,IAAA,EAAM,MAAA,EAAQ,cAAc,aAAA,EAAe,SAAA,EAAW,MAAM,IAAI,CAAA;AAC5F,MAAA,cAAA,CAAe,EAAE,CAAA,GAAA,CAAK,cAAA,CAAe,EAAE,KAAK,CAAA,IAAK,CAAA;AAAA,IACnD;AAEA,IAAA,MAAM,IAAI,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,KAAK,OAAA,GAAU,CAAA,CAAA;AAEvE,IAAA,IAAI,CAAA,GAAI,QAAA;AACR,IAAA,KAAA,IAAS,EAAA,GAAK,CAAA,EAAG,EAAA,GAAK,CAAA,EAAG,EAAA,EAAA,EAAM;AAC7B,MAAA,IAAI,OAAO,EAAA,EAAI;AACf,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AACrD,MAAA,IAAI,MAAM,CAAA,EAAG;AACb,MAAA,MAAM,IAAA,GAAO,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,CAAA,GAAI,EAAA;AAC/D,MAAA,IAAI,IAAA,GAAO,GAAG,CAAA,GAAI,IAAA;AAAA,IACpB;AAEA,IAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,CAAC,CAAA,IAAK,MAAM,QAAA,EAAU;AAE3C,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,CAAC,CAAA;AAC3B,IAAA,MAAM,CAAA,GAAI,KAAA,GAAQ,CAAA,GAAA,CAAK,CAAA,GAAI,KAAK,KAAA,GAAQ,CAAA;AAExC,IAAA,MAAM,IAAI,CAAA,GAAI,IAAA;AACd,IAAA,MAAM,IAAI,GAAA,GAAM,CAAA;AAChB,IAAA,IAAA,GAAO,IAAI,GAAA,GAAM,CAAA;AACjB,IAAA,GAAA,GAAM,CAAA;AAAA,EACR;AAEA,EAAA,OAAO,GAAA,GAAM,CAAA;AACf;AAEA,SAAS,yBAAA,CAA0B,CAAA,EAAW,MAAA,EAAgB,OAAA,EAAoC;AAChG,EAAA,MAAM,WAAW,MAAA,CAAO,IAAA;AACxB,EAAA,IAAI,WAAW,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,wCAAA;AAAA,MACA,WAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,MAAA,GAAS,wBAAA,CAAyB,MAAA,EAAQ,QAAQ,CAAA;AACxD,EAAA,MAAM,IAAA,GAAO,8BAAA,CAA+B,CAAA,EAAG,QAAQ,CAAA;AAEvD,EAAA,MAAM,CAAA,GAAI,OAAA,GAAU,OAAA,CAAQ,MAAA,GAAS,QAAA;AACrC,EAAA,IAAI,IAAI,CAAA,EAAG;AACT,IAAA,MAAM,IAAI,qBAAA,CAAsB,wCAAA,EAA0C,WAAA,EAAa,CAAC,CAAA;AAAA,EAC1F;AAGA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,MAAM,OAAA,GAAU,SAAA,CAAU,OAAA,EAAS,CAAA,EAAG,SAAS,CAAA,GAAI,CAAA;AACzD,IAAA,MAAM,EAAA,GAAK,IAAA,CAAK,GAAA,EAAK,GAAG,CAAA;AACxB,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,EAAE,KAAK,IAAA,CAAK,GAAA,CAAI,EAAE,CAAA,GAAI,KAAA,EAAO;AAChD,MAAA,MAAM,IAAI,mBAAA;AAAA,QACR,0DAA0D,MAAA,CAAO,EAAE,CAAC,CAAA,KAAA,EAAQ,GAAG,IAAI,GAAG,CAAA,CAAA;AAAA,OACxF;AAAA,IACF;AAAA,EACF;AAEA,EAAA,MAAM,cAAA,GAAiB,IAAI,UAAA,CAAW,CAAC,CAAA;AACvC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,MAAM,OAAA,GAAU,SAAA,CAAU,OAAA,EAAS,CAAA,EAAG,SAAS,CAAA,GAAI,CAAA;AACzD,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,MAAA,CAAO,KAAA,EAAO,KAAK,cAAc,CAAA;AACrD,IAAA,cAAA,CAAe,CAAC,CAAA,GAAI,CAAA;AAAA,EACtB;AAEA,EAAA,MAAM,EAAA,GAAK,eAAe,cAAc,CAAA;AACxC,EAAA,MAAM,QAAQ,EAAA,CAAG,KAAA;AACjB,EAAA,MAAM,IAAI,EAAA,CAAG,SAAA;AAEb,EAAA,IAAI,CAAA,GAAI,CAAA,IAAK,CAAA,GAAI,CAAA,GAAI,CAAA,EAAG;AACtB,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,sDAAA;AAAA,MACA,YAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,YAAA,GAAe,IAAI,UAAA,CAAW,CAAC,CAAA;AACrC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACtC,IAAA,YAAA,CAAa,EAAE,CAAA,GAAA,CAAK,YAAA,CAAa,EAAE,KAAK,CAAA,IAAK,CAAA;AAAA,EAC/C;AAEA,EAAA,MAAM,cAAA,GAAiB,IAAI,YAAA,CAAa,CAAC,CAAA;AAEzC,EAAA,IAAI,GAAA,GAAM,CAAA;AACV,EAAA,IAAI,IAAA,GAAO,CAAA;AAEX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACtC,IAAA,MAAM,OAAA,GAAU,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AAC1D,IAAA,IAAI,WAAW,CAAA,EAAG;AAElB,IAAA,cAAA,CAAe,KAAK,CAAC,CAAA;AAErB,IAAA,MAAM,OAAO,OAAA,GAAU,SAAA,CAAU,OAAA,EAAS,CAAA,EAAG,SAAS,CAAA,GAAI,CAAA;AAE1D,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,IAAI,MAAM,CAAA,EAAG;AAEb,MAAA,MAAM,OAAO,OAAA,GAAU,SAAA,CAAU,OAAA,EAAS,CAAA,EAAG,SAAS,CAAA,GAAI,CAAA;AAC1D,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AAEtC,MAAA,MAAM,CAAA,GAAI,IAAA,CAAK,IAAA,EAAM,IAAI,CAAA;AACzB,MAAA,cAAA,CAAe,EAAE,CAAA,GAAA,CAAK,cAAA,CAAe,EAAE,KAAK,CAAA,IAAK,CAAA;AAAA,IACnD;AAEA,IAAA,MAAM,IAAI,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,KAAK,OAAA,GAAU,CAAA,CAAA;AAEvE,IAAA,IAAI,CAAA,GAAI,QAAA;AACR,IAAA,KAAA,IAAS,EAAA,GAAK,CAAA,EAAG,EAAA,GAAK,CAAA,EAAG,EAAA,EAAA,EAAM;AAC7B,MAAA,IAAI,OAAO,EAAA,EAAI;AACf,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AACrD,MAAA,IAAI,MAAM,CAAA,EAAG;AACb,MAAA,MAAM,IAAA,GAAO,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,CAAA,GAAI,EAAA;AAC/D,MAAA,IAAI,IAAA,GAAO,GAAG,CAAA,GAAI,IAAA;AAAA,IACpB;AAEA,IAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,CAAC,CAAA,IAAK,MAAM,QAAA,EAAU;AAE3C,IAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,CAAC,CAAA;AAC3B,IAAA,MAAM,CAAA,GAAI,KAAA,GAAQ,CAAA,GAAA,CAAK,CAAA,GAAI,KAAK,KAAA,GAAQ,CAAA;AAExC,IAAA,MAAM,IAAI,CAAA,GAAI,IAAA;AACd,IAAA,MAAM,IAAI,GAAA,GAAM,CAAA;AAChB,IAAA,IAAA,GAAO,IAAI,GAAA,GAAM,CAAA;AACjB,IAAA,GAAA,GAAM,CAAA;AAAA,EACR;AAEA,EAAA,OAAO,GAAA,GAAM,CAAA;AACf;AA+BO,SAAS,eAAA,CACd,CAAA,EACA,MAAA,EACA,MAAA,GAA2B,aAC3B,OAAA,EACQ;AACR,EAAA,MAAM,aAAa,OAAA,EAAS,UAAA;AAC5B,EAAA,MAAM,cAAc,OAAA,EAAS,WAAA;AAE7B,EAAA,MAAM,WAAW,MAAA,CAAO,IAAA;AACxB,EAAA,IAAI,WAAW,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,wCAAA;AAAA,MACA,WAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,OAAA,GAAU,GAAA;AAChB,EAAA,IAAI,UAAA,KAAe,MAAA,IAAa,QAAA,GAAW,OAAA,EAAS;AAClD,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,6CAA0C,QAAQ,CAAA,8DAAA,CAAA;AAAA,MAClD,YAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,IAAI,eAAe,MAAA,EAAW;AAC5B,IAAA,IAAI,CAAC,OAAO,QAAA,CAAS,UAAU,KAAK,CAAC,MAAA,CAAO,SAAA,CAAU,UAAU,CAAA,EAAG;AACjE,MAAA,MAAM,IAAI,qBAAA,CAAsB,+BAAA,EAAiC,YAAA,EAAc,UAAU,CAAA;AAAA,IAC3F;AACA,IAAA,IAAI,UAAA,GAAa,CAAA,IAAK,UAAA,GAAa,QAAA,EAAU;AAC3C,MAAA,MAAM,IAAI,qBAAA;AAAA,QACR,sDAAA;AAAA,QACA,YAAA;AAAA,QACA;AAAA,OACF;AAAA,IACF;AAAA,EACF;AAEA,EAAA,MAAM,OAAA,GACJ,eAAe,MAAA,IAAa,UAAA,GAAa,WACrC,sBAAA,CAAuB,QAAA,EAAU,UAAA,EAAY,WAAW,CAAA,GACxD,IAAA;AAEN,EAAA,IAAI,WAAW,WAAA,EAAa,OAAO,uBAAA,CAAwB,CAAA,EAAG,QAAQ,OAAO,CAAA;AAC7E,EAAA,IAAI,WAAW,aAAA,EAAe,OAAO,yBAAA,CAA0B,CAAA,EAAG,QAAQ,OAAO,CAAA;AAEjF,EAAA,MAAM,IAAI,qBAAA;AAAA,IACR,CAAA,qBAAA,EAAwB,MAAA,CAAO,MAAM,CAAC,CAAA,uCAAA,CAAA;AAAA,IACtC,QAAA;AAAA,IACA;AAAA,GACF;AACF;AAuBO,SAAS,iBAAA,CACd,CAAA,EACA,MAAA,EACA,MAAA,GAA2B,WAAA,EACnB;AACR,EAAA,MAAM,WAAW,MAAA,CAAO,IAAA;AACxB,EAAA,IAAI,WAAW,CAAA,EAAG;AAChB,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,wCAAA;AAAA,MACA,WAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,GAAA,GAAM,wBAAA,CAAyB,MAAA,EAAQ,QAAQ,CAAA;AACrD,EAAA,MAAM,QAAQ,GAAA,CAAI,KAAA;AAClB,EAAA,MAAM,IAAI,GAAA,CAAI,SAAA;AAEd,EAAA,MAAM,WAAA,GAAc,IAAI,YAAA,CAAa,QAAQ,CAAA;AAE7C,EAAA,MAAM,YAAA,GAAe,IAAI,UAAA,CAAW,CAAC,CAAA;AACrC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACtC,IAAA,YAAA,CAAa,EAAE,CAAA,GAAA,CAAK,YAAA,CAAa,EAAE,KAAK,CAAA,IAAK,CAAA;AAAA,EAC/C;AAEA,EAAA,MAAM,cAAA,GAAiB,IAAI,YAAA,CAAa,CAAC,CAAA;AAEzC,EAAA,IAAI,WAAW,WAAA,EAAa;AAC1B,IAAA,MAAM,EAAE,MAAM,SAAA,EAAW,YAAA,EAAc,eAAe,MAAA,EAAO,GAAI,mBAAmB,CAAC,CAAA;AAErF,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACtC,MAAA,MAAM,OAAA,GAAU,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AAC1D,MAAA,IAAI,WAAW,CAAA,EAAG;AAChB,QAAA,WAAA,CAAY,CAAC,CAAA,GAAI,CAAA;AACjB,QAAA;AAAA,MACF;AAEA,MAAA,cAAA,CAAe,KAAK,CAAC,CAAA;AAErB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,IAAI,MAAM,CAAA,EAAG;AACb,QAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AAEtC,QAAA,MAAM,CAAA,GAAI,kBAAkB,IAAA,EAAM,MAAA,EAAQ,cAAc,aAAA,EAAe,SAAA,EAAW,GAAG,CAAC,CAAA;AACtF,QAAA,cAAA,CAAe,EAAE,CAAA,GAAA,CAAK,cAAA,CAAe,EAAE,KAAK,CAAA,IAAK,CAAA;AAAA,MACnD;AAEA,MAAA,MAAM,IAAI,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,KAAK,OAAA,GAAU,CAAA,CAAA;AAEvE,MAAA,IAAI,CAAA,GAAI,QAAA;AACR,MAAA,KAAA,IAAS,EAAA,GAAK,CAAA,EAAG,EAAA,GAAK,CAAA,EAAG,EAAA,EAAA,EAAM;AAC7B,QAAA,IAAI,OAAO,EAAA,EAAI;AACf,QAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AACrD,QAAA,IAAI,MAAM,CAAA,EAAG;AACb,QAAA,MAAM,IAAA,GAAO,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,CAAA,GAAI,EAAA;AAC/D,QAAA,IAAI,IAAA,GAAO,GAAG,CAAA,GAAI,IAAA;AAAA,MACpB;AAEA,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,CAAC,CAAA,IAAK,MAAM,QAAA,EAAU;AACzC,QAAA,WAAA,CAAY,CAAC,CAAA,GAAI,CAAA;AACjB,QAAA;AAAA,MACF;AAEA,MAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,CAAC,CAAA;AAC3B,MAAA,WAAA,CAAY,CAAC,CAAA,GAAI,KAAA,GAAQ,CAAA,GAAA,CAAK,CAAA,GAAI,KAAK,KAAA,GAAQ,CAAA;AAAA,IACjD;AAEA,IAAA,OAAO,OAAO,WAAW,CAAA;AAAA,EAC3B;AAEA,EAAA,IAAI,WAAW,aAAA,EAAe;AAC5B,IAAA,MAAM,IAAA,GAAO,8BAAA,CAA+B,CAAA,EAAG,QAAQ,CAAA;AAEvD,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,EAAA,GAAK,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACpB,MAAA,IAAI,CAAC,OAAO,QAAA,CAAS,EAAE,KAAK,IAAA,CAAK,GAAA,CAAI,EAAE,CAAA,GAAI,KAAA,EAAO;AAChD,QAAA,MAAM,IAAI,mBAAA;AAAA,UACR,0DAA0D,MAAA,CAAO,EAAE,CAAC,CAAA,KAAA,EAAQ,CAAC,IAAI,CAAC,CAAA,CAAA;AAAA,SACpF;AAAA,MACF;AAAA,IACF;AAEA,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACtC,MAAA,MAAM,OAAA,GAAU,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AAC1D,MAAA,IAAI,WAAW,CAAA,EAAG;AAChB,QAAA,WAAA,CAAY,CAAC,CAAA,GAAI,CAAA;AACjB,QAAA;AAAA,MACF;AAEA,MAAA,cAAA,CAAe,KAAK,CAAC,CAAA;AAErB,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,QAAA,IAAI,MAAM,CAAA,EAAG;AACb,QAAA,MAAM,EAAA,GAAK,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AAEtC,QAAA,MAAM,CAAA,GAAI,IAAA,CAAK,CAAA,EAAG,CAAC,CAAA;AACnB,QAAA,cAAA,CAAe,EAAE,CAAA,GAAA,CAAK,cAAA,CAAe,EAAE,KAAK,CAAA,IAAK,CAAA;AAAA,MACnD;AAEA,MAAA,MAAM,IAAI,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,KAAK,OAAA,GAAU,CAAA,CAAA;AAEvE,MAAA,IAAI,CAAA,GAAI,QAAA;AACR,MAAA,KAAA,IAAS,EAAA,GAAK,CAAA,EAAG,EAAA,GAAK,CAAA,EAAG,EAAA,EAAA,EAAM;AAC7B,QAAA,IAAI,OAAO,EAAA,EAAI;AACf,QAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,EAAA,EAAI,cAAc,CAAA;AACrD,QAAA,IAAI,MAAM,CAAA,EAAG;AACb,QAAA,MAAM,IAAA,GAAO,SAAA,CAAU,cAAA,EAAgB,EAAA,EAAI,gBAAgB,CAAA,GAAI,EAAA;AAC/D,QAAA,IAAI,IAAA,GAAO,GAAG,CAAA,GAAI,IAAA;AAAA,MACpB;AAEA,MAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,CAAC,CAAA,IAAK,MAAM,QAAA,EAAU;AACzC,QAAA,WAAA,CAAY,CAAC,CAAA,GAAI,CAAA;AACjB,QAAA;AAAA,MACF;AAEA,MAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,GAAA,CAAI,CAAA,EAAG,CAAC,CAAA;AAC3B,MAAA,WAAA,CAAY,CAAC,CAAA,GAAI,KAAA,GAAQ,CAAA,GAAA,CAAK,CAAA,GAAI,KAAK,KAAA,GAAQ,CAAA;AAAA,IACjD;AAEA,IAAA,OAAO,OAAO,WAAW,CAAA;AAAA,EAC3B;AAEA,EAAA,MAAM,IAAI,qBAAA;AAAA,IACR,CAAA,qBAAA,EAAwB,MAAA,CAAO,MAAM,CAAC,CAAA,uCAAA,CAAA;AAAA,IACtC,QAAA;AAAA,IACA;AAAA,GACF;AACF;AAsBO,SAAS,kBAAA,CAAmB,GAAW,MAAA,EAAwB;AACpE,EAAA,MAAM,EAAE,MAAM,QAAA,EAAU,SAAA,EAAW,cAAc,aAAA,EAAe,MAAA,EAAO,GAAI,kBAAA,CAAmB,CAAC,CAAA;AAC/F,EAAA,IAAI,QAAA,KAAa,GAAG,OAAO,CAAA;AAE3B,EAAA,IAAI,MAAA,CAAO,SAAS,QAAA,EAAU;AAC5B,IAAA,MAAM,IAAI,WAAW,4CAA4C,CAAA;AAAA,EACnE;AAEA,EAAA,MAAM,GAAA,GAAM,YAAA,CAAa,MAAA,EAAQ,QAAQ,CAAA;AACzC,EAAA,MAAM,QAAQ,GAAA,CAAI,KAAA;AAClB,EAAA,MAAM,IAAI,GAAA,CAAI,SAAA;AACd,EAAA,IAAI,CAAA,GAAI,GAAG,OAAO,CAAA;AAElB,EAAA,MAAM,SAAA,GAAY,IAAI,KAAA,CAAoB,CAAC,CAAA;AAC3C,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,YAAe,CAAC,CAAA,GAAI,IAAI,YAAA,CAAa,SAAS,CAAA;AAErE,EAAA,MAAM,YAAA,GAAe,IAAI,UAAA,CAAW,CAAC,CAAA;AAErC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACrC,IAAA,YAAA,CAAa,CAAC,CAAA,GAAA,CAAK,YAAA,CAAa,CAAC,KAAK,CAAA,IAAK,CAAA;AAE3C,IAAA,MAAM,IAAA,GAAO,SAAS,CAAA,GAAI,YAAA;AAC1B,IAAA,MAAM,QAAA,GAAW,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AACpD,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,CAAA,GAAI,iBAAA,CAAkB,IAAA,EAAM,IAAA,GAAO,IAAI,aAAa,CAAA;AAC1D,MAAA,kBAAA,CAAmB,GAAG,GAAA,EAAK,CAAA,OAAA,EAAU,CAAC,CAAA,UAAA,EAAa,CAAC,CAAA,CAAE,CAAA;AACtD,MAAA,QAAA,CAAS,CAAC,CAAA,GAAA,CAAK,QAAA,CAAS,CAAC,KAAK,CAAA,IAAK,CAAA;AAAA,IACrC;AAAA,EACF;AAEA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,CAAA,EAAG,cAAc,CAAA;AACpD,IAAA,IAAI,MAAM,CAAA,EAAG;AACb,IAAA,MAAM,QAAA,GAAW,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AACpD,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,QAAA,CAAS,CAAC,CAAA,GAAA,CAAK,QAAA,CAAS,CAAC,KAAK,CAAA,IAAK,EAAA;AAAA,IACrC;AAAA,EACF;AAEA,EAAA,MAAM,UAAA,GAAa,IAAI,YAAA,CAAa,CAAC,CAAA;AAErC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACrC,IAAA,MAAM,QAAA,GAAW,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAEpD,IAAA,IAAI,MAAA,GAAS,CAAA;AACb,IAAA,MAAM,IAAA,GAAO,SAAS,CAAA,GAAI,YAAA;AAC1B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,CAAA,GAAI,iBAAA,CAAkB,IAAA,EAAM,IAAA,GAAO,IAAI,aAAa,CAAA;AAC1D,MAAA,kBAAA,CAAmB,GAAG,GAAA,EAAK,CAAA,OAAA,EAAU,CAAC,CAAA,UAAA,EAAa,CAAC,CAAA,CAAE,CAAA;AACtD,MAAA,MAAM,CAAA,GAAI,CAAA,IAAK,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA,CAAA;AAC9B,MAAA,MAAA,IAAU,CAAA,GAAI,CAAA;AAAA,IAChB;AAEA,IAAA,UAAA,CAAW,CAAC,KAAK,UAAA,CAAW,CAAC,KAAK,CAAA,IAAK,IAAA,CAAK,KAAK,MAAM,CAAA;AAAA,EACzD;AAEA,EAAA,MAAM,CAAA,GAAI,IAAI,YAAA,CAAa,CAAC,CAAA;AAC5B,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,CAAA,EAAG,cAAc,CAAA;AACpD,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,UAAA,EAAY,CAAA,EAAG,YAAY,CAAA;AAChD,IAAA,CAAA,CAAE,CAAC,CAAA,GAAI,EAAA,GAAK,CAAA,GAAI,KAAK,EAAA,GAAK,CAAA;AAAA,EAC5B;AAEA,EAAA,IAAI,EAAA,GAAK,CAAA;AAET,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,IAAI,WAAW,MAAA,CAAO,iBAAA;AACtB,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAE9C,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,MAAA,IAAI,MAAM,CAAA,EAAG;AACb,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAE9C,MAAA,IAAI,MAAA,GAAS,CAAA;AACb,MAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,QAAA,MAAM,KAAK,EAAA,CAAG,CAAC,KAAK,CAAA,KAAM,EAAA,CAAG,CAAC,CAAA,IAAK,CAAA,CAAA;AACnC,QAAA,MAAA,IAAU,CAAA,GAAI,CAAA;AAAA,MAChB;AAEA,MAAA,MAAM,IAAA,GAAO,IAAA,CAAK,IAAA,CAAK,MAAM,CAAA;AAC7B,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,CAAA,EAAG,CAAA,EAAG,GAAG,CAAA;AAC9B,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,CAAA,EAAG,CAAA,EAAG,GAAG,CAAA;AAC9B,MAAA,MAAM,QAAQ,IAAA,KAAS,CAAA,GAAI,MAAA,CAAO,iBAAA,GAAA,CAAqB,KAAK,EAAA,IAAM,IAAA;AAClE,MAAA,IAAI,KAAA,GAAQ,UAAU,QAAA,GAAW,KAAA;AAAA,IACnC;AAEA,IAAA,EAAA,IAAM,QAAA;AAAA,EACR;AAEA,EAAA,OAAO,EAAA,GAAK,CAAA;AACd;AAsBO,SAAS,qBAAA,CAAsB,GAAW,MAAA,EAAwB;AACvE,EAAA,MAAM,EAAE,MAAM,QAAA,EAAU,SAAA,EAAW,cAAc,aAAA,EAAe,MAAA,EAAO,GAAI,kBAAA,CAAmB,CAAC,CAAA;AAC/F,EAAA,IAAI,QAAA,KAAa,GAAG,OAAO,CAAA;AAE3B,EAAA,IAAI,MAAA,CAAO,SAAS,QAAA,EAAU;AAC5B,IAAA,MAAM,IAAI,WAAW,4CAA4C,CAAA;AAAA,EACnE;AAEA,EAAA,MAAM,GAAA,GAAM,YAAA,CAAa,MAAA,EAAQ,QAAQ,CAAA;AACzC,EAAA,MAAM,QAAQ,GAAA,CAAI,KAAA;AAClB,EAAA,MAAM,IAAI,GAAA,CAAI,SAAA;AAEd,EAAA,MAAM,WAAA,GAAc,IAAI,YAAA,CAAa,SAAS,CAAA;AAE9C,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,IAAA,MAAM,IAAA,GAAO,SAAS,CAAA,GAAI,YAAA;AAC1B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,CAAA,GAAI,iBAAA,CAAkB,IAAA,EAAM,IAAA,GAAO,IAAI,aAAa,CAAA;AAC1D,MAAA,kBAAA,CAAmB,GAAG,GAAA,EAAK,CAAA,OAAA,EAAU,CAAC,CAAA,UAAA,EAAa,CAAC,CAAA,CAAE,CAAA;AACtD,MAAA,WAAA,CAAY,CAAC,CAAA,GAAA,CAAK,WAAA,CAAY,CAAC,KAAK,CAAA,IAAK,CAAA;AAAA,IAC3C;AAAA,EACF;AACA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,IAAA,WAAA,CAAY,CAAC,CAAA,GAAA,CAAK,WAAA,CAAY,CAAC,KAAK,CAAA,IAAK,QAAA;AAAA,EAC3C;AAEA,EAAA,MAAM,SAAA,GAAY,IAAI,KAAA,CAAoB,CAAC,CAAA;AAC3C,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,YAAe,CAAC,CAAA,GAAI,IAAI,YAAA,CAAa,SAAS,CAAA;AACrE,EAAA,MAAM,YAAA,GAAe,IAAI,UAAA,CAAW,CAAC,CAAA;AAErC,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACrC,IAAA,YAAA,CAAa,CAAC,CAAA,GAAA,CAAK,YAAA,CAAa,CAAC,KAAK,CAAA,IAAK,CAAA;AAE3C,IAAA,MAAM,IAAA,GAAO,SAAS,CAAA,GAAI,YAAA;AAC1B,IAAA,MAAM,QAAA,GAAW,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AACpD,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,CAAA,GAAI,iBAAA,CAAkB,IAAA,EAAM,IAAA,GAAO,IAAI,aAAa,CAAA;AAC1D,MAAA,kBAAA,CAAmB,GAAG,GAAA,EAAK,CAAA,OAAA,EAAU,CAAC,CAAA,UAAA,EAAa,CAAC,CAAA,CAAE,CAAA;AACtD,MAAA,QAAA,CAAS,CAAC,CAAA,GAAA,CAAK,QAAA,CAAS,CAAC,KAAK,CAAA,IAAK,CAAA;AAAA,IACrC;AAAA,EACF;AAEA,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,CAAA,EAAG,cAAc,CAAA;AACpD,IAAA,IAAI,MAAM,CAAA,EAAG;AACb,IAAA,MAAM,QAAA,GAAW,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AACpD,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,QAAA,CAAS,CAAC,CAAA,GAAA,CAAK,QAAA,CAAS,CAAC,KAAK,CAAA,IAAK,EAAA;AAAA,IACrC;AAAA,EACF;AAEA,EAAA,IAAI,IAAA,GAAO,CAAA;AACX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,CAAA,EAAG,CAAA,EAAA,EAAK;AAC1B,IAAA,MAAM,EAAA,GAAK,SAAA,CAAU,YAAA,EAAc,CAAA,EAAG,cAAc,CAAA;AACpD,IAAA,IAAI,MAAM,CAAA,EAAG;AAEb,IAAA,MAAM,QAAA,GAAW,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AACpD,IAAA,IAAI,MAAA,GAAS,CAAA;AACb,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,KAAK,QAAA,CAAS,CAAC,KAAK,CAAA,KAAM,WAAA,CAAY,CAAC,CAAA,IAAK,CAAA,CAAA;AAClD,MAAA,MAAA,IAAU,CAAA,GAAI,CAAA;AAAA,IAChB;AACA,IAAA,IAAA,IAAQ,EAAA,GAAK,MAAA;AAAA,EACf;AAEA,EAAA,IAAI,IAAA,GAAO,CAAA;AACX,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,QAAA,EAAU,CAAA,EAAA,EAAK;AACjC,IAAA,MAAM,CAAA,GAAI,SAAA,CAAU,KAAA,EAAO,CAAA,EAAG,OAAO,CAAA;AACrC,IAAA,MAAM,QAAA,GAAW,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAEpD,IAAA,MAAM,IAAA,GAAO,SAAS,CAAA,GAAI,YAAA;AAC1B,IAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,EAAW,CAAA,EAAA,EAAK;AAClC,MAAA,MAAM,CAAA,GAAI,iBAAA,CAAkB,IAAA,EAAM,IAAA,GAAO,IAAI,aAAa,CAAA;AAC1D,MAAA,kBAAA,CAAmB,GAAG,GAAA,EAAK,CAAA,OAAA,EAAU,CAAC,CAAA,UAAA,EAAa,CAAC,CAAA,CAAE,CAAA;AACtD,MAAA,MAAM,CAAA,GAAI,CAAA,IAAK,QAAA,CAAS,CAAC,CAAA,IAAK,CAAA,CAAA;AAC9B,MAAA,IAAA,IAAQ,CAAA,GAAI,CAAA;AAAA,IACd;AAAA,EACF;AAEA,EAAA,IAAI,CAAA,GAAI,CAAA,IAAK,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAChC,EAAA,OAAO,IAAA,IAAQ,CAAA,GAAI,CAAA,CAAA,IAAM,IAAA,IAAQ,QAAA,GAAW,CAAA,CAAA,CAAA;AAC9C;AAsBO,SAAS,iBAAA,CAAkB,YAAoB,UAAA,EAA4B;AAChF,EAAA,cAAA,CAAe,UAAA,EAAY,UAAA,EAAY,YAAA,EAAc,YAAY,CAAA;AAEjE,EAAA,MAAM,IAAI,UAAA,CAAW,IAAA;AACrB,EAAA,IAAI,CAAA,IAAK,GAAG,OAAO,CAAA;AAEnB,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,UAAA,EAAY,UAAU,CAAA;AAC1D,EAAA,MAAM,EAAE,gBAAA,EAAkB,iBAAA,EAAmB,SAAA,EAAW,WAAU,GAAI,KAAA;AAEtE,EAAA,IAAI,OAAA,GAAU,CAAA;AAEd,EAAA,IAAI,gBAAA,EAAkB;AACpB,IAAA,KAAA,IAAS,GAAA,GAAM,CAAA,EAAG,GAAA,GAAM,gBAAA,CAAiB,QAAQ,GAAA,EAAA,EAAO;AACtD,MAAA,MAAM,GAAA,GAAM,SAAA,CAAU,gBAAA,EAAkB,GAAA,EAAK,kBAAkB,CAAA;AAC/D,MAAA,IAAI,GAAA,GAAM,CAAA,EAAG,OAAA,IAAW,KAAA,CAAM,GAAG,CAAA;AAAA,IACnC;AAAA,EACF,WAAW,iBAAA,EAAmB;AAC5B,IAAA,KAAA,MAAW,GAAA,IAAO,iBAAA,CAAkB,MAAA,EAAO,EAAG;AAC5C,MAAA,OAAA,IAAW,MAAM,GAAG,CAAA;AAAA,IACtB;AAAA,EACF;AAEA,EAAA,IAAI,WAAA,GAAc,CAAA;AAClB,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,CAAU,QAAQ,CAAA,EAAA,EAAK;AACzC,IAAA,WAAA,IAAe,KAAA,CAAM,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAC,CAAA;AAAA,EAC3D;AAEA,EAAA,IAAI,WAAA,GAAc,CAAA;AAClB,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,CAAU,QAAQ,CAAA,EAAA,EAAK;AACzC,IAAA,WAAA,IAAe,KAAA,CAAM,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAC,CAAA;AAAA,EAC3D;AAEA,EAAA,MAAM,UAAA,GAAa,MAAM,CAAC,CAAA;AAC1B,EAAA,IAAI,UAAA,KAAe,GAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,aAAA,GAAiB,cAAc,WAAA,GAAe,UAAA;AACpD,EAAA,MAAM,QAAA,GAAA,CAAY,cAAc,WAAA,IAAe,CAAA;AAE/C,EAAA,MAAM,QAAQ,QAAA,GAAW,aAAA;AACzB,EAAA,IAAI,KAAA,KAAU,GAAG,OAAO,CAAA;AAExB,EAAA,OAAA,CAAQ,UAAU,aAAA,IAAiB,KAAA;AACrC;AAsBO,SAAS,uBAAA,CACd,UAAA,EACA,UAAA,EACA,aAAA,GAA+B,YAAA,EACvB;AACR,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,UAAA,EAAY,UAAU,CAAA;AAC1D,