@kolisachint/hoocode-agent 0.4.164 → 0.4.166

package/dist/core/embsearch/embsearch-service.js.map

@@ -1 +1 @@
-{"version":3,"file":"embsearch-service.js","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,IAAI,CAAC;AAClC,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AACtC,OAAO,EAAE,UAAU,EAAE,MAAM,8BAA8B,CAAC;AAC1D,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,eAAe,EAAE,MAAM,aAAa,CAAC;AAC9C,OAAO,EACN,cAAc,EAEd,oBAAoB,EACpB,iBAAiB,EACjB,WAAW,EAEX,aAAa,EACb,aAAa,GACb,MAAM,iBAAiB,CAAC;AACzB,OAAO,EAAqB,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAE7D;oDACoD;AACpD,MAAM,eAAe,GAAG,EAAE,CAAC;AAC3B,+EAA+E;AAC/E,MAAM,cAAc,GAAG,EAAE,CAAC;AAC1B,0FAAwF;AACxF,MAAM,aAAa,GAAG,cAAc,CAAC;AAgCrC,MAAM,OAAO,gBAAgB;IACX,OAAO,CAA0B;IAC1C,MAAM,CAA8B;IACpC,IAAI,CAAwB;IAC5B,KAAK,GAAmB,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;IAC1C,QAAQ,GAAG,KAAK,CAAC;IAEzB,YAAY,OAAgC,EAAE;QAC7C,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;IAAA,CACvB;IAED,QAAQ,GAAmB;QAC1B,OAAO,IAAI,CAAC,KAAK,CAAC;IAAA,CAClB;IAED,oFAAoF;IACpF,WAAW,GAAY;QACtB,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,OAAO,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,UAAU,CAAC;IAAA,CACvE;IAEO,QAAQ,CAAC,KAAqB,EAAQ;QAC7C,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC,KAAK,CAAC,CAAC;IAAA,CACjC;IAEO,KAAK,CAAC,aAAa,GAAgC;QAC1D,IAAI,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC;QAChC,CAAC;QACD,6EAA6E;QAC7E,8EAA8E;QAC9E,gEAAgE;QAChE,OAAO,MAAM,UAAU,CAAC,WAAW,EAAE,IAAI,EAAE,CAAC,aAAa,EAAE,UAAU,EAAE,EAAE,CAAC;YACzE,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,aAAa,EAAE,aAAa,EAAE,UAAU,EAAE,CAAC,CAAC;QAAA,CACnE,CAAC,CAAC;IAAA,CACH;IAED;;;OAGG;IACH,KAAK,CAAC,KAAK,CAAC,MAAoB,EAAiB;QAChD,IAAI,CAAC;YACJ,MAAM,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QACxB,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,MAAM,GAAG,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC1D,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,aAAa,EAAE,MAAM,EAAE,CAAC,CAAC;YAChD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;QAC1B,CAAC;IAAA,CACD;IAEO,KAAK,CAAC,GAAG,CAAC,MAAoB,EAAiB;QACtD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC;QAC1C,IAAI,CAAC,MAAM,EAAE,CAAC;YACb,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,aAAa;gBACpB,MAAM,EAAE,kEAAkE;aAC1E,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QAChD,IAAI,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,cAAc,EAAE,CAAC;YACnD,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,SAAS;gBAChB,MAAM,EAAE,yBAAyB,IAAI,CAAC,UAAU,MAAM,IAAI,CAAC,OAAO,CAAC,cAAc,SAAS;aAC1F,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,QAAQ,GAAG,oBAAoB,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QACxD,IAAI,CAAC,MAAM,GAAG,IAAI,eAAe,CAAC,EAAE,UAAU,EAAE,MAAM,EAAE,SAAS,EAAE,iBAAiB,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;QAClG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;QAE1B,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QACtC,IAAI,IAAI,CAAC,OAAO,KAAK,aAAa,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,+EAA+E,CAAC,CAAC;QAClG,CAAC;QAED,+EAA6E;QAC7E,IAAI,CAAC,IAAI,GAAG,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;QACpG,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAElC,MAAM,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC;IAAA,CAC3D;IAEO,KAAK,CAAC,iBAAiB,CAAC,KAAqB,EAAE,QAAgB,EAAE,MAAoB,EAAiB;QAC7G,MAAM,IAAI,GAAG,IAAI,CAAC,IAAK,CAAC;QACxB,MAAM,MAAM,GAAG,IAAI,CAAC,MAAO,CAAC;QAE5B,0EAA0E;QAC1E,MAAM,OAAO,GAAiE,EAAE,CAAC;QACjF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YAC1B,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnC,IAAI,KAAK,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI;gBAAE,SAAS;YAClF,IAAI,OAAe,CAAC;YACpB,IAAI,CAAC;gBACJ,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC3C,CAAC;YAAC,MAAM,CAAC;gBACR,SAAS;YACV,CAAC;YACD,MAAM,IAAI,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;YAClC,IAAI,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;gBAClC,oDAAkD;gBAClD,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;gBAC7B,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;gBACvB,SAAS;YACV,CAAC;YACD,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;QAEzE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACnD,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;YAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACtE,OAAO;QACR,CAAC;QAED,+DAA+D;QAC/D,MAAM,IAAI,GAA4F,EAAE,CAAC;QACzG,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,KAAK,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,OAAO,EAAE,CAAC;YAC/C,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC5C,IAAI,CAAC,IAAI,CAAC;gBACT,GAAG,EAAE,IAAI,CAAC,GAAG;gBACb,QAAQ,EAAE;oBACT,OAAO,EAAE,IAAI,CAAC,OAAO;oBACrB,IAAI,EAAE,IAAI,CAAC,IAAI;oBACf,IAAI;oBACJ,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;iBACnD;gBACD,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;aACvD,CAAC,CAAC;YACH,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC;QAC9B,CAAC;QAED,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;QAElE,4DAA4D;QAC5D,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;YAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAC3F,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC;QAED,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;YACzB,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;gBAAE,OAAO;YAC7C,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,MAAM,IAAI,CAAC,CAAC;YAC/D,mEAAmE;YACnE,KAAK,IAAI,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,GAAG,aAAa,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAEjG,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,IAAI,eAAe,EAAE,CAAC;gBAC7E,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;oBAAE,OAAO;gBAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,eAAe,CAAC,CAAC;gBAClE,MAAM,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBACzB,IAAI,IAAI,KAAK,CAAC,MAAM,CAAC;gBACrB,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;gBAC/D,uDAAuD;gBACvD,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,cAAc,CAAC,CAAC,CAAC;YACrE,CAAC;YACD,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC;QACtC,CAAC;QAED,MAAM,MAAM,CAAC,OAAO,EAAE,CAAC;QACvB,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;QACpB,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAAA,CACtE;IAEO,WAAW,CAAC,IAAe,EAAU;QAC5C,IAAI,CAAC,GAAG,CAAC,CAAC;QACV,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC;QAC9E,OAAO,CAAC,CAAC;IAAA,CACT;IAED,4DAA4D;IAC5D,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,CAAC,GAAG,EAAE,EAA0B;QAC3D,OAAO,MAAM,IAAI,CAAC,YAAY,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;