EAAA,MAAM,EAAE,CAAA,EAAG,SAAA,EAAW,SAAA,EAAU,GAAI,KAAA;AACpC,EAAA,IAAI,CAAA,IAAK,GAAG,OAAO,CAAA;AAEnB,EAAA,MAAM,EAAA,GAAK,iCAAiC,KAAK,CAAA;AACjD,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,SAAA,EAAW,CAAC,CAAA;AAChD,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,SAAA,EAAW,CAAC,CAAA;AAChD,EAAA,MAAM,GAAA,GAAM,0BAA0B,KAAK,CAAA;AAE3C,EAAA,MAAM,UAAA,GAAa,cAAA,CAAe,KAAA,EAAO,KAAA,EAAO,aAAa,CAAA;AAC7D,EAAA,IAAI,IAAA,CAAK,GAAA,CAAI,UAAU,CAAA,GAAI,OAAO,OAAO,CAAA;AAEzC,EAAA,MAAM,QAAQ,UAAA,GAAa,GAAA;AAC3B,EAAA,IAAI,IAAA,CAAK,GAAA,CAAI,KAAK,CAAA,GAAI,OAAO,OAAO,CAAA;AAEpC,EAAA,MAAM,GAAA,GAAA,CAAO,KAAK,GAAA,IAAO,KAAA;AACzB,EAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,GAAG,GAAG,OAAO,CAAA;AAClC,EAAA,IAAI,GAAA,GAAM,GAAG,OAAO,CAAA;AACpB,EAAA,IAAI,GAAA,GAAM,IAAI,OAAO,EAAA;AACrB,EAAA,OAAO,GAAA;AACT;AAsBO,SAAS,yBAAA,CACd,UAAA,EACA,UAAA,EACA,aAAA,GAA+B,YAAA,EACvB;AACR,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,UAAA,EAAY,UAAU,CAAA;AAC1D,EAAA,MAAM,EAAE,CAAA,EAAG,SAAA,EAAW,SAAA,EAAU,GAAI,KAAA;AAEpC,EAAA,MAAM,EAAA,GAAK,iCAAiC,KAAK,CAAA;AACjD,EAAA,MAAM,EAAA,GAAK,qBAAA,CAAsB,SAAA,EAAW,CAAC,CAAA;AAC7C,EAAA,MAAM,EAAA,GAAK,qBAAA,CAAsB,SAAA,EAAW,CAAC,CAAA;AAE7C,EAAA,IAAI,EAAA,KAAO,CAAA,IAAK,EAAA,KAAO,CAAA,EAAG;AACxB,IAAA,OAAO,EAAA,KAAO,CAAA,IAAK,EAAA,KAAO,CAAA,GAAI,CAAA,GAAM,CAAA;AAAA,EACtC;AAEA,EAAA,MAAM,UAAA,GAAa,cAAA,CAAe,EAAA,EAAI,EAAA,EAAI,aAAa,CAAA;AACvD,EAAA,IAAI,UAAA,KAAe,GAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,MAAM,EAAA,GAAK,UAAA;AACjB,EAAA,IAAI,GAAA,GAAM,GAAG,OAAO,CAAA;AACpB,EAAA,IAAI,GAAA,GAAM,GAAG,OAAO,CAAA;AACpB,EAAA,OAAO,GAAA;AACT;AAqBO,SAAS,mBAAA,CAAoB,YAAoB,UAAA,EAA4B;AAClF,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,UAAA,EAAY,UAAU,CAAA;AAC1D,EAAA,MAAM,EAAE,gBAAA,EAAkB,iBAAA,EAAmB,SAAA,EAAW,SAAA,EAAW,GAAE,GAAI,KAAA;AAEzE,EAAA,IAAI,CAAA,KAAM,GAAG,OAAO,CAAA;AAEpB,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,IAAI,gBAAA,EAAkB;AACpB,IAAA,KAAA,IAAS,GAAA,GAAM,CAAA,EAAG,GAAA,GAAM,gBAAA,CAAiB,QAAQ,GAAA,EAAA,EAAO;AACtD,MAAA,MAAM,GAAA,GAAM,SAAA,CAAU,gBAAA,EAAkB,GAAA,EAAK,kBAAkB,CAAA;AAC/D,MAAA,IAAI,GAAA,GAAM,CAAA,EAAG,EAAA,IAAM,KAAA,CAAM,GAAG,CAAA;AAAA,IAC9B;AAAA,EACF,WAAW,iBAAA,EAAmB;AAC5B,IAAA,KAAA,MAAW,GAAA,IAAO,iBAAA,CAAkB,MAAA,EAAO,EAAG;AAC5C,MAAA,EAAA,IAAM,MAAM,GAAG,CAAA;AAAA,IACjB;AAAA,EACF;AAEA,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,CAAU,MAAA,EAAQ,CAAA,EAAA,EAAK,EAAA,IAAM,KAAA,CAAM,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAC,CAAA;AAE3F,EAAA,IAAI,EAAA,GAAK,CAAA;AACT,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,SAAA,CAAU,MAAA,EAAQ,CAAA,EAAA,EAAK,EAAA,IAAM,KAAA,CAAM,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAC,CAAA;AAE3F,EAAA,IAAI,EAAA,KAAO,CAAA,IAAK,EAAA,KAAO,CAAA,EAAG,OAAO,CAAA;AACjC,EAAA,OAAO,EAAA,GAAK,IAAA,CAAK,IAAA,CAAK,EAAA,GAAK,EAAE,CAAA;AAC/B;AAsBO,SAAS,gBAAA,CAAiB,YAAoB,UAAA,EAA4B;AAC/E,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,UAAA,EAAY,UAAU,CAAA;AAC1D,EAAA,MAAM,EAAE,gBAAA,EAAkB,iBAAA,EAAmB,WAAW,SAAA,EAAW,KAAA,EAAO,GAAE,GAAI,KAAA;AAEhF,EAAA,IAAI,CAAA,KAAM,GAAG,OAAO,CAAA;AAEpB,EAAA,IAAI,GAAA,GAAM,CAAA;AAEV,EAAA,IAAI,gBAAA,EAAkB;AACpB,IAAA,KAAA,IAAS,GAAA,GAAM,CAAA,EAAG,GAAA,GAAM,gBAAA,CAAiB,QAAQ,GAAA,EAAA,EAAO;AACtD,MAAA,MAAM,GAAA,GAAM,SAAA,CAAU,gBAAA,EAAkB,GAAA,EAAK,kBAAkB,CAAA;AAC/D,MAAA,IAAI,OAAO,CAAA,EAAG;AAEd,MAAA,MAAM,IAAI,GAAA,GAAM,IAAA,CAAK,KAAA,CAAM,GAAA,GAAM,KAAK,CAAA,GAAI,KAAA;AAC1C,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,IAAI,EAAA,GAAK,GAAG,GAAA,IAAQ,GAAA,GAAM,IAAK,IAAA,CAAK,GAAA,CAAI,MAAM,EAAE,CAAA;AAAA,IAClD;AAAA,EACF,WAAW,iBAAA,EAAmB;AAC5B,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,GAAG,CAAA,IAAK,iBAAA,EAAmB;AAC1C,MAAA,IAAI,OAAO,CAAA,EAAG;AAEd,MAAA,MAAM,IAAI,GAAA,GAAM,IAAA,CAAK,KAAA,CAAM,GAAA,GAAM,KAAK,CAAA,GAAI,KAAA;AAC1C,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,IAAI,EAAA,GAAK,GAAG,GAAA,IAAQ,GAAA,GAAM,IAAK,IAAA,CAAK,GAAA,CAAI,MAAM,EAAE,CAAA;AAAA,IAClD;AAAA,EACF;AAEA,EAAA,MAAM,EAAA,GAAK,qBAAA,CAAsB,SAAA,EAAW,CAAC,CAAA;AAC7C,EAAA,OAAO,EAAA,KAAO,CAAA,GAAI,CAAA,GAAM,CAAA,GAAM,GAAA,GAAM,EAAA;AACtC;AAsBO,SAAS,iBAAA,CAAkB,YAAoB,UAAA,EAA4B;AAChF,EAAA,MAAM,KAAA,GAAQ,qBAAA,CAAsB,UAAA,EAAY,UAAU,CAAA;AAC1D,EAAA,MAAM,EAAE,gBAAA,EAAkB,iBAAA,EAAmB,WAAW,SAAA,EAAW,KAAA,EAAO,GAAE,GAAI,KAAA;AAEhF,EAAA,IAAI,CAAA,KAAM,GAAG,OAAO,CAAA;AAEpB,EAAA,IAAI,GAAA,GAAM,CAAA;AAEV,EAAA,IAAI,gBAAA,EAAkB;AACpB,IAAA,KAAA,IAAS,GAAA,GAAM,CAAA,EAAG,GAAA,GAAM,gBAAA,CAAiB,QAAQ,GAAA,EAAA,EAAO;AACtD,MAAA,MAAM,GAAA,GAAM,SAAA,CAAU,gBAAA,EAAkB,GAAA,EAAK,kBAAkB,CAAA;AAC/D,MAAA,IAAI,OAAO,CAAA,EAAG;AAEd,MAAA,MAAM,CAAA,GAAI,IAAA,CAAK,KAAA,CAAM,GAAA,GAAM,KAAK,CAAA;AAChC,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,IAAI,EAAA,GAAK,GAAG,GAAA,IAAQ,GAAA,GAAM,IAAK,IAAA,CAAK,GAAA,CAAI,MAAM,EAAE,CAAA;AAAA,IAClD;AAAA,EACF,WAAW,iBAAA,EAAmB;AAC5B,IAAA,KAAA,MAAW,CAAC,GAAA,EAAK,GAAG,CAAA,IAAK,iBAAA,EAAmB;AAC1C,MAAA,IAAI,OAAO,CAAA,EAAG;AAEd,MAAA,MAAM,CAAA,GAAI,IAAA,CAAK,KAAA,CAAM,GAAA,GAAM,KAAK,CAAA;AAChC,MAAA,MAAM,EAAA,GAAK,SAAA,CAAU,SAAA,EAAW,CAAA,EAAG,WAAW,CAAA;AAC9C,MAAA,IAAI,EAAA,GAAK,GAAG,GAAA,IAAQ,GAAA,GAAM,IAAK,IAAA,CAAK,GAAA,CAAI,MAAM,EAAE,CAAA;AAAA,IAClD;AAAA,EACF;AAEA,EAAA,MAAM,EAAA,GAAK,qBAAA,CAAsB,SAAA,EAAW,CAAC,CAAA;AAC7C,EAAA,OAAO,EAAA,KAAO,CAAA,GAAI,CAAA,GAAM,CAAA,GAAM,GAAA,GAAM,EAAA;AACtC;AA0BO,SAAS,aAAA,CAAc,UAAA,EAAoB,UAAA,EAAoB,IAAA,GAAO,CAAA,EAAa;AACxF,EAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,IAAI,CAAA,IAAK,QAAQ,CAAA,EAAG;AACvC,IAAA,MAAM,IAAI,qBAAA,CAAsB,uCAAA,EAAyC,MAAA,EAAQ,IAAI,CAAA;AAAA,EACvF;AAEA,EAAA,MAAM,CAAA,GAAI,gBAAA,CAAiB,UAAA,EAAY,UAAU,CAAA;AACjD,EAAA,MAAM,CAAA,GAAI,iBAAA,CAAkB,UAAA,EAAY,UAAU,CAAA;AAElD,EAAA,IAAI,CAAA,GAAI,CAAA,KAAM,CAAA,EAAG,OAAO,CAAA;AACxB,EAAA,OAAA,CAAS,CAAA,GAAI,IAAA,IAAQ,CAAA,GAAI,CAAA,IAAM,OAAO,CAAA,GAAI,CAAA,CAAA;AAC5C;;;ACp8CA,SAAS,wBAAA,CAAyB,GAAW,IAAA,EAAiC;AAC5E,EAAA,IAAI,CAAA,CAAE,UAAU,QAAA,EAAU;AACxB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,wBAAA,CAA0B,CAAA;AAAA,EACxD;AACA,EAAA,IAAI,CAAA,CAAE,UAAU,OAAA,EAAS;AACvB,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,8CAAA,CAAgD,CAAA;AAAA,EAC9E;AAEA,EAAA,MAAM,OAAO,CAAA,CAAE,IAAA;AACf,EAAA,IAAI,CAAC,YAAA,CAAa,IAAI,KAAK,CAAC,mBAAA,CAAoB,IAAI,CAAA,EAAG;AACrD,IAAA,MAAM,IAAI,UAAA,CAAW,CAAA,EAAG,IAAI,CAAA,wBAAA,CAA0B,CAAA;AAAA,EACxD;AAEA,EAAA,OAAO,IAAA;AACT;AAEA,SAAS,WAAA,CACP,IAAA,EACA,SAAA,EACA,KAAA,EACA,IAAA,EACA;AACA,EAAA,MAAM,KAAA,GAAQ,iBAAA,CAAkB,IAAA,EAAM,SAAA,CAAU,KAAK,CAAC,CAAA;AACtD,EAAA,kBAAA,CAAmB,KAAA,EAAO,IAAA,EAAM,CAAA,MAAA,EAAS,KAAK,CAAA,CAAE,CAAA;AAChD,EAAA,OAAO,KAAA;AACT;AAgCO,SAAS,GAAA,CAAI,OAAe,KAAA,EAAuB;AACxD,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AAEzD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,eAAA,GAAkB,CAAA;AACtB,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,IAAA,GACJ,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO,CAAA,GAC7C,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO,CAAA;AAC/C,IAAA,eAAA,IAAmB,IAAA,GAAO,IAAA;AAAA,EAC5B;AAEA,EAAA,OAAO,kBAAkB,KAAA,CAAM,IAAA;AACjC;AAgCO,SAAS,IAAA,CAAK,OAAe,KAAA,EAAuB;AACzD,EAAA,OAAO,IAAA,CAAK,IAAA,CAAK,GAAA,CAAI,KAAA,EAAO,KAAK,CAAC,CAAA;AACpC;AAgCO,SAAS,GAAA,CAAI,OAAe,KAAA,EAAuB;AACxD,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AAEzD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,WAAA,GAAc,CAAA;AAClB,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,IAAA,GACJ,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO,CAAA,GAC7C,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO,CAAA;AAC/C,IAAA,WAAA,IAAe,IAAA,CAAK,IAAI,IAAI,CAAA;AAAA,EAC9B;AAEA,EAAA,OAAO,cAAc,KAAA,CAAM,IAAA;AAC7B;AAqCO,SAAS,OAAA,CAAQ,OAAe,KAAA,EAAuB;AAC5D,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,IAAI,KAAA,CAAM,SAAS,CAAA,EAAG;AACpB,IAAA,MAAM,IAAI,qBAAA,CAAsB,sCAAA,EAAwC,OAAA,EAAS,MAAM,IAAI,CAAA;AAAA,EAC7F;AAEA,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,OAAA,GAAU,CAAA;AACd,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,OAAA,IAAW,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO,CAAA;AAAA,EAC1D;AACA,EAAA,MAAM,IAAA,GAAO,UAAU,KAAA,CAAM,IAAA;AAE7B,EAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,EAAA,IAAI,KAAA,GAAQ,CAAA;AACZ,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,KAAA,IAAA,CAAU,UAAU,OAAA,KAAY,CAAA;AAChC,IAAA,KAAA,IAAA,CAAU,UAAU,IAAA,KAAS,CAAA;AAAA,EAC/B;AAIA,EAAA,IAAI,UAAU,CAAA,EAAG;AACf,IAAA,OAAO,KAAA,KAAU,IAAI,CAAA,GAAM,CAAA;AAAA,EAC7B;AAEA,EAAA,OAAO,IAAI,KAAA,GAAQ,KAAA;AACrB;AAuCO,SAAS,eAAA,CAAgB,KAAA,EAAe,KAAA,EAAe,SAAA,EAA2B;AACvF,EAAA,IAAI,CAAC,MAAA,CAAO,QAAA,CAAS,SAAS,CAAA,IAAK,CAAC,MAAA,CAAO,SAAA,CAAU,SAAS,CAAA,IAAK,SAAA,GAAY,CAAA,EAAG;AAChF,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,0CAAA;AAAA,MACA,WAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,IAAI,KAAA,CAAM,IAAA;AAChB,EAAA,MAAM,CAAA,GAAI,SAAA;AAGV,EAAA,IAAI,CAAA,IAAK,IAAI,CAAA,EAAG;AACd,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,CAAA,kEAAA,EAAkE,CAAC,CAAA,IAAA,EAAO,CAAC,CAAA,CAAA;AAAA,MAC3E,WAAA;AAAA,MACA;AAAA,KACF;AAAA,EACF;AAEA,EAAA,MAAM,EAAA,GAAK,OAAA,CAAQ,KAAA,EAAO,KAAK,CAAA;AAE/B,EAAA,OAAO,KAAM,CAAA,GAAI,EAAA,KAAO,CAAA,GAAI,CAAA,CAAA,IAAO,IAAI,CAAA,GAAI,CAAA,CAAA;AAC7C;AAqCO,SAAS,IAAA,CAAK,OAAe,KAAA,EAAuB;AACzD,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,eAAA,GAAkB,CAAA;AACtB,EAAA,IAAI,YAAA,GAAe,CAAA;AACnB,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,IAAI,YAAY,CAAA,EAAG;AACjB,MAAA,eAAA,IAAmB,IAAA,CAAK,GAAA,CAAA,CAAK,OAAA,GAAU,OAAA,IAAW,OAAO,CAAA;AACzD,MAAA,YAAA,EAAA;AAAA,IACF;AAAA,EACF;AAEA,EAAA,IAAI,iBAAiB,CAAA,EAAG;AACtB,IAAA,OAAO,CAAA;AAAA,EACT;AAEA,EAAA,OAAQ,kBAAkB,YAAA,GAAgB,GAAA;AAC5C;AAgCO,SAAS,mBAAA,CAAoB,OAAe,KAAA,EAAuB;AACxE,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AAEzD,EAAA,IAAI,KAAA,CAAM,IAAA,KAAS,CAAA,EAAG,OAAO,CAAA;AAE7B,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,MAAM,SAAmB,EAAC;AAC1B,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAO,IAAA,CAAK,GAAA;AAAA,MAChB,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO,IAC3C,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO;AAAA,KACjD;AACA,IAAA,MAAA,CAAO,KAAK,IAAI,CAAA;AAAA,EAClB;AAEA,EAAA,MAAA,CAAO,IAAA,CAAK,CAAC,CAAA,EAAG,CAAA,KAAM,IAAI,CAAC,CAAA;AAC3B,EAAA,MAAM,GAAA,GAAM,IAAA,CAAK,KAAA,CAAM,MAAA,CAAO,SAAS,CAAC,CAAA;AACxC,EAAA,OAAO,OAAO,MAAA,GAAS,CAAA,KAAM,CAAA,GACxB,MAAA,CAAO,GAAG,CAAA,IAAK,CAAA,GAAA,CAAA,CACd,MAAA,CAAO,GAAA,GAAM,CAAC,CAAA,IAAK,CAAA,KAAM,MAAA,CAAO,GAAG,KAAK,CAAA,CAAA,IAAM,CAAA;AACtD;AAgCO,SAAS,QAAA,CAAS,OAAe,KAAA,EAAuB;AAC7D,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AAEzD,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,MAAA,GAAS,CAAA;AACb,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAO,IAAA,CAAK,GAAA;AAAA,MAChB,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO,IAC3C,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,CAAA,EAAG,OAAO;AAAA,KACjD;AACA,IAAA,MAAA,GAAS,IAAA,CAAK,GAAA,CAAI,MAAA,EAAQ,IAAI,CAAA;AAAA,EAChC;AAEA,EAAA,OAAO,MAAA;AACT;AAkCO,SAAS,sBAAA,CAAuB,OAAe,KAAA,EAAuB;AAC3E,EAAA,qBAAA,CAAsB,KAAA,EAAO,KAAA,EAAO,OAAA,EAAS,OAAO,CAAA;AACpD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,MAAM,SAAA,GAAY,wBAAA,CAAyB,KAAA,EAAO,OAAO,CAAA;AACzD,EAAA,IAAI,KAAA,CAAM,SAAS,CAAA,EAAG;AACpB,IAAA,MAAM,IAAI,qBAAA;AAAA,MACR,qDAAA;AAAA,MACA,OAAA;AAAA,MACA,KAAA,CAAM;AAAA,KACR;AAAA,EACF;AAEA,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAC5C,EAAA,MAAM,UAAA,GAAa,oBAAoB,KAAK,CAAA;AAE5C,EAAA,IAAI,OAAA,GAAU,CAAA;AACd,EAAA,IAAI,WAAA,GAAc,CAAA;AAClB,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,OAAA,IAAW,OAAA;AACX,IAAA,WAAA,IAAe,OAAA,GAAU,OAAA;AAAA,EAC3B;AACA,EAAA,MAAM,QAAA,GAAW,UAAU,KAAA,CAAM,IAAA;AACjC,EAAA,MAAM,YAAA,GAAe,cAAc,KAAA,CAAM,IAAA;AAEzC,EAAA,IAAI,WAAA,GAAc,CAAA;AAClB,EAAA,IAAI,OAAA,GAAU,CAAA;AACd,EAAA,KAAA,IAAS,CAAA,GAAI,CAAA,EAAG,CAAA,GAAI,KAAA,CAAM,MAAM,CAAA,EAAA,EAAK;AACnC,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,MAAM,OAAA,GAAU,WAAA,CAAY,SAAA,EAAW,UAAA,EAAY,GAAG,OAAO,CAAA;AAC7D,IAAA,MAAM,WAAW,OAAA,GAAU,OAAA;AAC3B,IAAA,WAAA,IAAA,CAAgB,WAAW,YAAA,KAAiB,CAAA;AAC5C,IAAA,OAAA,IAAA,CAAY,UAAU,QAAA,KAAa,CAAA;AAAA,EACrC;AAIA,EAAA,IAAI,YAAY,CAAA,EAAG;AACjB,IAAA,OAAO,WAAA,KAAgB,IAAI,CAAA,GAAM,CAAA;AAAA,EACnC;AAEA,EAAA,OAAO,IAAI,WAAA,GAAc,OAAA;AAC3B","file":"chunk-FJYLIGJX.js","sourcesContent":["export {\n  accuracy,\n  averagePrecisionScore,\n  balancedAccuracyScore,\n  classificationReport,\n  cohenKappaScore,\n  confusionMatrix,\n  f1Score,\n  fbetaScore,\n  hammingLoss,\n  jaccardScore,\n  logLoss,\n  matthewsCorrcoef,\n  precision,\n  precisionRecallCurve,\n  recall,\n  rocAucScore,\n  rocCurve,\n} from \"./classification\";\nexport {\n  adjustedMutualInfoScore,\n  adjustedRandScore,\n  calinskiHarabaszScore,\n  completenessScore,\n  daviesBouldinScore,\n  fowlkesMallowsScore,\n  homogeneityScore,\n  normalizedMutualInfoScore,\n  silhouetteSamples,\n  silhouetteScore,\n  vMeasureScore,\n} from \"./clustering\";\nexport {\n  adjustedR2Score,\n  explainedVarianceScore,\n  mae,\n  mape,\n  maxError,\n  medianAbsoluteError,\n  mse,\n  r2Score,\n  rmse,\n} from \"./regression\";\n","/**\n * Shared internal helpers for the metrics module.\n *\n * These functions are used across classification, regression, and clustering metrics\n * to handle strided tensor access and input validation.\n *\n * @internal\n */\n\nimport { DataValidationError, ShapeError } from \"../core/errors\";\nimport type { Tensor } from \"../ndarray\";\n\n/**\n * Converts a flat (logical) index to a physical buffer offset.\n *\n * @internal\n */\nexport type FlatOffsetter = (flatIndex: number) => number;\n\n/**\n * Compute row-major logical strides from a tensor shape.\n *\n * @internal\n */\nexport function computeLogicalStrides(shape: readonly number[]): number[] {\n  const strides = new Array<number>(shape.length);\n  let stride = 1;\n  for (let i = shape.length - 1; i >= 0; i--) {\n    const dim = shape[i];\n    if (dim === undefined) {\n      throw new ShapeError(\"Tensor shape must be fully defined\");\n    }\n    strides[i] = stride;\n    stride *= dim;\n  }\n  return strides;\n}\n\n/**\n * Build a function that maps flat (logical) indices to physical buffer offsets.\n *\n * Handles arbitrary strides (views, slices, transposes).\n *\n * @internal\n */\nexport function createFlatOffsetter(t: Tensor): FlatOffsetter {\n  const base = t.offset;\n\n  if (t.ndim <= 1) {\n    const stride0 = t.strides[0] ?? 1;\n    return (flatIndex: number) => base + flatIndex * stride0;\n  }\n\n  const logicalStrides = computeLogicalStrides(t.shape);\n  const strides = t.strides;\n\n  return (flatIndex: number) => {\n    let rem = flatIndex;\n    let offset = base;\n\n    for (let axis = 0; axis < logicalStrides.length; axis++) {\n      const axisLogicalStride = logicalStrides[axis] ?? 1;\n      const coord = Math.floor(rem / axisLogicalStride);\n      rem -= coord * axisLogicalStride;\n      offset += coord * (strides[axis] ?? 0);\n    }\n\n    return offset;\n  };\n}\n\n/**\n * Assert that a numeric value is finite.\n *\n * @param value - The value to check\n * @param name - Name of the tensor for error messages\n * @param detail - Additional detail for the error message (e.g., \"index 5\")\n *\n * @internal\n */\nexport function assertFiniteNumber(value: number, name: string, detail: string): void {\n  if (!Number.isFinite(value)) {\n    throw new DataValidationError(\n      `${name} must contain only finite numbers; found ${String(value)} at ${detail}`\n    );\n  }\n}\n\n/**\n * Assert that a tensor is 1D or a column vector.\n *\n * @internal\n */\nexport function assertVectorLike(t: Tensor, name: string): void {\n  if (t.ndim <= 1) return;\n  if (t.ndim === 2 && (t.shape[1] ?? 0) === 1) return;\n  throw new ShapeError(`${name} must be 1D or a column vector`);\n}\n\n/**\n * Assert that two tensors have the same size and are vector-like.\n *\n * @internal\n */\nexport function assertSameSizeVectors(a: Tensor, b: Tensor, nameA: string, nameB: string): void {\n  if (a.size !== b.size) {\n    throw new ShapeError(\n      `${nameA} (size ${a.size}) and ${nameB} (size ${b.size}) must have same size`\n    );\n  }\n  assertVectorLike(a, nameA);\n  assertVectorLike(b, nameB);\n}\n\n/**\n * Assert that two tensors have the same size (without vector-like check).\n *\n * @internal\n */\nexport function assertSameSize(a: Tensor, b: Tensor, nameA: string, nameB: string): void {\n  if (a.size !== b.size) {\n    throw new ShapeError(\n      `${nameA} (size ${a.size}) and ${nameB} (size ${b.size}) must have same size`\n    );\n  }\n}\n","import {\n  getBigIntElement,\n  getNumericElement,\n  getStringElement,\n  isNumericTypedArray,\n  isTypedArray,\n  type NumericTypedArray,\n} from \"../core\";\nimport { DTypeError, InvalidParameterError } from \"../core/errors\";\nimport { type Tensor, tensor } from \"../ndarray\";\nimport {\n  assertFiniteNumber,\n  assertSameSizeVectors,\n  createFlatOffsetter,\n  type FlatOffsetter,\n} from \"./_internal\";\n\nfunction getNumericLabelData(t: Tensor): NumericTypedArray {\n  if (t.dtype === \"string\") {\n    throw new DTypeError(\"metrics do not support string labels\");\n  }\n  if (t.dtype === \"int64\") {\n    throw new DTypeError(\"metrics do not support int64 tensors\");\n  }\n\n  const data = t.data;\n  if (!isTypedArray(data) || !isNumericTypedArray(data)) {\n    throw new DTypeError(\"metrics require numeric tensors\");\n  }\n  return data;\n}\n\nfunction readNumericLabel(\n  data: NumericTypedArray,\n  offsetter: FlatOffsetter,\n  index: number,\n  name: string\n) {\n  const value = getNumericElement(data, offsetter(index));\n  assertFiniteNumber(value, name, `index ${index}`);\n  return value;\n}\n\nfunction ensureBinaryValue(value: number, name: string, index: number): void {\n  if (value !== 0 && value !== 1) {\n    throw new InvalidParameterError(\n      `${name} must contain only binary values (0 or 1); found ${String(value)} at index ${index}`,\n      name,\n      value\n    );\n  }\n}\n\nfunction assertBinaryLabels(yTrue: Tensor, yPred: Tensor): void {\n  if (yTrue.dtype === \"string\" || yPred.dtype === \"string\") {\n    throw new InvalidParameterError(\n      \"classificationReport requires binary numeric labels (0 or 1)\",\n      \"yTrue\"\n    );\n  }\n\n  const yTrueData = getNumericLabelData(yTrue);\n  const yPredData = getNumericLabelData(yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n    const predVal = readNumericLabel(yPredData, predOffset, i, \"yPred\");\n    ensureBinaryValue(trueVal, \"yTrue\", i);\n    ensureBinaryValue(predVal, \"yPred\", i);\n  }\n}\n\ntype LabelKind = \"string\" | \"int64\" | \"numeric\";\n\nfunction assertComparableLabelTypes(yTrue: Tensor, yPred: Tensor): LabelKind {\n  const trueKind =\n    yTrue.dtype === \"string\" ? \"string\" : yTrue.dtype === \"int64\" ? \"int64\" : \"numeric\";\n  const predKind =\n    yPred.dtype === \"string\" ? \"string\" : yPred.dtype === \"int64\" ? \"int64\" : \"numeric\";\n\n  if (trueKind !== predKind) {\n    throw new DTypeError(\"yTrue and yPred must use compatible label types\");\n  }\n\n  return trueKind;\n}\n\nfunction readComparableLabel(t: Tensor, offsetter: FlatOffsetter, index: number, name: string) {\n  const offset = offsetter(index);\n  const data = t.data;\n\n  if (Array.isArray(data)) {\n    return getStringElement(data, offset);\n  }\n\n  if (data instanceof BigInt64Array) {\n    return getBigIntElement(data, offset);\n  }\n\n  if (!isTypedArray(data) || !isNumericTypedArray(data)) {\n    throw new DTypeError(`${name} must be numeric or string labels`);\n  }\n\n  const value = getNumericElement(data, offset);\n  assertFiniteNumber(value, name, `index ${index}`);\n  return value;\n}\n\ntype ClassStats = {\n  tp: number;\n  fp: number;\n  fn: number;\n  support: number;\n};\n\nfunction buildClassStats(yTrue: Tensor, yPred: Tensor) {\n  assertComparableLabelTypes(yTrue, yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  const stats = new Map<number | string | bigint, ClassStats>();\n  let totalTp = 0;\n  let totalFp = 0;\n  let totalFn = 0;\n\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readComparableLabel(yTrue, trueOffset, i, \"yTrue\");\n    const predVal = readComparableLabel(yPred, predOffset, i, \"yPred\");\n\n    let trueStats = stats.get(trueVal);\n    if (!trueStats) {\n      trueStats = { tp: 0, fp: 0, fn: 0, support: 0 };\n      stats.set(trueVal, trueStats);\n    }\n\n    let predStats = stats.get(predVal);\n    if (!predStats) {\n      predStats = { tp: 0, fp: 0, fn: 0, support: 0 };\n      stats.set(predVal, predStats);\n    }\n\n    trueStats.support += 1;\n\n    if (trueVal === predVal) {\n      trueStats.tp += 1;\n      totalTp += 1;\n    } else {\n      predStats.fp += 1;\n      trueStats.fn += 1;\n      totalFp += 1;\n      totalFn += 1;\n    }\n  }\n\n  const classes = Array.from(stats.keys()).sort((a, b) => {\n    if (typeof a === \"number\" && typeof b === \"number\") return a - b;\n    if (typeof a === \"string\" && typeof b === \"string\") return a.localeCompare(b);\n    if (typeof a === \"bigint\" && typeof b === \"bigint\") return a === b ? 0 : a < b ? -1 : 1;\n    return String(a).localeCompare(String(b));\n  });\n  return { classes, stats, totalTp, totalFp, totalFn };\n}\n\n/**\n * Calculates the accuracy classification score.\n *\n * Accuracy is the fraction of predictions that match the true labels.\n * It's the most intuitive performance measure but can be misleading\n * for imbalanced datasets.\n *\n * **Formula**: accuracy = (correct predictions) / (total predictions)\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated targets as returned by a classifier\n * @returns Accuracy score in range [0, 1], where 1 is perfect accuracy\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue and yPred use incompatible label types\n * @throws {DataValidationError} If numeric labels contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { accuracy, tensor } from 'deepbox/core';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const acc = accuracy(yTrue, yPred); // 0.8 (4 out of 5 correct)\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.accuracy_score.html | Scikit-learn accuracy_score}\n */\nexport function accuracy(yTrue: Tensor, yPred: Tensor): number {\n  // Validate input tensors have same size\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  // Handle empty input - return 0 for undefined accuracy\n  if (yTrue.size === 0) return 0;\n\n  assertComparableLabelTypes(yTrue, yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  // Count correct predictions\n  let correct = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readComparableLabel(yTrue, trueOffset, i, \"yTrue\");\n    const predVal = readComparableLabel(yPred, predOffset, i, \"yPred\");\n    if (trueVal === predVal) {\n      correct++;\n    }\n  }\n\n  // Return fraction of correct predictions\n  return correct / yTrue.size;\n}\n\n/**\n * Calculates the precision classification score.\n *\n * Precision is the ratio of true positives to all positive predictions.\n * It answers: \"Of all samples predicted as positive, how many are actually positive?\"\n * High precision means low false positive rate.\n *\n * **Formula**: precision = TP / (TP + FP)\n *\n * **Time Complexity**: O(n) for binary, O(n*k) for multiclass where k is number of classes\n * **Space Complexity**: O(k) for multiclass\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated targets as returned by a classifier\n * @param average - Averaging strategy: 'binary', 'micro', 'macro', 'weighted', or null\n *   - 'binary': Calculate metrics for positive class only (default)\n *   - 'micro': Calculate metrics globally by counting total TP, FP, FN\n *   - 'macro': Calculate metrics for each class, return unweighted mean\n *   - 'weighted': Calculate metrics for each class, return weighted mean by support\n *   - null: Return array of scores for each class\n * @returns Precision score(s) in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n * @throws {InvalidParameterError} If average is invalid or binary labels are not in {0,1}\n *\n * @example\n * ```ts\n * import { precision, tensor } from 'deepbox/core';\n *\n * // Binary classification\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const prec = precision(yTrue, yPred); // 1.0 (2 TP, 0 FP)\n *\n * // Multiclass\n * const yTrueMulti = tensor([0, 1, 2, 0, 1, 2]);\n * const yPredMulti = tensor([0, 2, 1, 0, 0, 1]);\n * const precMacro = precision(yTrueMulti, yPredMulti, 'macro');\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_score.html | Scikit-learn precision_score}\n */\nexport function precision(yTrue: Tensor, yPred: Tensor): number;\nexport function precision(\n  yTrue: Tensor,\n  yPred: Tensor,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\"\n): number;\nexport function precision(yTrue: Tensor, yPred: Tensor, average: null): number[];\nexport function precision(\n  yTrue: Tensor,\n  yPred: Tensor,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\" | null = \"binary\"\n): number | number[] {\n  // Validate input tensors have same size\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  // Handle empty input\n  if (yTrue.size === 0) return average === null ? [] : 0;\n\n  // Binary classification (assumes positive class is 1)\n  if (average === \"binary\") {\n    if (yTrue.dtype === \"string\" || yPred.dtype === \"string\") {\n      throw new InvalidParameterError(\n        \"Binary average requires numeric labels (0/1). Use 'macro', 'micro', or 'weighted' for string labels.\"\n      );\n    }\n    const yTrueData = getNumericLabelData(yTrue);\n    const yPredData = getNumericLabelData(yPred);\n    const trueOffset = createFlatOffsetter(yTrue);\n    const predOffset = createFlatOffsetter(yPred);\n\n    let tp = 0;\n    let fp = 0;\n\n    for (let i = 0; i < yTrue.size; i++) {\n      const trueVal = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n      const predVal = readNumericLabel(yPredData, predOffset, i, \"yPred\");\n      ensureBinaryValue(trueVal, \"yTrue\", i);\n      ensureBinaryValue(predVal, \"yPred\", i);\n\n      if (predVal === 1) {\n        if (trueVal === 1) {\n          tp++;\n        } else {\n          fp++;\n        }\n      }\n    }\n\n    return tp + fp === 0 ? 0 : tp / (tp + fp);\n  }\n\n  const { classes, stats, totalTp, totalFp } = buildClassStats(yTrue, yPred);\n\n  const precisions: number[] = [];\n  const supports: number[] = [];\n  for (const cls of classes) {\n    const classStats = stats.get(cls);\n    const tp = classStats?.tp ?? 0;\n    const fp = classStats?.fp ?? 0;\n    const support = classStats?.support ?? 0;\n    precisions.push(tp + fp === 0 ? 0 : tp / (tp + fp));\n    supports.push(support);\n  }\n\n  if (average === null) {\n    return precisions;\n  }\n\n  if (average === \"micro\") {\n    return totalTp + totalFp === 0 ? 0 : totalTp / (totalTp + totalFp);\n  }\n\n  if (average === \"macro\") {\n    const sum = precisions.reduce((acc, val) => acc + val, 0);\n    return precisions.length === 0 ? 0 : sum / precisions.length;\n  }\n\n  if (average === \"weighted\") {\n    let weightedSum = 0;\n    let totalSupport = 0;\n\n    for (let i = 0; i < precisions.length; i++) {\n      weightedSum += (precisions[i] ?? 0) * (supports[i] ?? 0);\n      totalSupport += supports[i] ?? 0;\n    }\n\n    return totalSupport === 0 ? 0 : weightedSum / totalSupport;\n  }\n\n  throw new InvalidParameterError(\n    `Invalid average parameter: ${average}. Must be one of: 'binary', 'micro', 'macro', 'weighted', or null`,\n    \"average\",\n    average\n  );\n}\n\n/**\n * Calculates the recall classification score (sensitivity, true positive rate).\n *\n * Recall is the ratio of true positives to all actual positive samples.\n * It answers: \"Of all actual positive samples, how many did we correctly identify?\"\n * High recall means low false negative rate.\n *\n * **Formula**: recall = TP / (TP + FN)\n *\n * **Time Complexity**: O(n) for binary, O(n*k) for multiclass where k is number of classes\n * **Space Complexity**: O(k) for multiclass\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated targets as returned by a classifier\n * @param average - Averaging strategy: 'binary', 'micro', 'macro', 'weighted', or null\n * @returns Recall score(s) in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n * @throws {InvalidParameterError} If average is invalid or binary labels are not in {0,1}\n *\n * @example\n * ```ts\n * import { recall, tensor } from 'deepbox/core';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const rec = recall(yTrue, yPred); // 0.667 (2 out of 3 positives found)\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.recall_score.html | Scikit-learn recall_score}\n */\nexport function recall(yTrue: Tensor, yPred: Tensor): number;\nexport function recall(\n  yTrue: Tensor,\n  yPred: Tensor,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\"\n): number;\nexport function recall(yTrue: Tensor, yPred: Tensor, average: null): number[];\nexport function recall(\n  yTrue: Tensor,\n  yPred: Tensor,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\" | null = \"binary\"\n): number | number[] {\n  // Validate input tensors have same size\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  // Handle empty input\n  if (yTrue.size === 0) return average === null ? [] : 0;\n\n  // Binary classification (assumes positive class is 1)\n  if (average === \"binary\") {\n    if (yTrue.dtype === \"string\" || yPred.dtype === \"string\") {\n      throw new InvalidParameterError(\n        \"Binary average requires numeric labels (0/1). Use 'macro', 'micro', or 'weighted' for string labels.\"\n      );\n    }\n    const yTrueData = getNumericLabelData(yTrue);\n    const yPredData = getNumericLabelData(yPred);\n    const trueOffset = createFlatOffsetter(yTrue);\n    const predOffset = createFlatOffsetter(yPred);\n\n    let tp = 0;\n    let fn = 0;\n\n    for (let i = 0; i < yTrue.size; i++) {\n      const trueVal = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n      const predVal = readNumericLabel(yPredData, predOffset, i, \"yPred\");\n      ensureBinaryValue(trueVal, \"yTrue\", i);\n      ensureBinaryValue(predVal, \"yPred\", i);\n\n      if (trueVal === 1) {\n        if (predVal === 1) {\n          tp++;\n        } else {\n          fn++;\n        }\n      }\n    }\n\n    return tp + fn === 0 ? 0 : tp / (tp + fn);\n  }\n\n  const { classes, stats, totalTp, totalFn } = buildClassStats(yTrue, yPred);\n\n  const recalls: number[] = [];\n  const supports: number[] = [];\n  for (const cls of classes) {\n    const classStats = stats.get(cls);\n    const tp = classStats?.tp ?? 0;\n    const fn = classStats?.fn ?? 0;\n    const support = classStats?.support ?? 0;\n    recalls.push(tp + fn === 0 ? 0 : tp / (tp + fn));\n    supports.push(support);\n  }\n\n  if (average === null) {\n    return recalls;\n  }\n\n  if (average === \"micro\") {\n    return totalTp + totalFn === 0 ? 0 : totalTp / (totalTp + totalFn);\n  }\n\n  if (average === \"macro\") {\n    const sum = recalls.reduce((acc, val) => acc + val, 0);\n    return recalls.length === 0 ? 0 : sum / recalls.length;\n  }\n\n  if (average === \"weighted\") {\n    let weightedSum = 0;\n    let totalSupport = 0;\n\n    for (let i = 0; i < recalls.length; i++) {\n      weightedSum += (recalls[i] ?? 0) * (supports[i] ?? 0);\n      totalSupport += supports[i] ?? 0;\n    }\n\n    return totalSupport === 0 ? 0 : weightedSum / totalSupport;\n  }\n\n  throw new InvalidParameterError(\n    `Invalid average parameter: ${average}. Must be one of: 'binary', 'micro', 'macro', 'weighted', or null`,\n    \"average\",\n    average\n  );\n}\n\n/**\n * Calculates the F1 score (harmonic mean of precision and recall).\n *\n * F1 score is the harmonic mean of precision and recall, providing a single\n * metric that balances both concerns. It's especially useful when you need\n * to balance false positives and false negatives.\n *\n * **Formula**: F1 = 2 * (precision * recall) / (precision + recall)\n *\n * **Time Complexity**: O(n) for binary, O(n*k) for multiclass\n * **Space Complexity**: O(k) for multiclass\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated targets as returned by a classifier\n * @param average - Averaging strategy: 'binary', 'micro', 'macro', 'weighted', or null\n * @returns F1 score(s) in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n * @throws {InvalidParameterError} If average is invalid or binary labels are not in {0,1}\n *\n * @example\n * ```ts\n * import { f1Score, tensor } from 'deepbox/core';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const f1 = f1Score(yTrue, yPred); // 0.8\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.f1_score.html | Scikit-learn f1_score}\n */\nexport function f1Score(yTrue: Tensor, yPred: Tensor): number;\nexport function f1Score(\n  yTrue: Tensor,\n  yPred: Tensor,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\"\n): number;\nexport function f1Score(yTrue: Tensor, yPred: Tensor, average: null): number[];\nexport function f1Score(\n  yTrue: Tensor,\n  yPred: Tensor,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\" | null = \"binary\"\n): number | number[] {\n  // For binary and micro, computing F1 from scalar P and R is correct.\n  // For macro and weighted, we must compute per-class F1 first, then average,\n  // because the harmonic mean is non-linear: avg(F1_i) ≠ F1(avg(P_i), avg(R_i)).\n  if (average === \"binary\" || average === \"micro\") {\n    const p = precision(yTrue, yPred, average);\n    const r = recall(yTrue, yPred, average);\n    return p + r === 0 ? 0 : (2 * p * r) / (p + r);\n  }\n\n  // Compute per-class precision and recall\n  const prec = precision(yTrue, yPred, null);\n  const rec = recall(yTrue, yPred, null);\n\n  // Compute per-class F1 scores\n  const f1Scores: number[] = [];\n  for (let i = 0; i < prec.length; i++) {\n    const p = prec[i] ?? 0;\n    const r = rec[i] ?? 0;\n    f1Scores.push(p + r === 0 ? 0 : (2 * p * r) / (p + r));\n  }\n\n  if (average === null) {\n    return f1Scores;\n  }\n\n  if (f1Scores.length === 0) return 0;\n\n  if (average === \"macro\") {\n    const sum = f1Scores.reduce((acc, val) => acc + val, 0);\n    return sum / f1Scores.length;\n  }\n\n  if (average === \"weighted\") {\n    const { classes, stats } = buildClassStats(yTrue, yPred);\n    let weightedSum = 0;\n    let totalSupport = 0;\n    for (let i = 0; i < f1Scores.length; i++) {\n      const cls = classes[i];\n      const support = cls !== undefined ? (stats.get(cls)?.support ?? 0) : 0;\n      weightedSum += (f1Scores[i] ?? 0) * support;\n      totalSupport += support;\n    }\n    return totalSupport === 0 ? 0 : weightedSum / totalSupport;\n  }\n\n  throw new InvalidParameterError(\n    `Invalid average parameter: ${average}. Must be one of: 'binary', 'micro', 'macro', 'weighted', or null`,\n    \"average\",\n    average\n  );\n}\n\n/**\n * Calculates the F-beta score.\n *\n * F-beta score is a weighted harmonic mean of precision and recall, where\n * beta controls the trade-off between precision and recall.\n * - beta < 1: More weight on precision\n * - beta = 1: Equal weight (equivalent to F1 score)\n * - beta > 1: More weight on recall\n *\n * **Formula**: F_beta = (1 + beta²) * (precision * recall) / (beta² * precision + recall)\n *\n * **Time Complexity**: O(n) for binary, O(n*k) for multiclass\n * **Space Complexity**: O(k) for multiclass\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated targets as returned by a classifier\n * @param beta - Weight of recall vs precision (beta > 1 favors recall, beta < 1 favors precision)\n * @param average - Averaging strategy: 'binary', 'micro', 'macro', 'weighted', or null\n * @returns F-beta score(s) in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { fbetaScore, tensor } from 'deepbox/core';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const fb2 = fbetaScore(yTrue, yPred, 2); // Favors recall\n * const fb05 = fbetaScore(yTrue, yPred, 0.5); // Favors precision\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.fbeta_score.html | Scikit-learn fbeta_score}\n */\nexport function fbetaScore(yTrue: Tensor, yPred: Tensor, beta: number): number;\nexport function fbetaScore(\n  yTrue: Tensor,\n  yPred: Tensor,\n  beta: number,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\"\n): number;\nexport function fbetaScore(yTrue: Tensor, yPred: Tensor, beta: number, average: null): number[];\nexport function fbetaScore(\n  yTrue: Tensor,\n  yPred: Tensor,\n  beta: number,\n  average: \"binary\" | \"micro\" | \"macro\" | \"weighted\" | null = \"binary\"\n) {\n  if (!Number.isFinite(beta) || beta <= 0) {\n    throw new InvalidParameterError(\"beta must be a positive finite number\", \"beta\", beta);\n  }\n\n  const betaSq = beta * beta;\n\n  // For binary and micro, computing fbeta from scalar P and R is correct.\n  // For macro and weighted, we must compute per-class fbeta first, then average,\n  // because the weighted harmonic mean is non-linear.\n  if (average === \"binary\" || average === \"micro\") {\n    const p = precision(yTrue, yPred, average);\n    const r = recall(yTrue, yPred, average);\n    return p + r === 0 ? 0 : ((1 + betaSq) * p * r) / (betaSq * p + r);\n  }\n\n  // Compute per-class precision and recall\n  const prec = precision(yTrue, yPred, null);\n  const rec = recall(yTrue, yPred, null);\n\n  // Compute per-class fbeta scores\n  const fbetaScores: number[] = [];\n  for (let i = 0; i < prec.length; i++) {\n    const p = prec[i] ?? 0;\n    const r = rec[i] ?? 0;\n    fbetaScores.push(p + r === 0 ? 0 : ((1 + betaSq) * p * r) / (betaSq * p + r));\n  }\n\n  if (average === null) {\n    return fbetaScores;\n  }\n\n  if (fbetaScores.length === 0) return 0;\n\n  if (average === \"macro\") {\n    const sum = fbetaScores.reduce((acc, val) => acc + val, 0);\n    return sum / fbetaScores.length;\n  }\n\n  if (average === \"weighted\") {\n    const { classes, stats } = buildClassStats(yTrue, yPred);\n    let weightedSum = 0;\n    let totalSupport = 0;\n    for (let i = 0; i < fbetaScores.length; i++) {\n      const cls = classes[i];\n      const support = cls !== undefined ? (stats.get(cls)?.support ?? 0) : 0;\n      weightedSum += (fbetaScores[i] ?? 0) * support;\n      totalSupport += support;\n    }\n    return totalSupport === 0 ? 0 : weightedSum / totalSupport;\n  }\n\n  throw new InvalidParameterError(\n    `Invalid average parameter: ${average}. Must be one of: 'binary', 'micro', 'macro', 'weighted', or null`,\n    \"average\",\n    average\n  );\n}\n\n/**\n * Computes the confusion matrix to evaluate classification accuracy.\n *\n * A confusion matrix is a table showing the counts of correct and incorrect\n * predictions broken down by each class. Rows represent true labels,\n * columns represent predicted labels.\n *\n * **Time Complexity**: O(n + k²) where n is number of samples, k is number of classes\n * **Space Complexity**: O(k²)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated targets as returned by a classifier\n * @returns Confusion matrix as a 2D tensor of shape [n_classes, n_classes]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { confusionMatrix, tensor } from 'deepbox/core';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const cm = confusionMatrix(yTrue, yPred);\n * // [[2, 0],\n * //  [1, 2]]\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.confusion_matrix.html | Scikit-learn confusion_matrix}\n */\nexport function confusionMatrix(yTrue: Tensor, yPred: Tensor): Tensor {\n  // Validate input tensors have same size\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  assertComparableLabelTypes(yTrue, yPred);\n\n  if (yTrue.size === 0) {\n    return tensor([]).reshape([0, 0]);\n  }\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  // Collect all unique labels from both true and predicted values\n  const labelSet = new Set<number | string | bigint>();\n  for (let i = 0; i < yTrue.size; i++) {\n    labelSet.add(readComparableLabel(yTrue, trueOffset, i, \"yTrue\"));\n    labelSet.add(readComparableLabel(yPred, predOffset, i, \"yPred\"));\n  }\n\n  // Sort labels for consistent ordering\n  const labels = Array.from(labelSet).sort((a, b) => {\n    if (typeof a === \"number\" && typeof b === \"number\") return a - b;\n    if (typeof a === \"string\" && typeof b === \"string\") return a.localeCompare(b);\n    if (typeof a === \"bigint\" && typeof b === \"bigint\") return a === b ? 0 : a < b ? -1 : 1;\n    return String(a).localeCompare(String(b));\n  });\n\n  // Create mapping from label to matrix index\n  const labelToIndex = new Map<number | string | bigint, number>();\n  for (let i = 0; i < labels.length; i++) {\n    const label = labels[i];\n    if (label === undefined) continue;\n    labelToIndex.set(label, i);\n  }\n\n  // Initialize confusion matrix with zeros\n  const nClasses = labels.length;\n  const matrix = Array.from({ length: nClasses }, () => new Array<number>(nClasses).fill(0));\n\n  // Populate confusion matrix by counting predictions\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueLabel = readComparableLabel(yTrue, trueOffset, i, \"yTrue\");\n    const predLabel = readComparableLabel(yPred, predOffset, i, \"yPred\");\n\n    // Get matrix indices for this true/pred label pair\n    const r = labelToIndex.get(trueLabel);\n    const c = labelToIndex.get(predLabel);\n    if (r === undefined || c === undefined) continue;\n\n    // Increment count at [true_label, pred_label]\n    const row = matrix[r];\n    if (row) row[c] = (row[c] ?? 0) + 1;\n  }\n\n  // Convert to tensor and return\n  return tensor(matrix);\n}\n\n/**\n * Generates a text classification report showing main classification metrics.\n *\n * Provides a comprehensive summary including precision, recall, F1-score,\n * and accuracy for the classification task.\n *\n * **Time Complexity**: O(n * k) where n is the number of samples and k is the number of classes\n * **Space Complexity**: O(k) where k is the number of classes\n *\n * @param yTrue - Ground truth (correct) binary target values (0 or 1)\n * @param yPred - Estimated binary targets as returned by a classifier (0 or 1)\n * @returns Formatted string report with per-class and aggregate classification metrics\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or int64)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n * @throws {InvalidParameterError} If labels are not binary (0 or 1)\n *\n * @example\n * ```ts\n * import { classificationReport, tensor } from 'deepbox/core';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * console.log(classificationReport(yTrue, yPred));\n * // Classification Report:\n * //   Precision: 1.0000\n * //   Recall:    0.6667\n * //   F1-Score:  0.8000\n * //   Accuracy:  0.8000\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.classification_report.html | Scikit-learn classification_report}\n */\nexport function classificationReport(yTrue: Tensor, yPred: Tensor): string {\n  // Validate input tensors have same size\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  // Handle empty input\n  if (yTrue.size === 0) return \"Classification Report:\\n  (empty)\";\n\n  assertBinaryLabels(yTrue, yPred);\n\n  // Calculate per-class metrics\n  const { classes, stats } = buildClassStats(yTrue, yPred);\n  const precs = precision(yTrue, yPred, null);\n  const recs = recall(yTrue, yPred, null);\n  const f1s = f1Score(yTrue, yPred, null);\n  const acc = accuracy(yTrue, yPred);\n\n  // Determine width for class names\n  const maxClassLen = Math.max(...classes.map((c) => String(c).length), \"Class\".length);\n  const colWidth = Math.max(12, maxClassLen + 2);\n\n  // Header\n  let report = \"Classification Report:\\n\";\n  report +=\n    \"Class\".padEnd(colWidth) +\n    \"Precision\".padEnd(12) +\n    \"Recall\".padEnd(12) +\n    \"F1-Score\".padEnd(12) +\n    \"Support\\n\";\n  report += `${\"-\".repeat(colWidth + 36 + 7)}\\n`;\n\n  let totalSupport = 0;\n  let weightedPrec = 0;\n  let weightedRec = 0;\n  let weightedF1 = 0;\n  let macroPrec = 0;\n  let macroRec = 0;\n  let macroF1 = 0;\n\n  for (const [i, cls] of classes.entries()) {\n    const p = precs[i] ?? 0;\n    const r = recs[i] ?? 0;\n    const f1 = f1s[i] ?? 0;\n    const s = stats.get(cls)?.support ?? 0;\n\n    totalSupport += s;\n    weightedPrec += p * s;\n    weightedRec += r * s;\n    weightedF1 += f1 * s;\n    macroPrec += p;\n    macroRec += r;\n    macroF1 += f1;\n\n    report +=\n      String(cls).padEnd(colWidth) +\n      p.toFixed(4).padEnd(12) +\n      r.toFixed(4).padEnd(12) +\n      f1.toFixed(4).padEnd(12) +\n      String(s) +\n      \"\\n\";\n  }\n\n  report += \"\\n\";\n\n  // Averages\n\n  const nClasses = classes.length;\n  if (nClasses > 0) {\n    macroPrec /= nClasses;\n    macroRec /= nClasses;\n    macroF1 /= nClasses;\n  } else {\n    macroPrec = 0;\n    macroRec = 0;\n    macroF1 = 0;\n  }\n\n  weightedPrec = totalSupport === 0 ? 0 : weightedPrec / totalSupport;\n  weightedRec = totalSupport === 0 ? 0 : weightedRec / totalSupport;\n  weightedF1 = totalSupport === 0 ? 0 : weightedF1 / totalSupport;\n\n  report +=\n    \"Accuracy\".padEnd(colWidth) +\n    \"\".padEnd(12) +\n    \"\".padEnd(12) +\n    acc.toFixed(4).padEnd(12) +\n    String(totalSupport) +\n    \"\\n\";\n  report +=\n    \"Macro Avg\".padEnd(colWidth) +\n    macroPrec.toFixed(4).padEnd(12) +\n    macroRec.toFixed(4).padEnd(12) +\n    macroF1.toFixed(4).padEnd(12) +\n    String(totalSupport) +\n    \"\\n\";\n  report +=\n    \"Weighted Avg\".padEnd(colWidth) +\n    weightedPrec.toFixed(4).padEnd(12) +\n    weightedRec.toFixed(4).padEnd(12) +\n    weightedF1.toFixed(4).padEnd(12) +\n    String(totalSupport);\n\n  return report;\n}\n\n/**\n * ROC curve data.\n *\n * Computes Receiver Operating Characteristic (ROC) curve for binary classification.\n * The ROC curve shows the trade-off between true positive rate and false positive rate\n * at various threshold settings.\n *\n * **Returns**: [fpr, tpr, thresholds]\n * - fpr: False positive rates\n * - tpr: True positive rates\n * - thresholds: Decision thresholds (in descending order)\n *\n * **Edge Cases**:\n * - Returns empty tensors if yTrue contains only one class\n * - Handles tied scores by grouping them at the same threshold\n *\n * @param yTrue - Ground truth binary labels (must be 0 or 1)\n * @param yScore - Target scores (higher score = more likely positive class)\n * @returns Tuple of [fpr, tpr, thresholds] tensors\n *\n * @throws {ShapeError} If yTrue and yScore have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If inputs are not numeric (string or int64)\n * @throws {InvalidParameterError} If yTrue contains non-binary values\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_curve.html | Scikit-learn roc_curve}\n */\nexport function rocCurve(yTrue: Tensor, yScore: Tensor): [Tensor, Tensor, Tensor] {\n  assertSameSizeVectors(yTrue, yScore, \"yTrue\", \"yScore\");\n\n  const n = yTrue.size;\n\n  if (n === 0) return [tensor([]), tensor([]), tensor([])];\n\n  const yTrueData = getNumericLabelData(yTrue);\n  const yScoreData = getNumericLabelData(yScore);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const scoreOffset = createFlatOffsetter(yScore);\n\n  // Create pairs of (score, label)\n  const pairs: Array<{ score: number; label: number }> = [];\n  let nPos = 0;\n  let nNeg = 0;\n  for (let i = 0; i < n; i++) {\n    const label = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n    ensureBinaryValue(label, \"yTrue\", i);\n    const score = readNumericLabel(yScoreData, scoreOffset, i, \"yScore\");\n    pairs.push({ score, label });\n    if (label === 1) nPos++;\n    else nNeg++;\n  }\n\n  // Sort by score descending\n  pairs.sort((a, b) => b.score - a.score);\n\n  if (nPos === 0 || nNeg === 0) return [tensor([]), tensor([]), tensor([])];\n\n  const fpr = [0];\n  const tpr = [0];\n  const thresholds = [Infinity];\n\n  let tp = 0;\n  let fp = 0;\n  let idx = 0;\n  while (idx < pairs.length) {\n    const threshold = pairs[idx]?.score ?? 0;\n\n    // Consume all samples with the same score as one step.\n    while (idx < pairs.length && (pairs[idx]?.score ?? 0) === threshold) {\n      const label = pairs[idx]?.label ?? 0;\n      if (label === 1) tp++;\n      else fp++;\n      idx++;\n    }\n\n    fpr.push(fp / nNeg);\n    tpr.push(tp / nPos);\n    thresholds.push(threshold);\n  }\n\n  return [tensor(fpr), tensor(tpr), tensor(thresholds)];\n}\n\n/**\n * Area Under ROC Curve (AUC-ROC).\n *\n * Computes the Area Under the Receiver Operating Characteristic Curve.\n * AUC represents the probability that a randomly chosen positive sample\n * is ranked higher than a randomly chosen negative sample.\n *\n * **Range**: [0, 1], where 1 is perfect and 0.5 is random.\n *\n * **Time Complexity**: O(n log n) due to sorting\n * **Space Complexity**: O(n)\n *\n * @param yTrue - Ground truth binary labels (must be 0 or 1)\n * @param yScore - Target scores (higher score = more likely positive class)\n * @returns AUC score in range [0, 1], or 0.5 if ROC curve cannot be computed\n *\n * @throws {ShapeError} If yTrue and yScore have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If inputs are not numeric (string or int64)\n * @throws {InvalidParameterError} If yTrue contains non-binary values\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { rocAucScore, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 0, 1, 1]);\n * const yScore = tensor([0.1, 0.4, 0.35, 0.8]);\n * const auc = rocAucScore(yTrue, yScore); // ~0.75\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html | Scikit-learn roc_auc_score}\n */\nexport function rocAucScore(yTrue: Tensor, yScore: Tensor): number {\n  const curves = rocCurve(yTrue, yScore);\n  const fprT = curves[0];\n  const tprT = curves[1];\n  if (!fprT || !tprT || fprT.size === 0 || tprT.size === 0) return 0.5;\n\n  const fprData = getNumericLabelData(fprT);\n  const tprData = getNumericLabelData(tprT);\n  const fprOffset = createFlatOffsetter(fprT);\n  const tprOffset = createFlatOffsetter(tprT);\n\n  let auc = 0;\n  let prevX = 0;\n  let prevY = 0;\n  for (let i = 1; i < fprT.size; i++) {\n    const x = readNumericLabel(fprData, fprOffset, i, \"fpr\");\n    const y = readNumericLabel(tprData, tprOffset, i, \"tpr\");\n    auc += (x - prevX) * ((y + prevY) / 2);\n    prevX = x;\n    prevY = y;\n  }\n\n  return auc;\n}\n\n/**\n * Precision-Recall curve.\n *\n * Computes precision-recall pairs for different probability thresholds.\n * Useful for evaluating classifiers on imbalanced datasets where ROC curves\n * may be overly optimistic.\n *\n * **Returns**: [precision, recall, thresholds] as a tuple of tensors\n * - precision: Precision values at each threshold\n * - recall: Recall values at each threshold\n * - thresholds: Decision thresholds (in descending order)\n *\n * **Time Complexity**: O(n log n) due to sorting\n * **Space Complexity**: O(n)\n *\n * @param yTrue - Ground truth binary labels (0 or 1)\n * @param yScore - Target scores (higher score = more likely positive class)\n * @returns Tuple of [precision, recall, thresholds] tensors\n *\n * @throws {ShapeError} If yTrue and yScore have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If inputs are not numeric (string or int64)\n * @throws {InvalidParameterError} If yTrue contains non-binary values\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { precisionRecallCurve, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 0, 1, 1]);\n * const yScore = tensor([0.1, 0.4, 0.35, 0.8]);\n * const [prec, rec, thresh] = precisionRecallCurve(yTrue, yScore);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_curve.html | Scikit-learn precision_recall_curve}\n */\nexport function precisionRecallCurve(yTrue: Tensor, yScore: Tensor): [Tensor, Tensor, Tensor] {\n  assertSameSizeVectors(yTrue, yScore, \"yTrue\", \"yScore\");\n\n  const n = yTrue.size;\n\n  if (n === 0) return [tensor([]), tensor([]), tensor([])];\n\n  const yTrueData = getNumericLabelData(yTrue);\n  const yScoreData = getNumericLabelData(yScore);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const scoreOffset = createFlatOffsetter(yScore);\n\n  // Create pairs of (score, label)\n  const pairs: Array<{ score: number; label: number }> = [];\n  let nPos = 0;\n  for (let i = 0; i < n; i++) {\n    const label = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n    ensureBinaryValue(label, \"yTrue\", i);\n    const score = readNumericLabel(yScoreData, scoreOffset, i, \"yScore\");\n    pairs.push({ score, label });\n    if (label === 1) nPos++;\n  }\n\n  // Sort by score descending\n  pairs.sort((a, b) => b.score - a.score);\n\n  if (nPos === 0) return [tensor([]), tensor([]), tensor([])];\n\n  const prec = [1];\n  const rec = [0];\n  const thresholds = [Infinity];\n\n  let tp = 0;\n  let fp = 0;\n  let idx = 0;\n  while (idx < pairs.length) {\n    const threshold = pairs[idx]?.score ?? 0;\n    while (idx < pairs.length && (pairs[idx]?.score ?? 0) === threshold) {\n      const label = pairs[idx]?.label ?? 0;\n      if (label === 1) tp++;\n      else fp++;\n      idx++;\n    }\n\n    const precisionVal = tp + fp === 0 ? 1 : tp / (tp + fp);\n    const recallVal = tp / nPos;\n    prec.push(precisionVal);\n    rec.push(recallVal);\n    thresholds.push(threshold);\n  }\n\n  return [tensor(prec), tensor(rec), tensor(thresholds)];\n}\n\n/**\n * Average precision score.\n *\n * Computes the average precision (AP) from prediction scores. AP summarizes\n * a precision-recall curve as the weighted mean of precisions achieved at\n * each threshold, with the increase in recall from the previous threshold\n * used as the weight.\n *\n * **Range**: [0, 1], where 1 is perfect.\n *\n * **Time Complexity**: O(n log n) due to sorting\n * **Space Complexity**: O(n)\n *\n * @param yTrue - Ground truth binary labels (0 or 1)\n * @param yScore - Target scores (higher score = more likely positive class)\n * @returns Average precision score in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yScore have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If inputs are not numeric (string or int64)\n * @throws {InvalidParameterError} If yTrue contains non-binary values\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { averagePrecisionScore, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 0, 1, 1]);\n * const yScore = tensor([0.1, 0.4, 0.35, 0.8]);\n * const ap = averagePrecisionScore(yTrue, yScore);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.average_precision_score.html | Scikit-learn average_precision_score}\n */\nexport function averagePrecisionScore(yTrue: Tensor, yScore: Tensor): number {\n  const curves = precisionRecallCurve(yTrue, yScore);\n  const precT = curves[0];\n  const recT = curves[1];\n  if (!precT || !recT || precT.size === 0 || recT.size === 0) return 0;\n\n  const precData = getNumericLabelData(precT);\n  const recData = getNumericLabelData(recT);\n  const precOffset = createFlatOffsetter(precT);\n  const recOffset = createFlatOffsetter(recT);\n\n  let ap = 0;\n  let prevRecall = readNumericLabel(recData, recOffset, 0, \"recall\");\n  for (let i = 1; i < recT.size; i++) {\n    const recall = readNumericLabel(recData, recOffset, i, \"recall\");\n    const precision = readNumericLabel(precData, precOffset, i, \"precision\");\n    const deltaRecall = recall - prevRecall;\n    if (deltaRecall > 0) {\n      ap += deltaRecall * precision;\n    }\n    prevRecall = recall;\n  }\n\n  return ap;\n}\n\n/**\n * Log loss (logistic loss, cross-entropy loss).\n *\n * Measures the performance of a classification model where the prediction is a probability\n * value between 0 and 1. Lower log loss indicates better predictions.\n *\n * **Formula**: -1/n * Σ(y_true * log(y_pred) + (1 - y_true) * log(1 - y_pred))\n *\n * **Edge Cases**:\n * - Predictions are clipped to [1e-15, 1-1e-15] to avoid log(0)\n * - Returns 0 for empty inputs\n *\n * @param yTrue - Ground truth binary labels (0 or 1)\n * @param yPred - Predicted probabilities (must be in range [0, 1])\n * @returns Log loss value (lower is better, 0 is perfect)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If inputs are not numeric (string or int64)\n * @throws {InvalidParameterError} If yTrue is not binary or yPred is outside [0, 1]\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.log_loss.html | Scikit-learn log_loss}\n */\nexport function logLoss(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  if (yTrue.size === 0) return 0;\n\n  const yTrueData = getNumericLabelData(yTrue);\n  const yPredData = getNumericLabelData(yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  const eps = 1e-15;\n  let loss = 0;\n\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n    ensureBinaryValue(trueVal, \"yTrue\", i);\n    const predRaw = readNumericLabel(yPredData, predOffset, i, \"yPred\");\n    if (predRaw < 0 || predRaw > 1) {\n      throw new InvalidParameterError(\n        `yPred must contain probabilities in range [0, 1], found ${String(predRaw)} at index ${i}`,\n        \"yPred\",\n        predRaw\n      );\n    }\n    const predVal = Math.max(eps, Math.min(1 - eps, predRaw));\n    loss -= trueVal * Math.log(predVal) + (1 - trueVal) * Math.log(1 - predVal);\n  }\n\n  return loss / yTrue.size;\n}\n\n/**\n * Hamming loss.\n *\n * Computes the fraction of labels that are incorrectly predicted.\n * For binary classification, this equals 1 - accuracy.\n *\n * **Formula**: hamming_loss = (incorrect predictions) / (total predictions)\n *\n * **Range**: [0, 1], where 0 is perfect.\n *\n * **Time Complexity**: O(n)\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth target values\n * @param yPred - Estimated targets as returned by a classifier\n * @returns Hamming loss in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue and yPred use incompatible label types\n * @throws {DataValidationError} If numeric labels contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { hammingLoss, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const loss = hammingLoss(yTrue, yPred); // 0.2\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.hamming_loss.html | Scikit-learn hamming_loss}\n */\nexport function hammingLoss(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  if (yTrue.size === 0) return 0;\n\n  assertComparableLabelTypes(yTrue, yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let errors = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readComparableLabel(yTrue, trueOffset, i, \"yTrue\");\n    const predVal = readComparableLabel(yPred, predOffset, i, \"yPred\");\n    if (trueVal !== predVal) {\n      errors++;\n    }\n  }\n\n  return errors / yTrue.size;\n}\n\n/**\n * Jaccard similarity score (Intersection over Union).\n *\n * Computes the Jaccard similarity coefficient between two binary label sets.\n * Also known as the Jaccard index or Intersection over Union (IoU).\n *\n * **Formula**: jaccard = TP / (TP + FP + FN)\n *\n * **Edge Cases**:\n * - If both yTrue and yPred contain no positive labels (TP + FP + FN = 0),\n *   returns 1 to reflect perfect similarity of empty sets.\n *\n * **Range**: [0, 1], where 1 is perfect.\n *\n * **Time Complexity**: O(n)\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth binary labels (0 or 1)\n * @param yPred - Predicted binary labels (0 or 1)\n * @returns Jaccard score in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n * @throws {InvalidParameterError} If labels are not binary (0 or 1)\n *\n * @example\n * ```ts\n * import { jaccardScore, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const score = jaccardScore(yTrue, yPred); // 0.667\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.jaccard_score.html | Scikit-learn jaccard_score}\n */\nexport function jaccardScore(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  if (yTrue.size === 0) return 1;\n\n  const yTrueData = getNumericLabelData(yTrue);\n  const yPredData = getNumericLabelData(yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let tp = 0,\n    fp = 0,\n    fn = 0;\n\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n    const predVal = readNumericLabel(yPredData, predOffset, i, \"yPred\");\n    ensureBinaryValue(trueVal, \"yTrue\", i);\n    ensureBinaryValue(predVal, \"yPred\", i);\n\n    if (trueVal === 1 && predVal === 1) tp++;\n    else if (trueVal === 0 && predVal === 1) fp++;\n    else if (trueVal === 1 && predVal === 0) fn++;\n  }\n\n  return tp + fp + fn === 0 ? 1 : tp / (tp + fp + fn);\n}\n\n/**\n * Matthews correlation coefficient (MCC).\n *\n * Computes the Matthews correlation coefficient, a balanced measure that\n * can be used even if the classes are of very different sizes. MCC is\n * considered one of the best metrics for binary classification.\n *\n * **Formula**: MCC = (TP*TN - FP*FN) / sqrt((TP+FP)(TP+FN)(TN+FP)(TN+FN))\n *\n * **Range**: [-1, 1], where 1 is perfect, 0 is random, -1 is inverse.\n *\n * **Time Complexity**: O(n)\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth binary labels (0 or 1)\n * @param yPred - Predicted binary labels (0 or 1)\n * @returns MCC score in range [-1, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { matthewsCorrcoef, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const mcc = matthewsCorrcoef(yTrue, yPred); // ~0.667\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.matthews_corrcoef.html | Scikit-learn matthews_corrcoef}\n */\nexport function matthewsCorrcoef(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  if (yTrue.size === 0) return 0;\n\n  const yTrueData = getNumericLabelData(yTrue);\n  const yPredData = getNumericLabelData(yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let tp = 0,\n    tn = 0,\n    fp = 0,\n    fn = 0;\n\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n    const predVal = readNumericLabel(yPredData, predOffset, i, \"yPred\");\n    ensureBinaryValue(trueVal, \"yTrue\", i);\n    ensureBinaryValue(predVal, \"yPred\", i);\n\n    if (trueVal === 1 && predVal === 1) tp++;\n    else if (trueVal === 0 && predVal === 0) tn++;\n    else if (trueVal === 0 && predVal === 1) fp++;\n    else if (trueVal === 1 && predVal === 0) fn++;\n  }\n\n  const numerator = tp * tn - fp * fn;\n  const denominator = Math.sqrt((tp + fp) * (tp + fn) * (tn + fp) * (tn + fn));\n\n  return denominator === 0 ? 0 : numerator / denominator;\n}\n\n/**\n * Cohen's kappa score.\n *\n * Computes Cohen's kappa, a statistic that measures inter-annotator agreement.\n * It is generally thought to be a more robust measure than simple percent\n * agreement since it takes into account the possibility of agreement by chance.\n *\n * **Formula**: kappa = (p_o - p_e) / (1 - p_e)\n * - p_o: observed agreement\n * - p_e: expected agreement by chance\n *\n * **Range**: [-1, 1], where 1 is perfect agreement, 0 is chance, <0 is worse than chance.\n *\n * **Time Complexity**: O(n)\n * **Space Complexity**: O(k) where k is number of classes\n *\n * @param yTrue - Ground truth labels\n * @param yPred - Predicted labels\n * @returns Kappa score in range [-1, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { cohenKappaScore, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 1, 1, 0, 1]);\n * const yPred = tensor([0, 1, 0, 0, 1]);\n * const kappa = cohenKappaScore(yTrue, yPred); // ~0.615\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.cohen_kappa_score.html | Scikit-learn cohen_kappa_score}\n */\nexport function cohenKappaScore(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  const n = yTrue.size;\n\n  if (n === 0) return 0;\n\n  const yTrueData = getNumericLabelData(yTrue);\n  const yPredData = getNumericLabelData(yPred);\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  // Observed agreement and class marginals\n  let po = 0;\n  const trueCount = new Map<number, number>();\n  const predCount = new Map<number, number>();\n\n  for (let i = 0; i < n; i++) {\n    const t = readNumericLabel(yTrueData, trueOffset, i, \"yTrue\");\n    const p = readNumericLabel(yPredData, predOffset, i, \"yPred\");\n    if (t === p) {\n      po++;\n    }\n    trueCount.set(t, (trueCount.get(t) ?? 0) + 1);\n    predCount.set(p, (predCount.get(p) ?? 0) + 1);\n  }\n\n  po /= n;\n\n  let pe = 0;\n  const allClasses = new Set([...trueCount.keys(), ...predCount.keys()]);\n  for (const c of allClasses) {\n    const trueProb = (trueCount.get(c) ?? 0) / n;\n    const predProb = (predCount.get(c) ?? 0) / n;\n    pe += trueProb * predProb;\n  }\n\n  const denom = 1 - pe;\n  if (denom === 0) return po === 1 ? 1 : 0;\n  return (po - pe) / denom;\n}\n\n/**\n * Balanced accuracy score.\n *\n * Computes the balanced accuracy, which is the macro-averaged recall.\n * It is useful for imbalanced datasets where regular accuracy can be misleading.\n *\n * **Formula**: balanced_accuracy = (1/n_classes) * Σ(recall_per_class)\n *\n * **Range**: [0, 1], where 1 is perfect.\n *\n * **Time Complexity**: O(n * k) where k is number of classes\n * **Space Complexity**: O(k)\n *\n * @param yTrue - Ground truth labels\n * @param yPred - Predicted labels\n * @returns Balanced accuracy score in range [0, 1]\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If labels are not numeric (string or bigint)\n * @throws {DataValidationError} If labels contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { balancedAccuracyScore, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([0, 0, 0, 0, 1]); // Imbalanced\n * const yPred = tensor([0, 0, 0, 0, 0]); // Predicts all 0\n * const bacc = balancedAccuracyScore(yTrue, yPred); // 0.5 (not 0.8!)\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.balanced_accuracy_score.html | Scikit-learn balanced_accuracy_score}\n */\nexport function balancedAccuracyScore(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n\n  if (yTrue.size === 0) return 0;\n\n  const { classes, stats } = buildClassStats(yTrue, yPred);\n  let sumRecall = 0;\n  let classCount = 0;\n\n  for (const cls of classes) {\n    const classStats = stats.get(cls);\n    const support = classStats?.support ?? 0;\n    if (support === 0) continue;\n    const tp = classStats?.tp ?? 0;\n    const fn = classStats?.fn ?? 0;\n    const recall = tp + fn > 0 ? tp / (tp + fn) : 0;\n    sumRecall += recall;\n    classCount++;\n  }\n\n  return classCount === 0 ? 0 : sumRecall / classCount;\n}\n","import {\n  getNumericElement,\n  isNumericTypedArray,\n  isTypedArray,\n  type NumericTypedArray,\n} from \"../core\";\nimport { DataValidationError, DTypeError, InvalidParameterError, ShapeError } from \"../core/errors\";\nimport { type Tensor, tensor } from \"../ndarray\";\nimport {\n  assertFiniteNumber,\n  assertSameSize,\n  computeLogicalStrides,\n  createFlatOffsetter,\n  type FlatOffsetter,\n} from \"./_internal\";\n\nfunction isBigIntTypedArray(x: unknown): x is BigInt64Array | BigUint64Array {\n  return x instanceof BigInt64Array || x instanceof BigUint64Array;\n}\n\nfunction readIndex<T>(arr: ArrayLike<T>, index: number, name: string): T {\n  const v = arr[index];\n  if (v === undefined) {\n    throw new DataValidationError(`${name} index out of range: ${index}`);\n  }\n  return v;\n}\n\nfunction getNumericTensorData(t: Tensor, name: string): NumericTypedArray {\n  if (t.dtype === \"string\") {\n    throw new DTypeError(`${name} must be numeric (string tensors not supported)`);\n  }\n  if (t.dtype === \"int64\") {\n    throw new DTypeError(`${name} must be numeric (int64 tensors not supported)`);\n  }\n  const data = t.data;\n  if (!isTypedArray(data) || !isNumericTypedArray(data)) {\n    throw new DTypeError(`${name} must be a numeric tensor`);\n  }\n  return data;\n}\n\nfunction getFeatureAccessor(X: Tensor) {\n  const data = getNumericTensorData(X, \"X\");\n\n  if (X.ndim === 0 || X.ndim > 2) {\n    throw new ShapeError(\"X must be a 1D or 2D tensor\");\n  }\n\n  const nSamples = X.shape[0] ?? 0;\n  const nFeatures = X.ndim === 1 ? 1 : (X.shape[1] ?? 0);\n  if (nFeatures === 0) {\n    throw new ShapeError(\"X must have at least one feature\");\n  }\n\n  const logicalStrides = computeLogicalStrides(X.shape);\n  const logical0 = logicalStrides[0];\n  if (logical0 === undefined) throw new ShapeError(\"Invalid logical strides\");\n\n  const sampleStride = X.strides[0] ?? logical0 ?? nFeatures;\n\n  const featureStride = X.ndim === 1 ? 0 : (X.strides[1] ?? logicalStrides[1] ?? 1);\n\n  if (X.ndim === 2 && featureStride === 0) {\n    throw new ShapeError(\"X must have a non-degenerate feature stride\");\n  }\n\n  return {\n    data,\n    nSamples,\n    nFeatures,\n    sampleStride,\n    featureStride,\n    offset: X.offset,\n  };\n}\n\ntype EncodedLabels = {\n  readonly codes: Int32Array;\n  readonly nClusters: number;\n};\n\nfunction readLabelValue(\n  labels: Tensor,\n  offsetter: FlatOffsetter,\n  index: number,\n  name: string\n): string | number | boolean | bigint {\n  const data = labels.data;\n  const flat = offsetter(index);\n\n  if (isBigIntTypedArray(data)) {\n    const v = data[flat];\n    if (v === undefined) {\n      throw new DataValidationError(`${name} must contain a value for index ${index}`);\n    }\n    return v;\n  }\n\n  if (isTypedArray(data) && isNumericTypedArray(data)) {\n    const v = getNumericElement(data, flat);\n    assertFiniteNumber(v, name, `index ${index}`);\n\n    if ((labels.dtype === \"float32\" || labels.dtype === \"float64\") && !Number.isInteger(v)) {\n      throw new DataValidationError(\n        `${name} must contain discrete labels; found non-integer ${String(v)} at index ${index}`\n      );\n    }\n\n    if (labels.dtype === \"int64\" && (!Number.isInteger(v) || !Number.isSafeInteger(v))) {\n      throw new DataValidationError(\n        `${name} contains an int64 value that cannot be represented safely as a number at index ${index}`\n      );\n    }\n\n    return v;\n  }\n\n  if (Array.isArray(data)) {\n    const v = data[flat];\n    if (v === undefined || v === null) {\n      throw new DataValidationError(`${name} must contain a value for index ${index}`);\n    }\n\n    if (typeof v === \"string\") return v;\n    if (typeof v === \"boolean\") return v;\n\n    if (typeof v === \"number\") {\n      assertFiniteNumber(v, name, `index ${index}`);\n\n      if ((labels.dtype === \"float32\" || labels.dtype === \"float64\") && !Number.isInteger(v)) {\n        throw new DataValidationError(\n          `${name} must contain discrete labels; found non-integer ${String(v)} at index ${index}`\n        );\n      }\n\n      if (labels.dtype === \"int64\" && (!Number.isInteger(v) || !Number.isSafeInteger(v))) {\n        throw new DataValidationError(\n          `${name} contains an int64 value that cannot be represented safely as a number at index ${index}`\n        );\n      }\n\n      return v;\n    }\n\n    throw new DTypeError(`${name} must contain primitive labels (string, boolean, number, bigint)`);\n  }\n\n  throw new DTypeError(`${name} has unsupported backing storage for labels`);\n}\n\nfunction encodeLabels(labels: Tensor, name: string): EncodedLabels {\n  if (labels.dtype === \"string\") {\n    throw new DTypeError(`${name}: string labels not supported for clustering metrics`);\n  }\n\n  const n = labels.size;\n  const codes = new Int32Array(n);\n  const offsetter = createFlatOffsetter(labels);\n\n  const map = new Map<string | number | boolean | bigint, number>();\n  let next = 0;\n\n  for (let i = 0; i < n; i++) {\n    const raw = readLabelValue(labels, offsetter, i, name);\n\n    const existing = map.get(raw);\n    if (existing === undefined) {\n      map.set(raw, next);\n      codes[i] = next;\n      next++;\n    } else {\n      codes[i] = existing;\n    }\n  }\n\n  return { codes, nClusters: next };\n}\n\nfunction comb2(x: number): number {\n  return x <= 1 ? 0 : (x * (x - 1)) / 2;\n}\n\ntype ContingencyStats = {\n  readonly contingencyDense: Int32Array | null;\n  readonly contingencySparse: Map<number, number> | null;\n  readonly trueCount: Int32Array;\n  readonly predCount: Int32Array;\n  readonly nTrue: number;\n  readonly nPred: number;\n  readonly n: number;\n};\n\nfunction buildContingencyStats(labelsTrue: Tensor, labelsPred: Tensor): ContingencyStats {\n  assertSameSize(labelsTrue, labelsPred, \"labelsTrue\", \"labelsPred\");\n\n  const n = labelsTrue.size;\n  const encT = encodeLabels(labelsTrue, \"labelsTrue\");\n  const encP = encodeLabels(labelsPred, \"labelsPred\");\n\n  const trueCodes = encT.codes;\n  const predCodes = encP.codes;\n  const nTrue = encT.nClusters;\n  const nPred = encP.nClusters;\n\n  const trueCount = new Int32Array(nTrue);\n  const predCount = new Int32Array(nPred);\n\n  for (let i = 0; i < n; i++) {\n    const t = readIndex(trueCodes, i, \"trueCodes\");\n    const p = readIndex(predCodes, i, \"predCodes\");\n    trueCount[t] = (trueCount[t] ?? 0) + 1;\n    predCount[p] = (predCount[p] ?? 0) + 1;\n  }\n\n  const denseSize = nTrue * nPred;\n  const maxDenseCells = 4_000_000; // ~16MB Int32\n\n  if (denseSize > 0 && denseSize <= maxDenseCells) {\n    const contingency = new Int32Array(denseSize);\n    for (let i = 0; i < n; i++) {\n      const t = readIndex(trueCodes, i, \"trueCodes\");\n      const p = readIndex(predCodes, i, \"predCodes\");\n      const idx = t * nPred + p;\n      contingency[idx] = (contingency[idx] ?? 0) + 1;\n    }\n    return {\n      contingencyDense: contingency,\n      contingencySparse: null,\n      trueCount,\n      predCount,\n      nTrue,\n      nPred,\n      n,\n    };\n  }\n\n  const contingency = new Map<number, number>();\n  for (let i = 0; i < n; i++) {\n    const t = readIndex(trueCodes, i, \"trueCodes\");\n    const p = readIndex(predCodes, i, \"predCodes\");\n    const key = t * nPred + p;\n    contingency.set(key, (contingency.get(key) ?? 0) + 1);\n  }\n\n  return {\n    contingencyDense: null,\n    contingencySparse: contingency,\n    trueCount,\n    predCount,\n    nTrue,\n    nPred,\n    n,\n  };\n}\n\nfunction entropyFromCountArray(counts: Int32Array, n: number): number {\n  if (n === 0) return 0;\n  let h = 0;\n\n  for (let i = 0; i < counts.length; i++) {\n    const c = readIndex(counts, i, \"counts\");\n    if (c > 0) {\n      const p = c / n;\n      h -= p * Math.log(p);\n    }\n  }\n  return h;\n}\n\nfunction mutualInformationFromContingency(stats: ContingencyStats): number {\n  const { contingencyDense, contingencySparse, trueCount, predCount, nPred, n } = stats;\n  if (n === 0) return 0;\n\n  let mi = 0;\n\n  if (contingencyDense) {\n    for (let idx = 0; idx < contingencyDense.length; idx++) {\n      const nij = readIndex(contingencyDense, idx, \"contingencyDense\");\n      if (nij <= 0) continue;\n\n      const t = Math.floor(idx / nPred);\n      const p = idx - t * nPred;\n\n      const ni = readIndex(trueCount, t, \"trueCount\");\n      const nj = readIndex(predCount, p, \"predCount\");\n      if (ni > 0 && nj > 0) {\n        mi += (nij / n) * Math.log((n * nij) / (ni * nj));\n      }\n    }\n    return mi;\n  }\n\n  if (contingencySparse) {\n    for (const [key, nij] of contingencySparse) {\n      if (nij <= 0) continue;\n\n      const t = Math.floor(key / nPred);\n      const p = key - t * nPred;\n\n      const ni = readIndex(trueCount, t, \"trueCount\");\n      const nj = readIndex(predCount, p, \"predCount\");\n      if (ni > 0 && nj > 0) {\n        mi += (nij / n) * Math.log((n * nij) / (ni * nj));\n      }\n    }\n    return mi;\n  }\n\n  return 0;\n}\n\nfunction buildLogFactorials(n: number): Float64Array {\n  const out = new Float64Array(n + 1);\n  for (let i = 1; i <= n; i++) {\n    out[i] = (out[i - 1] ?? 0) + Math.log(i);\n  }\n  return out;\n}\n\nfunction logCombination(n: number, k: number, logFactorials: Float64Array): number {\n  if (k < 0 || k > n) return Number.NEGATIVE_INFINITY;\n  const a = readIndex(logFactorials, n, \"logFactorials\");\n  const b = readIndex(logFactorials, k, \"logFactorials\");\n  const c = readIndex(logFactorials, n - k, \"logFactorials\");\n  return a - b - c;\n}\n\nconst LOG_EXP_UNDERFLOW_CUTOFF = -745;\n\nfunction expectedMutualInformation(stats: ContingencyStats): number {\n  const { trueCount, predCount, n } = stats;\n  if (n <= 1) return 0;\n\n  const rowSums = Array.from(trueCount);\n  const colSums = Array.from(predCount);\n  const logFactorials = buildLogFactorials(n);\n\n  let emi = 0;\n  let comp = 0;\n\n  for (const a of rowSums) {\n    if (a <= 0) continue;\n    for (const b of colSums) {\n      if (b <= 0) continue;\n\n      const nijMin = Math.max(1, a + b - n);\n      const nijMax = Math.min(a, b);\n      if (nijMin > nijMax) continue;\n\n      const logDenominator = logCombination(n, b, logFactorials);\n\n      for (let nij = nijMin; nij <= nijMax; nij++) {\n        const logProbability =\n          logCombination(a, nij, logFactorials) +\n          logCombination(n - a, b - nij, logFactorials) -\n          logDenominator;\n\n        if (logProbability < LOG_EXP_UNDERFLOW_CUTOFF) continue;\n\n        const probability = Math.exp(logProbability);\n        if (!Number.isFinite(probability) || probability === 0) continue;\n\n        const miTerm = (nij / n) * Math.log((n * nij) / (a * b));\n        const y = probability * miTerm - comp;\n        const t = emi + y;\n        comp = t - emi - y;\n        emi = t;\n      }\n    }\n  }\n\n  return emi;\n}\n\ntype AverageMethod = \"min\" | \"geometric\" | \"arithmetic\" | \"max\";\n\nfunction averageEntropy(hTrue: number, hPred: number, method: AverageMethod): number {\n  if (method === \"min\") return Math.min(hTrue, hPred);\n  if (method === \"max\") return Math.max(hTrue, hPred);\n  if (method === \"geometric\") return Math.sqrt(hTrue * hPred);\n  return (hTrue + hPred) / 2;\n}\n\nfunction euclideanDistance(\n  data: NumericTypedArray,\n  offset: number,\n  sampleStride: number,\n  featureStride: number,\n  nFeatures: number,\n  i: number,\n  j: number\n): number {\n  let sum = 0;\n  const baseI = offset + i * sampleStride;\n  const baseJ = offset + j * sampleStride;\n\n  for (let k = 0; k < nFeatures; k++) {\n    const vi = getNumericElement(data, baseI + k * featureStride);\n    const vj = getNumericElement(data, baseJ + k * featureStride);\n    assertFiniteNumber(vi, \"X\", `sample ${i}, feature ${k}`);\n    assertFiniteNumber(vj, \"X\", `sample ${j}, feature ${k}`);\n    const d = vi - vj;\n    sum += d * d;\n  }\n  return Math.sqrt(sum);\n}\n\ntype SilhouetteMetric = \"euclidean\" | \"precomputed\";\n\nfunction getPrecomputedDistanceAccessor(X: Tensor, n: number) {\n  const data = getNumericTensorData(X, \"X\");\n\n  if (X.ndim !== 2) {\n    throw new ShapeError(\"X must be a 2D tensor for metric='precomputed'\");\n  }\n  const rows = X.shape[0] ?? 0;\n  const cols = X.shape[1] ?? 0;\n  if (rows !== n || cols !== n) {\n    throw new ShapeError(\n      \"For metric='precomputed', X must be a square [n_samples, n_samples] matrix\"\n    );\n  }\n\n  const logicalStrides = computeLogicalStrides(X.shape);\n  const rowStride = X.strides[0] ?? logicalStrides[0] ?? n;\n  const colStride = X.strides[1] ?? logicalStrides[1] ?? 1;\n\n  if (rowStride === 0 || colStride === 0) {\n    throw new ShapeError(\"Precomputed distance matrix must have non-degenerate strides\");\n  }\n\n  const base = X.offset;\n\n  return (i: number, j: number) => {\n    const v = getNumericElement(data, base + i * rowStride + j * colStride);\n    assertFiniteNumber(v, \"X\", `distance[${i},${j}]`);\n    if (v < 0) {\n      throw new DataValidationError(\n        `Precomputed distances must be non-negative; found ${String(v)} at [${i},${j}]`\n      );\n    }\n    return v;\n  };\n}\n\nfunction validateSilhouetteLabels(labels: Tensor, nSamples: number): EncodedLabels {\n  if (labels.size !== nSamples) {\n    throw new ShapeError(\"labels length must match number of samples\");\n  }\n\n  const enc = encodeLabels(labels, \"labels\");\n  const k = enc.nClusters;\n\n  if (k < 2 || k > nSamples - 1) {\n    throw new InvalidParameterError(\n      \"silhouette requires 2 <= n_clusters <= n_samples - 1\",\n      \"n_clusters\",\n      k\n    );\n  }\n\n  return enc;\n}\n\nfunction reservoirSampleIndices(n: number, k: number, seed: number | undefined): Int32Array {\n  let state = (seed ?? 0) >>> 0;\n  const hasSeed = seed !== undefined;\n\n  const randU32 = () => {\n    if (!hasSeed) return (Math.random() * 0x1_0000_0000) >>> 0;\n    state = (1664525 * state + 1013904223) >>> 0;\n    return state;\n  };\n\n  const randInt = (exclusiveMax: number) => randU32() % exclusiveMax;\n\n  const out = new Int32Array(k);\n  for (let i = 0; i < k; i++) out[i] = i;\n\n  for (let i = k; i < n; i++) {\n    const j = randInt(i + 1);\n    if (j < k) out[j] = i;\n  }\n  return out;\n}\n\nfunction reencodeSubset(codes: Int32Array): {\n  codes: Int32Array;\n  nClusters: number;\n} {\n  const out = new Int32Array(codes.length);\n  const map = new Map<number, number>();\n  let next = 0;\n\n  for (let i = 0; i < codes.length; i++) {\n    const v = readIndex(codes, i, \"codes\");\n    const existing = map.get(v);\n    if (existing === undefined) {\n      map.set(v, next);\n      out[i] = next;\n      next++;\n    } else {\n      out[i] = existing;\n    }\n  }\n\n  return { codes: out, nClusters: next };\n}\n\nfunction silhouetteMeanEuclidean(X: Tensor, labels: Tensor, indices: Int32Array | null): number {\n  const { data, nSamples, nFeatures, sampleStride, featureStride, offset } = getFeatureAccessor(X);\n\n  const n = indices ? indices.length : nSamples;\n  if (nSamples < 2 || n < 2) {\n    throw new InvalidParameterError(\"silhouette requires at least 2 samples\", \"n_samples\", n);\n  }\n\n  const encAll = validateSilhouetteLabels(labels, nSamples);\n\n  const subsetCodesRaw = new Int32Array(n);\n  for (let i = 0; i < n; i++) {\n    const src = indices ? readIndex(indices, i, \"indices\") : i;\n    const c = readIndex(encAll.codes, src, \"labels.codes\");\n    subsetCodesRaw[i] = c;\n  }\n\n  const re = reencodeSubset(subsetCodesRaw);\n  const codes = re.codes;\n  const k = re.nClusters;\n\n  if (k < 2 || k > n - 1) {\n    throw new InvalidParameterError(\n      \"silhouette requires 2 <= n_clusters <= n_samples - 1\",\n      \"n_clusters\",\n      k\n    );\n  }\n\n  const clusterSizes = new Int32Array(k);\n  for (let i = 0; i < n; i++) {\n    const ci = readIndex(codes, i, \"codes\");\n    clusterSizes[ci] = (clusterSizes[ci] ?? 0) + 1;\n  }\n\n  const sumsToClusters = new Float64Array(k);\n\n  let sum = 0;\n  let comp = 0;\n\n  for (let i = 0; i < n; i++) {\n    const ci = readIndex(codes, i, \"codes\");\n    const sizeOwn = readIndex(clusterSizes, ci, \"clusterSizes\");\n\n    if (sizeOwn <= 1) continue; // singleton => 0\n\n    sumsToClusters.fill(0);\n\n    const srcI = indices ? readIndex(indices, i, \"indices\") : i;\n\n    for (let j = 0; j < n; j++) {\n      if (i === j) continue;\n      const srcJ = indices ? readIndex(indices, j, \"indices\") : j;\n      const cj = readIndex(codes, j, \"codes\");\n\n      const d = euclideanDistance(data, offset, sampleStride, featureStride, nFeatures, srcI, srcJ);\n      sumsToClusters[cj] = (sumsToClusters[cj] ?? 0) + d;\n    }\n\n    const a = readIndex(sumsToClusters, ci, \"sumsToClusters\") / (sizeOwn - 1);\n\n    let b = Infinity;\n    for (let cl = 0; cl < k; cl++) {\n      if (cl === ci) continue;\n      const sz = readIndex(clusterSizes, cl, \"clusterSizes\");\n      if (sz <= 0) continue;\n      const mean = readIndex(sumsToClusters, cl, \"sumsToClusters\") / sz;\n      if (mean < b) b = mean;\n    }\n\n    if (!Number.isFinite(b) || b === Infinity) continue;\n\n    const denom = Math.max(a, b);\n    const s = denom > 0 ? (b - a) / denom : 0;\n\n    const y = s - comp;\n    const t = sum + y;\n    comp = t - sum - y;\n    sum = t;\n  }\n\n  return sum / n;\n}\n\nfunction silhouetteMeanPrecomputed(X: Tensor, labels: Tensor, indices: Int32Array | null): number {\n  const nSamples = labels.size;\n  if (nSamples < 2) {\n    throw new InvalidParameterError(\n      \"silhouette requires at least 2 samples\",\n      \"n_samples\",\n      nSamples\n    );\n  }\n\n  const encAll = validateSilhouetteLabels(labels, nSamples);\n  const dist = getPrecomputedDistanceAccessor(X, nSamples);\n\n  const n = indices ? indices.length : nSamples;\n  if (n < 2) {\n    throw new InvalidParameterError(\"silhouette requires at least 2 samples\", \"n_samples\", n);\n  }\n\n  // Validate diagonal ~ 0 for relevant indices\n  for (let i = 0; i < n; i++) {\n    const src = indices ? readIndex(indices, i, \"indices\") : i;\n    const d0 = dist(src, src);\n    if (!Number.isFinite(d0) || Math.abs(d0) > 1e-12) {\n      throw new DataValidationError(\n        `Precomputed distance matrix diagonal must be ~0; found ${String(d0)} at [${src},${src}]`\n      );\n    }\n  }\n\n  const subsetCodesRaw = new Int32Array(n);\n  for (let i = 0; i < n; i++) {\n    const src = indices ? readIndex(indices, i, \"indices\") : i;\n    const c = readIndex(encAll.codes, src, \"labels.codes\");\n    subsetCodesRaw[i] = c;\n  }\n\n  const re = reencodeSubset(subsetCodesRaw);\n  const codes = re.codes;\n  const k = re.nClusters;\n\n  if (k < 2 || k > n - 1) {\n    throw new InvalidParameterError(\n      \"silhouette requires 2 <= n_clusters <= n_samples - 1\",\n      \"n_clusters\",\n      k\n    );\n  }\n\n  const clusterSizes = new Int32Array(k);\n  for (let i = 0; i < n; i++) {\n    const ci = readIndex(codes, i, \"codes\");\n    clusterSizes[ci] = (clusterSizes[ci] ?? 0) + 1;\n  }\n\n  const sumsToClusters = new Float64Array(k);\n\n  let sum = 0;\n  let comp = 0;\n\n  for (let i = 0; i < n; i++) {\n    const ci = readIndex(codes, i, \"codes\");\n    const sizeOwn = readIndex(clusterSizes, ci, \"clusterSizes\");\n    if (sizeOwn <= 1) continue;\n\n    sumsToClusters.fill(0);\n\n    const srcI = indices ? readIndex(indices, i, \"indices\") : i;\n\n    for (let j = 0; j < n; j++) {\n      if (i === j) continue;\n\n      const srcJ = indices ? readIndex(indices, j, \"indices\") : j;\n      const cj = readIndex(codes, j, \"codes\");\n\n      const d = dist(srcI, srcJ);\n      sumsToClusters[cj] = (sumsToClusters[cj] ?? 0) + d;\n    }\n\n    const a = readIndex(sumsToClusters, ci, \"sumsToClusters\") / (sizeOwn - 1);\n\n    let b = Infinity;\n    for (let cl = 0; cl < k; cl++) {\n      if (cl === ci) continue;\n      const sz = readIndex(clusterSizes, cl, \"clusterSizes\");\n      if (sz <= 0) continue;\n      const mean = readIndex(sumsToClusters, cl, \"sumsToClusters\") / sz;\n      if (mean < b) b = mean;\n    }\n\n    if (!Number.isFinite(b) || b === Infinity) continue;\n\n    const denom = Math.max(a, b);\n    const s = denom > 0 ? (b - a) / denom : 0;\n\n    const y = s - comp;\n    const t = sum + y;\n    comp = t - sum - y;\n    sum = t;\n  }\n\n  return sum / n;\n}\n\n/**\n * Computes the mean Silhouette Coefficient over all samples.\n *\n * The Silhouette Coefficient for a sample measures how similar it is to its own\n * cluster compared to other clusters. Values range from -1 to 1, where higher\n * values indicate better-defined clusters.\n *\n * **Formula**: s(i) = (b(i) - a(i)) / max(a(i), b(i))\n * - a(i): mean intra-cluster distance for sample i\n * - b(i): mean nearest-cluster distance for sample i\n *\n * **Time Complexity**: O(n²) where n is the number of samples\n * **Space Complexity**: O(n + k) where k is the number of clusters\n *\n * @param X - Feature matrix of shape [n_samples, n_features], or a precomputed distance matrix\n * @param labels - Cluster labels for each sample\n * @param metric - Distance metric: 'euclidean' (default) or 'precomputed'\n * @param options - Optional parameters\n * @param options.sampleSize - Number of samples to use for approximation (required when n > 2000)\n * @param options.randomState - Seed for reproducible sampling\n * @returns Mean silhouette coefficient in range [-1, 1]\n *\n * @throws {InvalidParameterError} If fewer than 2 samples, invalid sampleSize, or unsupported metric\n * @throws {ShapeError} If labels length doesn't match samples, or X shape is invalid\n * @throws {DTypeError} If labels are string or X is non-numeric\n * @throws {DataValidationError} If X contains non-finite values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.silhouette_score.html | Scikit-learn silhouette_score}\n */\nexport function silhouetteScore(\n  X: Tensor,\n  labels: Tensor,\n  metric: SilhouetteMetric = \"euclidean\",\n  options?: { sampleSize?: number; randomState?: number }\n): number {\n  const sampleSize = options?.sampleSize;\n  const randomState = options?.randomState;\n\n  const nSamples = labels.size;\n  if (nSamples < 2) {\n    throw new InvalidParameterError(\n      \"silhouette requires at least 2 samples\",\n      \"n_samples\",\n      nSamples\n    );\n  }\n\n  const maxFull = 2000;\n  if (sampleSize === undefined && nSamples > maxFull) {\n    throw new InvalidParameterError(\n      `silhouetteScore is O(n²) and n_samples=${nSamples} is too large for full computation; provide options.sampleSize`,\n      \"sampleSize\",\n      sampleSize\n    );\n  }\n\n  if (sampleSize !== undefined) {\n    if (!Number.isFinite(sampleSize) || !Number.isInteger(sampleSize)) {\n      throw new InvalidParameterError(\"sampleSize must be an integer\", \"sampleSize\", sampleSize);\n    }\n    if (sampleSize < 2 || sampleSize > nSamples) {\n      throw new InvalidParameterError(\n        \"sampleSize must satisfy 2 <= sampleSize <= n_samples\",\n        \"sampleSize\",\n        sampleSize\n      );\n    }\n  }\n\n  const indices =\n    sampleSize !== undefined && sampleSize < nSamples\n      ? reservoirSampleIndices(nSamples, sampleSize, randomState)\n      : null;\n\n  if (metric === \"euclidean\") return silhouetteMeanEuclidean(X, labels, indices);\n  if (metric === \"precomputed\") return silhouetteMeanPrecomputed(X, labels, indices);\n\n  throw new InvalidParameterError(\n    `Unsupported metric: '${String(metric)}'. Must be 'euclidean' or 'precomputed'`,\n    \"metric\",\n    metric\n  );\n}\n\n/**\n * Computes the Silhouette Coefficient for each sample.\n *\n * Returns a tensor of per-sample silhouette values. Useful for identifying\n * samples that are well-clustered vs. poorly-clustered.\n *\n * **Time Complexity**: O(n²) where n is the number of samples\n * **Space Complexity**: O(n + k) where k is the number of clusters\n *\n * @param X - Feature matrix of shape [n_samples, n_features], or a precomputed distance matrix\n * @param labels - Cluster labels for each sample\n * @param metric - Distance metric: 'euclidean' (default) or 'precomputed'\n * @returns Tensor of silhouette coefficients in range [-1, 1] for each sample\n *\n * @throws {InvalidParameterError} If fewer than 2 samples or unsupported metric\n * @throws {ShapeError} If labels length doesn't match samples, or X shape is invalid\n * @throws {DTypeError} If labels are string or X is non-numeric\n * @throws {DataValidationError} If X contains non-finite values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.silhouette_samples.html | Scikit-learn silhouette_samples}\n */\nexport function silhouetteSamples(\n  X: Tensor,\n  labels: Tensor,\n  metric: SilhouetteMetric = \"euclidean\"\n): Tensor {\n  const nSamples = labels.size;\n  if (nSamples < 2) {\n    throw new InvalidParameterError(\n      \"silhouette requires at least 2 samples\",\n      \"n_samples\",\n      nSamples\n    );\n  }\n\n  const enc = validateSilhouetteLabels(labels, nSamples);\n  const codes = enc.codes;\n  const k = enc.nClusters;\n\n  const silhouettes = new Float64Array(nSamples);\n\n  const clusterSizes = new Int32Array(k);\n  for (let i = 0; i < nSamples; i++) {\n    const ci = readIndex(codes, i, \"codes\");\n    clusterSizes[ci] = (clusterSizes[ci] ?? 0) + 1;\n  }\n\n  const sumsToClusters = new Float64Array(k);\n\n  if (metric === \"euclidean\") {\n    const { data, nFeatures, sampleStride, featureStride, offset } = getFeatureAccessor(X);\n\n    for (let i = 0; i < nSamples; i++) {\n      const ci = readIndex(codes, i, \"codes\");\n      const sizeOwn = readIndex(clusterSizes, ci, \"clusterSizes\");\n      if (sizeOwn <= 1) {\n        silhouettes[i] = 0;\n        continue;\n      }\n\n      sumsToClusters.fill(0);\n\n      for (let j = 0; j < nSamples; j++) {\n        if (i === j) continue;\n        const cj = readIndex(codes, j, \"codes\");\n\n        const d = euclideanDistance(data, offset, sampleStride, featureStride, nFeatures, i, j);\n        sumsToClusters[cj] = (sumsToClusters[cj] ?? 0) + d;\n      }\n\n      const a = readIndex(sumsToClusters, ci, \"sumsToClusters\") / (sizeOwn - 1);\n\n      let b = Infinity;\n      for (let cl = 0; cl < k; cl++) {\n        if (cl === ci) continue;\n        const sz = readIndex(clusterSizes, cl, \"clusterSizes\");\n        if (sz <= 0) continue;\n        const mean = readIndex(sumsToClusters, cl, \"sumsToClusters\") / sz;\n        if (mean < b) b = mean;\n      }\n\n      if (!Number.isFinite(b) || b === Infinity) {\n        silhouettes[i] = 0;\n        continue;\n      }\n\n      const denom = Math.max(a, b);\n      silhouettes[i] = denom > 0 ? (b - a) / denom : 0;\n    }\n\n    return tensor(silhouettes);\n  }\n\n  if (metric === \"precomputed\") {\n    const dist = getPrecomputedDistanceAccessor(X, nSamples);\n\n    for (let i = 0; i < nSamples; i++) {\n      const d0 = dist(i, i);\n      if (!Number.isFinite(d0) || Math.abs(d0) > 1e-12) {\n        throw new DataValidationError(\n          `Precomputed distance matrix diagonal must be ~0; found ${String(d0)} at [${i},${i}]`\n        );\n      }\n    }\n\n    for (let i = 0; i < nSamples; i++) {\n      const ci = readIndex(codes, i, \"codes\");\n      const sizeOwn = readIndex(clusterSizes, ci, \"clusterSizes\");\n      if (sizeOwn <= 1) {\n        silhouettes[i] = 0;\n        continue;\n      }\n\n      sumsToClusters.fill(0);\n\n      for (let j = 0; j < nSamples; j++) {\n        if (i === j) continue;\n        const cj = readIndex(codes, j, \"codes\");\n\n        const d = dist(i, j);\n        sumsToClusters[cj] = (sumsToClusters[cj] ?? 0) + d;\n      }\n\n      const a = readIndex(sumsToClusters, ci, \"sumsToClusters\") / (sizeOwn - 1);\n\n      let b = Infinity;\n      for (let cl = 0; cl < k; cl++) {\n        if (cl === ci) continue;\n        const sz = readIndex(clusterSizes, cl, \"clusterSizes\");\n        if (sz <= 0) continue;\n        const mean = readIndex(sumsToClusters, cl, \"sumsToClusters\") / sz;\n        if (mean < b) b = mean;\n      }\n\n      if (!Number.isFinite(b) || b === Infinity) {\n        silhouettes[i] = 0;\n        continue;\n      }\n\n      const denom = Math.max(a, b);\n      silhouettes[i] = denom > 0 ? (b - a) / denom : 0;\n    }\n\n    return tensor(silhouettes);\n  }\n\n  throw new InvalidParameterError(\n    `Unsupported metric: '${String(metric)}'. Must be 'euclidean' or 'precomputed'`,\n    \"metric\",\n    metric\n  );\n}\n\n/**\n * Computes the Davies-Bouldin index.\n *\n * The Davies-Bouldin index measures the average similarity ratio of each cluster\n * with its most similar cluster. Lower values indicate better clustering.\n * Returns 0 when k < 2 or n_samples == 0.\n *\n * **Time Complexity**: O(n * d + k² * d) where n is samples, d is features, k is clusters\n * **Space Complexity**: O(k * d)\n *\n * @param X - Feature matrix of shape [n_samples, n_features]\n * @param labels - Cluster labels for each sample\n * @returns Davies-Bouldin index (lower is better, 0 is minimum)\n *\n * @throws {ShapeError} If labels length doesn't match samples, or X shape is invalid\n * @throws {DTypeError} If labels are string or X is non-numeric\n * @throws {DataValidationError} If X contains non-finite values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.davies_bouldin_score.html | Scikit-learn davies_bouldin_score}\n */\nexport function daviesBouldinScore(X: Tensor, labels: Tensor): number {\n  const { data, nSamples, nFeatures, sampleStride, featureStride, offset } = getFeatureAccessor(X);\n  if (nSamples === 0) return 0;\n\n  if (labels.size !== nSamples) {\n    throw new ShapeError(\"labels length must match number of samples\");\n  }\n\n  const enc = encodeLabels(labels, \"labels\");\n  const codes = enc.codes;\n  const k = enc.nClusters;\n  if (k < 2) return 0;\n\n  const centroids = new Array<Float64Array>(k);\n  for (let c = 0; c < k; c++) centroids[c] = new Float64Array(nFeatures);\n\n  const clusterSizes = new Int32Array(k);\n\n  for (let i = 0; i < nSamples; i++) {\n    const c = readIndex(codes, i, \"codes\");\n    clusterSizes[c] = (clusterSizes[c] ?? 0) + 1;\n\n    const base = offset + i * sampleStride;\n    const centroid = readIndex(centroids, c, \"centroids\");\n    for (let f = 0; f < nFeatures; f++) {\n      const v = getNumericElement(data, base + f * featureStride);\n      assertFiniteNumber(v, \"X\", `sample ${i}, feature ${f}`);\n      centroid[f] = (centroid[f] ?? 0) + v;\n    }\n  }\n\n  for (let c = 0; c < k; c++) {\n    const sz = readIndex(clusterSizes, c, \"clusterSizes\");\n    if (sz <= 0) continue;\n    const centroid = readIndex(centroids, c, \"centroids\");\n    for (let f = 0; f < nFeatures; f++) {\n      centroid[f] = (centroid[f] ?? 0) / sz;\n    }\n  }\n\n  const scatterSum = new Float64Array(k);\n\n  for (let i = 0; i < nSamples; i++) {\n    const c = readIndex(codes, i, \"codes\");\n    const centroid = readIndex(centroids, c, \"centroids\");\n\n    let distSq = 0;\n    const base = offset + i * sampleStride;\n    for (let f = 0; f < nFeatures; f++) {\n      const v = getNumericElement(data, base + f * featureStride);\n      assertFiniteNumber(v, \"X\", `sample ${i}, feature ${f}`);\n      const d = v - (centroid[f] ?? 0);\n      distSq += d * d;\n    }\n\n    scatterSum[c] = (scatterSum[c] ?? 0) + Math.sqrt(distSq);\n  }\n\n  const S = new Float64Array(k);\n  for (let c = 0; c < k; c++) {\n    const sz = readIndex(clusterSizes, c, \"clusterSizes\");\n    const sc = readIndex(scatterSum, c, \"scatterSum\");\n    S[c] = sz > 0 ? sc / sz : 0;\n  }\n\n  let db = 0;\n\n  for (let i = 0; i < k; i++) {\n    let maxRatio = Number.NEGATIVE_INFINITY;\n    const ci = readIndex(centroids, i, \"centroids\");\n\n    for (let j = 0; j < k; j++) {\n      if (i === j) continue;\n      const cj = readIndex(centroids, j, \"centroids\");\n\n      let distSq = 0;\n      for (let f = 0; f < nFeatures; f++) {\n        const d = (ci[f] ?? 0) - (cj[f] ?? 0);\n        distSq += d * d;\n      }\n\n      const dist = Math.sqrt(distSq);\n      const si = readIndex(S, i, \"S\");\n      const sj = readIndex(S, j, \"S\");\n      const ratio = dist === 0 ? Number.POSITIVE_INFINITY : (si + sj) / dist;\n      if (ratio > maxRatio) maxRatio = ratio;\n    }\n\n    db += maxRatio;\n  }\n\n  return db / k;\n}\n\n/**\n * Computes the Calinski-Harabasz index (Variance Ratio Criterion).\n *\n * The score is the ratio of between-cluster dispersion to within-cluster\n * dispersion. Higher values indicate better-defined clusters.\n * Returns 0 when k < 2, n_samples == 0, or within-group sum of squares is 0.\n *\n * **Time Complexity**: O(n * d) where n is samples and d is features\n * **Space Complexity**: O(k * d) where k is clusters\n *\n * @param X - Feature matrix of shape [n_samples, n_features]\n * @param labels - Cluster labels for each sample\n * @returns Calinski-Harabasz index (higher is better)\n *\n * @throws {ShapeError} If labels length doesn't match samples, or X shape is invalid\n * @throws {DTypeError} If labels are string or X is non-numeric\n * @throws {DataValidationError} If X contains non-finite values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.calinski_harabasz_score.html | Scikit-learn calinski_harabasz_score}\n */\nexport function calinskiHarabaszScore(X: Tensor, labels: Tensor): number {\n  const { data, nSamples, nFeatures, sampleStride, featureStride, offset } = getFeatureAccessor(X);\n  if (nSamples === 0) return 0;\n\n  if (labels.size !== nSamples) {\n    throw new ShapeError(\"labels length must match number of samples\");\n  }\n\n  const enc = encodeLabels(labels, \"labels\");\n  const codes = enc.codes;\n  const k = enc.nClusters;\n\n  const overallMean = new Float64Array(nFeatures);\n\n  for (let i = 0; i < nSamples; i++) {\n    const base = offset + i * sampleStride;\n    for (let f = 0; f < nFeatures; f++) {\n      const v = getNumericElement(data, base + f * featureStride);\n      assertFiniteNumber(v, \"X\", `sample ${i}, feature ${f}`);\n      overallMean[f] = (overallMean[f] ?? 0) + v;\n    }\n  }\n  for (let f = 0; f < nFeatures; f++) {\n    overallMean[f] = (overallMean[f] ?? 0) / nSamples;\n  }\n\n  const centroids = new Array<Float64Array>(k);\n  for (let c = 0; c < k; c++) centroids[c] = new Float64Array(nFeatures);\n  const clusterSizes = new Int32Array(k);\n\n  for (let i = 0; i < nSamples; i++) {\n    const c = readIndex(codes, i, \"codes\");\n    clusterSizes[c] = (clusterSizes[c] ?? 0) + 1;\n\n    const base = offset + i * sampleStride;\n    const centroid = readIndex(centroids, c, \"centroids\");\n    for (let f = 0; f < nFeatures; f++) {\n      const v = getNumericElement(data, base + f * featureStride);\n      assertFiniteNumber(v, \"X\", `sample ${i}, feature ${f}`);\n      centroid[f] = (centroid[f] ?? 0) + v;\n    }\n  }\n\n  for (let c = 0; c < k; c++) {\n    const sz = readIndex(clusterSizes, c, \"clusterSizes\");\n    if (sz <= 0) continue;\n    const centroid = readIndex(centroids, c, \"centroids\");\n    for (let f = 0; f < nFeatures; f++) {\n      centroid[f] = (centroid[f] ?? 0) / sz;\n    }\n  }\n\n  let bgss = 0;\n  for (let c = 0; c < k; c++) {\n    const sz = readIndex(clusterSizes, c, \"clusterSizes\");\n    if (sz <= 0) continue;\n\n    const centroid = readIndex(centroids, c, \"centroids\");\n    let distSq = 0;\n    for (let f = 0; f < nFeatures; f++) {\n      const d = (centroid[f] ?? 0) - (overallMean[f] ?? 0);\n      distSq += d * d;\n    }\n    bgss += sz * distSq;\n  }\n\n  let wgss = 0;\n  for (let i = 0; i < nSamples; i++) {\n    const c = readIndex(codes, i, \"codes\");\n    const centroid = readIndex(centroids, c, \"centroids\");\n\n    const base = offset + i * sampleStride;\n    for (let f = 0; f < nFeatures; f++) {\n      const v = getNumericElement(data, base + f * featureStride);\n      assertFiniteNumber(v, \"X\", `sample ${i}, feature ${f}`);\n      const d = v - (centroid[f] ?? 0);\n      wgss += d * d;\n    }\n  }\n\n  if (k < 2 || wgss === 0) return 0;\n  return bgss / (k - 1) / (wgss / (nSamples - k));\n}\n\n/**\n * Computes the Adjusted Rand Index (ARI).\n *\n * The ARI measures the similarity between two clusterings, adjusted for chance.\n * It ranges from -1 to 1, where 1 indicates perfect agreement, 0 indicates\n * random labeling, and negative values indicate worse than random.\n *\n * **Time Complexity**: O(n + k₁ * k₂) where k₁, k₂ are cluster counts\n * **Space Complexity**: O(k₁ * k₂)\n *\n * @param labelsTrue - Ground truth cluster labels\n * @param labelsPred - Predicted cluster labels\n * @returns Adjusted Rand Index in range [-1, 1]\n *\n * @throws {ShapeError} If labelsTrue and labelsPred have different sizes\n * @throws {DTypeError} If labels are string\n * @throws {DataValidationError} If labels contain non-finite or non-integer values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.adjusted_rand_score.html | Scikit-learn adjusted_rand_score}\n */\nexport function adjustedRandScore(labelsTrue: Tensor, labelsPred: Tensor): number {\n  assertSameSize(labelsTrue, labelsPred, \"labelsTrue\", \"labelsPred\");\n\n  const n = labelsTrue.size;\n  if (n <= 1) return 1;\n\n  const stats = buildContingencyStats(labelsTrue, labelsPred);\n  const { contingencyDense, contingencySparse, trueCount, predCount } = stats;\n\n  let sumComb = 0;\n\n  if (contingencyDense) {\n    for (let idx = 0; idx < contingencyDense.length; idx++) {\n      const nij = readIndex(contingencyDense, idx, \"contingencyDense\");\n      if (nij > 0) sumComb += comb2(nij);\n    }\n  } else if (contingencySparse) {\n    for (const nij of contingencySparse.values()) {\n      sumComb += comb2(nij);\n    }\n  }\n\n  let sumCombTrue = 0;\n  for (let i = 0; i < trueCount.length; i++) {\n    sumCombTrue += comb2(readIndex(trueCount, i, \"trueCount\"));\n  }\n\n  let sumCombPred = 0;\n  for (let j = 0; j < predCount.length; j++) {\n    sumCombPred += comb2(readIndex(predCount, j, \"predCount\"));\n  }\n\n  const totalPairs = comb2(n);\n  if (totalPairs === 0) return 1;\n\n  const expectedIndex = (sumCombTrue * sumCombPred) / totalPairs;\n  const maxIndex = (sumCombTrue + sumCombPred) / 2;\n\n  const denom = maxIndex - expectedIndex;\n  if (denom === 0) return 1;\n\n  return (sumComb - expectedIndex) / denom;\n}\n\n/**\n * Computes the Adjusted Mutual Information (AMI) between two clusterings.\n *\n * AMI adjusts the Mutual Information score to account for chance, providing\n * a normalized measure of agreement between two clusterings.\n *\n * **Time Complexity**: O(n + k₁ * k₂ * min(k₁, k₂))\n * **Space Complexity**: O(k₁ * k₂ + n)\n *\n * @param labelsTrue - Ground truth cluster labels\n * @param labelsPred - Predicted cluster labels\n * @param averageMethod - Method to compute the normalizer: 'min', 'geometric', 'arithmetic' (default), or 'max'\n * @returns Adjusted Mutual Information score, typically in range [0, 1]\n *\n * @throws {ShapeError} If labelsTrue and labelsPred have different sizes\n * @throws {DTypeError} If labels are string\n * @throws {DataValidationError} If labels contain non-finite or non-integer values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.adjusted_mutual_info_score.html | Scikit-learn adjusted_mutual_info_score}\n */\nexport function adjustedMutualInfoScore(\n  labelsTrue: Tensor,\n  labelsPred: Tensor,\n  averageMethod: AverageMethod = \"arithmetic\"\n): number {\n  const stats = buildContingencyStats(labelsTrue, labelsPred);\n  const { n, trueCount, predCount } = stats;\n  if (n <= 1) return 1;\n\n  const mi = mutualInformationFromContingency(stats);\n  const hTrue = entropyFromCountArray(trueCount, n);\n  const hPred = entropyFromCountArray(predCount, n);\n  const emi = expectedMutualInformation(stats);\n\n  const normalizer = averageEntropy(hTrue, hPred, averageMethod);\n  if (Math.abs(normalizer) < 1e-15) return 1;\n\n  const denom = normalizer - emi;\n  if (Math.abs(denom) < 1e-15) return 0;\n\n  const ami = (mi - emi) / denom;\n  if (!Number.isFinite(ami)) return 0;\n  if (ami > 1) return 1;\n  if (ami < -1) return -1;\n  return ami;\n}\n\n/**\n * Computes the Normalized Mutual Information (NMI) between two clusterings.\n *\n * NMI normalizes the Mutual Information score to scale between 0 and 1,\n * where 1 indicates perfect correlation between clusterings.\n *\n * **Time Complexity**: O(n + k₁ * k₂)\n * **Space Complexity**: O(k₁ * k₂)\n *\n * @param labelsTrue - Ground truth cluster labels\n * @param labelsPred - Predicted cluster labels\n * @param averageMethod - Method to compute the normalizer: 'min', 'geometric', 'arithmetic' (default), or 'max'\n * @returns Normalized Mutual Information score in range [0, 1]\n *\n * @throws {ShapeError} If labelsTrue and labelsPred have different sizes\n * @throws {DTypeError} If labels are string\n * @throws {DataValidationError} If labels contain non-finite or non-integer values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.normalized_mutual_info_score.html | Scikit-learn normalized_mutual_info_score}\n */\nexport function normalizedMutualInfoScore(\n  labelsTrue: Tensor,\n  labelsPred: Tensor,\n  averageMethod: AverageMethod = \"arithmetic\"\n): number {\n  const stats = buildContingencyStats(labelsTrue, labelsPred);\n  const { n, trueCount, predCount } = stats;\n\n  const mi = mutualInformationFromContingency(stats);\n  const ht = entropyFromCountArray(trueCount, n);\n  const hp = entropyFromCountArray(predCount, n);\n\n  if (ht === 0 || hp === 0) {\n    return ht === 0 && hp === 0 ? 1.0 : 0.0;\n  }\n\n  const normalizer = averageEntropy(ht, hp, averageMethod);\n  if (normalizer === 0) return 0;\n\n  const nmi = mi / normalizer;\n  if (nmi > 1) return 1;\n  if (nmi < 0) return 0;\n  return nmi;\n}\n\n/**\n * Computes the Fowlkes-Mallows Index (FMI).\n *\n * The FMI is the geometric mean of pairwise precision and recall.\n * It ranges from 0 to 1, where 1 indicates perfect agreement.\n *\n * **Time Complexity**: O(n + k₁ * k₂)\n * **Space Complexity**: O(k₁ * k₂)\n *\n * @param labelsTrue - Ground truth cluster labels\n * @param labelsPred - Predicted cluster labels\n * @returns Fowlkes-Mallows score in range [0, 1]\n *\n * @throws {ShapeError} If labelsTrue and labelsPred have different sizes\n * @throws {DTypeError} If labels are string\n * @throws {DataValidationError} If labels contain non-finite or non-integer values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.fowlkes_mallows_score.html | Scikit-learn fowlkes_mallows_score}\n */\nexport function fowlkesMallowsScore(labelsTrue: Tensor, labelsPred: Tensor): number {\n  const stats = buildContingencyStats(labelsTrue, labelsPred);\n  const { contingencyDense, contingencySparse, trueCount, predCount, n } = stats;\n\n  if (n === 0) return 1.0;\n\n  let tk = 0;\n  if (contingencyDense) {\n    for (let idx = 0; idx < contingencyDense.length; idx++) {\n      const nij = readIndex(contingencyDense, idx, \"contingencyDense\");\n      if (nij > 0) tk += comb2(nij);\n    }\n  } else if (contingencySparse) {\n    for (const nij of contingencySparse.values()) {\n      tk += comb2(nij);\n    }\n  }\n\n  let pk = 0;\n  for (let i = 0; i < trueCount.length; i++) pk += comb2(readIndex(trueCount, i, \"trueCount\"));\n\n  let qk = 0;\n  for (let j = 0; j < predCount.length; j++) qk += comb2(readIndex(predCount, j, \"predCount\"));\n\n  if (pk === 0 || qk === 0) return 0.0;\n  return tk / Math.sqrt(pk * qk);\n}\n\n/**\n * Computes the homogeneity score of a clustering.\n *\n * A clustering result satisfies homogeneity if all of its clusters contain\n * only data points which are members of a single class. Score ranges from\n * 0 to 1, where 1 indicates perfectly homogeneous clustering.\n *\n * **Time Complexity**: O(n + k₁ * k₂)\n * **Space Complexity**: O(k₁ * k₂)\n *\n * @param labelsTrue - Ground truth class labels\n * @param labelsPred - Predicted cluster labels\n * @returns Homogeneity score in range [0, 1]\n *\n * @throws {ShapeError} If labelsTrue and labelsPred have different sizes\n * @throws {DTypeError} If labels are string\n * @throws {DataValidationError} If labels contain non-finite or non-integer values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.homogeneity_score.html | Scikit-learn homogeneity_score}\n */\nexport function homogeneityScore(labelsTrue: Tensor, labelsPred: Tensor): number {\n  const stats = buildContingencyStats(labelsTrue, labelsPred);\n  const { contingencyDense, contingencySparse, predCount, trueCount, nPred, n } = stats;\n\n  if (n === 0) return 1.0;\n\n  let hck = 0;\n\n  if (contingencyDense) {\n    for (let idx = 0; idx < contingencyDense.length; idx++) {\n      const nij = readIndex(contingencyDense, idx, \"contingencyDense\");\n      if (nij <= 0) continue;\n\n      const p = idx - Math.floor(idx / nPred) * nPred;\n      const nj = readIndex(predCount, p, \"predCount\");\n      if (nj > 0) hck -= (nij / n) * Math.log(nij / nj);\n    }\n  } else if (contingencySparse) {\n    for (const [key, nij] of contingencySparse) {\n      if (nij <= 0) continue;\n\n      const p = key - Math.floor(key / nPred) * nPred;\n      const nj = readIndex(predCount, p, \"predCount\");\n      if (nj > 0) hck -= (nij / n) * Math.log(nij / nj);\n    }\n  }\n\n  const hc = entropyFromCountArray(trueCount, n);\n  return hc === 0 ? 1.0 : 1.0 - hck / hc;\n}\n\n/**\n * Computes the completeness score of a clustering.\n *\n * A clustering result satisfies completeness if all data points that are\n * members of a given class are assigned to the same cluster. Score ranges\n * from 0 to 1, where 1 indicates perfectly complete clustering.\n *\n * **Time Complexity**: O(n + k₁ * k₂)\n * **Space Complexity**: O(k₁ * k₂)\n *\n * @param labelsTrue - Ground truth class labels\n * @param labelsPred - Predicted cluster labels\n * @returns Completeness score in range [0, 1]\n *\n * @throws {ShapeError} If labelsTrue and labelsPred have different sizes\n * @throws {DTypeError} If labels are string\n * @throws {DataValidationError} If labels contain non-finite or non-integer values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.completeness_score.html | Scikit-learn completeness_score}\n */\nexport function completenessScore(labelsTrue: Tensor, labelsPred: Tensor): number {\n  const stats = buildContingencyStats(labelsTrue, labelsPred);\n  const { contingencyDense, contingencySparse, trueCount, predCount, nPred, n } = stats;\n\n  if (n === 0) return 1.0;\n\n  let hkc = 0;\n\n  if (contingencyDense) {\n    for (let idx = 0; idx < contingencyDense.length; idx++) {\n      const nij = readIndex(contingencyDense, idx, \"contingencyDense\");\n      if (nij <= 0) continue;\n\n      const t = Math.floor(idx / nPred);\n      const ni = readIndex(trueCount, t, \"trueCount\");\n      if (ni > 0) hkc -= (nij / n) * Math.log(nij / ni);\n    }\n  } else if (contingencySparse) {\n    for (const [key, nij] of contingencySparse) {\n      if (nij <= 0) continue;\n\n      const t = Math.floor(key / nPred);\n      const ni = readIndex(trueCount, t, \"trueCount\");\n      if (ni > 0) hkc -= (nij / n) * Math.log(nij / ni);\n    }\n  }\n\n  const hk = entropyFromCountArray(predCount, n);\n  return hk === 0 ? 1.0 : 1.0 - hkc / hk;\n}\n\n/**\n * Computes the V-measure score of a clustering.\n *\n * V-measure is the harmonic mean of homogeneity and completeness, controlled\n * by the beta parameter. When beta > 1, completeness is weighted more; when\n * beta < 1, homogeneity is weighted more.\n *\n * **Formula**: v = (1 + beta) * h * c / (beta * h + c)\n *\n * **Time Complexity**: O(n + k₁ * k₂)\n * **Space Complexity**: O(k₁ * k₂)\n *\n * @param labelsTrue - Ground truth class labels\n * @param labelsPred - Predicted cluster labels\n * @param beta - Weight of homogeneity vs completeness (default: 1.0)\n * @returns V-measure score in range [0, 1]\n *\n * @throws {ShapeError} If labelsTrue and labelsPred have different sizes\n * @throws {DTypeError} If labels are string\n * @throws {InvalidParameterError} If beta is not a positive finite number\n * @throws {DataValidationError} If labels contain non-finite or non-integer values\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.v_measure_score.html | Scikit-learn v_measure_score}\n */\nexport function vMeasureScore(labelsTrue: Tensor, labelsPred: Tensor, beta = 1.0): number {\n  if (!Number.isFinite(beta) || beta <= 0) {\n    throw new InvalidParameterError(\"beta must be a positive finite number\", \"beta\", beta);\n  }\n\n  const h = homogeneityScore(labelsTrue, labelsPred);\n  const c = completenessScore(labelsTrue, labelsPred);\n\n  if (h + c === 0) return 0;\n  return ((1 + beta) * h * c) / (beta * h + c);\n}\n","import {\n  getNumericElement,\n  isNumericTypedArray,\n  isTypedArray,\n  type NumericTypedArray,\n} from \"../core\";\nimport { DTypeError, InvalidParameterError } from \"../core/errors\";\nimport type { Tensor } from \"../ndarray\";\nimport {\n  assertFiniteNumber,\n  assertSameSizeVectors,\n  createFlatOffsetter,\n  type FlatOffsetter,\n} from \"./_internal\";\n\nfunction getNumericRegressionData(t: Tensor, name: string): NumericTypedArray {\n  if (t.dtype === \"string\") {\n    throw new DTypeError(`${name} must be numeric tensors`);\n  }\n  if (t.dtype === \"int64\") {\n    throw new DTypeError(`${name} must be numeric tensors (int64 not supported)`);\n  }\n\n  const data = t.data;\n  if (!isTypedArray(data) || !isNumericTypedArray(data)) {\n    throw new DTypeError(`${name} must be numeric tensors`);\n  }\n\n  return data;\n}\n\nfunction readNumeric(\n  data: NumericTypedArray,\n  offsetter: FlatOffsetter,\n  index: number,\n  name: string\n) {\n  const value = getNumericElement(data, offsetter(index));\n  assertFiniteNumber(value, name, `index ${index}`);\n  return value;\n}\n\n/**\n * Calculate Mean Squared Error (MSE).\n *\n * Measures the average squared difference between predictions and actual values.\n * MSE is sensitive to outliers due to squaring the errors.\n *\n * **Formula**: MSE = (1/n) * Σ(y_true - y_pred)²\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns MSE value (always non-negative, 0 is perfect)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { mse, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const error = mse(yTrue, yPred);  // 0.375\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_squared_error.html | Scikit-learn mean_squared_error}\n */\nexport function mse(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const yTrueData = getNumericRegressionData(yTrue, \"yTrue\");\n  const yPredData = getNumericRegressionData(yPred, \"yPred\");\n\n  if (yTrue.size === 0) return 0;\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let sumSquaredError = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const diff =\n      readNumeric(yTrueData, trueOffset, i, \"yTrue\") -\n      readNumeric(yPredData, predOffset, i, \"yPred\");\n    sumSquaredError += diff * diff;\n  }\n\n  return sumSquaredError / yTrue.size;\n}\n\n/**\n * Calculate Root Mean Squared Error (RMSE).\n *\n * Square root of MSE, expressed in the same units as the target variable.\n * RMSE is more interpretable than MSE as it's in the original scale.\n *\n * **Formula**: RMSE = √(MSE) = √((1/n) * Σ(y_true - y_pred)²)\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns RMSE value (always non-negative, 0 is perfect)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { rmse, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const error = rmse(yTrue, yPred);  // √0.375 ≈ 0.612\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_squared_error.html | Scikit-learn mean_squared_error}\n */\nexport function rmse(yTrue: Tensor, yPred: Tensor): number {\n  return Math.sqrt(mse(yTrue, yPred));\n}\n\n/**\n * Calculate Mean Absolute Error (MAE).\n *\n * Measures the average absolute difference between predictions and actual values.\n * MAE is more robust to outliers than MSE.\n *\n * **Formula**: MAE = (1/n) * Σ|y_true - y_pred|\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns MAE value (always non-negative, 0 is perfect)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { mae, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const error = mae(yTrue, yPred);  // 0.5\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_absolute_error.html | Scikit-learn mean_absolute_error}\n */\nexport function mae(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const yTrueData = getNumericRegressionData(yTrue, \"yTrue\");\n  const yPredData = getNumericRegressionData(yPred, \"yPred\");\n\n  if (yTrue.size === 0) return 0;\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let sumAbsError = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const diff =\n      readNumeric(yTrueData, trueOffset, i, \"yTrue\") -\n      readNumeric(yPredData, predOffset, i, \"yPred\");\n    sumAbsError += Math.abs(diff);\n  }\n\n  return sumAbsError / yTrue.size;\n}\n\n/**\n * Calculate R² (coefficient of determination) score.\n *\n * Represents the proportion of variance in the target variable that is\n * explained by the model. R² of 1 indicates perfect predictions, 0 indicates\n * the model is no better than predicting the mean, and negative values indicate\n * the model is worse than predicting the mean.\n *\n * **Formula**: R² = 1 - (SS_res / SS_tot)\n * - SS_res = Σ(y_true - y_pred)² (residual sum of squares)\n * - SS_tot = Σ(y_true - mean(y_true))² (total sum of squares)\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns R² score (1 is perfect, 0 is baseline, negative is worse than baseline)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {InvalidParameterError} If inputs are empty\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { r2Score, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const score = r2Score(yTrue, yPred);  // Close to 1 for good fit\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.r2_score.html | Scikit-learn r2_score}\n */\nexport function r2Score(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const yTrueData = getNumericRegressionData(yTrue, \"yTrue\");\n  const yPredData = getNumericRegressionData(yPred, \"yPred\");\n  if (yTrue.size === 0) {\n    throw new InvalidParameterError(\"r2Score requires at least one sample\", \"yTrue\", yTrue.size);\n  }\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let sumTrue = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    sumTrue += readNumeric(yTrueData, trueOffset, i, \"yTrue\");\n  }\n  const mean = sumTrue / yTrue.size;\n\n  let ssRes = 0;\n  let ssTot = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumeric(yTrueData, trueOffset, i, \"yTrue\");\n    const predVal = readNumeric(yPredData, predOffset, i, \"yPred\");\n    ssRes += (trueVal - predVal) ** 2;\n    ssTot += (trueVal - mean) ** 2;\n  }\n\n  // Handle constant targets (ssTot = 0)\n  // When all true values are identical, return 0.0 (no variance to explain)\n  if (ssTot === 0) {\n    return ssRes === 0 ? 1.0 : 0.0;\n  }\n\n  return 1 - ssRes / ssTot;\n}\n\n/**\n * Calculate Adjusted R² score.\n *\n * R² adjusted for the number of features in the model. Penalizes the addition\n * of features that don't improve the model. More appropriate than R² when\n * comparing models with different numbers of features.\n *\n * **Formula**: Adjusted R² = 1 - ((1 - R²) * (n - 1)) / (n - p - 1)\n * - n = number of samples\n * - p = number of features\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * **Constraints**: Requires n > p + 1 (more samples than features + 1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @param nFeatures - Number of features (predictors) used in the model\n * @returns Adjusted R² score\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {InvalidParameterError} If nFeatures is not a non-negative integer, n <= p + 1, or inputs are empty\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { adjustedR2Score, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const score = adjustedR2Score(yTrue, yPred, 2);  // Adjusted for 2 features\n * ```\n *\n * @see {@link https://en.wikipedia.org/wiki/Coefficient_of_determination#Adjusted_R2 | Wikipedia: Adjusted R²}\n */\nexport function adjustedR2Score(yTrue: Tensor, yPred: Tensor, nFeatures: number): number {\n  if (!Number.isFinite(nFeatures) || !Number.isInteger(nFeatures) || nFeatures < 0) {\n    throw new InvalidParameterError(\n      \"nFeatures must be a non-negative integer\",\n      \"nFeatures\",\n      nFeatures\n    );\n  }\n\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const n = yTrue.size;\n  const p = nFeatures;\n\n  // Validate sufficient samples\n  if (n <= p + 1) {\n    throw new InvalidParameterError(\n      `Adjusted R² requires n > p + 1 (samples > features + 1). Got n=${n}, p=${p}`,\n      \"nFeatures\",\n      nFeatures\n    );\n  }\n\n  const r2 = r2Score(yTrue, yPred);\n\n  return 1 - ((1 - r2) * (n - 1)) / (n - p - 1);\n}\n\n/**\n * Calculate Mean Absolute Percentage Error (MAPE).\n *\n * Measures the average absolute percentage difference between predictions\n * and actual values. Expressed as a percentage, making it scale-independent\n * and easy to interpret.\n *\n * **Formula**: MAPE = (100/m) * Σ|((y_true - y_pred) / y_true)|\n * where m is the number of non-zero targets.\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * **Important**: Zero values in yTrue are skipped. If all targets are zero,\n * this function returns 0.\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns MAPE value as percentage (0 is perfect, lower is better)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { mape, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const error = mape(yTrue, yPred);  // Percentage error\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_absolute_percentage_error.html | Scikit-learn MAPE}\n */\nexport function mape(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const yTrueData = getNumericRegressionData(yTrue, \"yTrue\");\n  const yPredData = getNumericRegressionData(yPred, \"yPred\");\n  if (yTrue.size === 0) return 0;\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let sumPercentError = 0;\n  let nonZeroCount = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumeric(yTrueData, trueOffset, i, \"yTrue\");\n    const predVal = readNumeric(yPredData, predOffset, i, \"yPred\");\n    if (trueVal !== 0) {\n      sumPercentError += Math.abs((trueVal - predVal) / trueVal);\n      nonZeroCount++;\n    }\n  }\n\n  if (nonZeroCount === 0) {\n    return 0;\n  }\n\n  return (sumPercentError / nonZeroCount) * 100;\n}\n\n/**\n * Calculate Median Absolute Error (MedAE).\n *\n * Measures the median of absolute differences between predictions and actual values.\n * More robust to outliers than MAE or MSE as it uses the median instead of mean.\n *\n * **Formula**: MedAE = median(|y_true - y_pred|)\n *\n * **Time Complexity**: O(n log n) due to sorting for median calculation\n * **Space Complexity**: O(n) for storing error array\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns Median absolute error (always non-negative, 0 is perfect)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { medianAbsoluteError, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const error = medianAbsoluteError(yTrue, yPred);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.median_absolute_error.html | Scikit-learn median_absolute_error}\n */\nexport function medianAbsoluteError(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const yTrueData = getNumericRegressionData(yTrue, \"yTrue\");\n  const yPredData = getNumericRegressionData(yPred, \"yPred\");\n\n  if (yTrue.size === 0) return 0;\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  const errors: number[] = [];\n  for (let i = 0; i < yTrue.size; i++) {\n    const diff = Math.abs(\n      readNumeric(yTrueData, trueOffset, i, \"yTrue\") -\n        readNumeric(yPredData, predOffset, i, \"yPred\")\n    );\n    errors.push(diff);\n  }\n\n  errors.sort((a, b) => a - b);\n  const mid = Math.floor(errors.length / 2);\n  return errors.length % 2 !== 0\n    ? (errors[mid] ?? 0)\n    : ((errors[mid - 1] ?? 0) + (errors[mid] ?? 0)) / 2;\n}\n\n/**\n * Calculate maximum residual error.\n *\n * Returns the maximum absolute difference between predictions and actual values.\n * Useful for identifying the worst-case prediction error.\n *\n * **Formula**: max_error = max(|y_true - y_pred|)\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns Maximum absolute error (always non-negative, 0 is perfect)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { maxError, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const error = maxError(yTrue, yPred);  // 1.0 (worst prediction)\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.max_error.html | Scikit-learn max_error}\n */\nexport function maxError(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const yTrueData = getNumericRegressionData(yTrue, \"yTrue\");\n  const yPredData = getNumericRegressionData(yPred, \"yPred\");\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let maxErr = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const diff = Math.abs(\n      readNumeric(yTrueData, trueOffset, i, \"yTrue\") -\n        readNumeric(yPredData, predOffset, i, \"yPred\")\n    );\n    maxErr = Math.max(maxErr, diff);\n  }\n\n  return maxErr;\n}\n\n/**\n * Calculate explained variance score.\n *\n * Measures the proportion of variance in the target variable that is explained\n * by the model. Similar to R² but uses variance instead of sum of squares.\n * Best possible score is 1.0, lower values are worse.\n *\n * **Formula**: explained_variance = 1 - Var(y_true - y_pred) / Var(y_true)\n *\n * **Time Complexity**: O(n) where n is the number of samples\n * **Space Complexity**: O(1)\n *\n * @param yTrue - Ground truth (correct) target values\n * @param yPred - Estimated target values\n * @returns Explained variance score (1.0 is perfect, lower is worse)\n *\n * @throws {ShapeError} If yTrue and yPred have different sizes or are not 1D/column vectors\n * @throws {DTypeError} If yTrue or yPred is non-numeric or int64\n * @throws {InvalidParameterError} If inputs are empty\n * @throws {DataValidationError} If inputs contain NaN or infinite values\n *\n * @example\n * ```ts\n * import { explainedVarianceScore, tensor } from 'deepbox/metrics';\n *\n * const yTrue = tensor([3, -0.5, 2, 7]);\n * const yPred = tensor([2.5, 0.0, 2, 8]);\n * const score = explainedVarianceScore(yTrue, yPred);\n * ```\n *\n * @see {@link https://scikit-learn.org/stable/modules/generated/sklearn.metrics.explained_variance_score.html | Scikit-learn explained_variance_score}\n */\nexport function explainedVarianceScore(yTrue: Tensor, yPred: Tensor): number {\n  assertSameSizeVectors(yTrue, yPred, \"yTrue\", \"yPred\");\n  const yTrueData = getNumericRegressionData(yTrue, \"yTrue\");\n  const yPredData = getNumericRegressionData(yPred, \"yPred\");\n  if (yTrue.size === 0) {\n    throw new InvalidParameterError(\n      \"explainedVarianceScore requires at least one sample\",\n      \"yTrue\",\n      yTrue.size\n    );\n  }\n\n  const trueOffset = createFlatOffsetter(yTrue);\n  const predOffset = createFlatOffsetter(yPred);\n\n  let sumTrue = 0;\n  let sumResidual = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumeric(yTrueData, trueOffset, i, \"yTrue\");\n    const predVal = readNumeric(yPredData, predOffset, i, \"yPred\");\n    sumTrue += trueVal;\n    sumResidual += trueVal - predVal;\n  }\n  const meanTrue = sumTrue / yTrue.size;\n  const meanResidual = sumResidual / yTrue.size;\n\n  let varResidual = 0;\n  let varTrue = 0;\n  for (let i = 0; i < yTrue.size; i++) {\n    const trueVal = readNumeric(yTrueData, trueOffset, i, \"yTrue\");\n    const predVal = readNumeric(yPredData, predOffset, i, \"yPred\");\n    const residual = trueVal - predVal;\n    varResidual += (residual - meanResidual) ** 2;\n    varTrue += (trueVal - meanTrue) ** 2;\n  }\n\n  // Handle constant targets (varTrue = 0)\n  // When all true values are identical, return 0.0 (no variance to explain)\n  if (varTrue === 0) {\n    return varResidual === 0 ? 1.0 : 0.0;\n  }\n\n  return 1 - varResidual / varTrue;\n}\n"]}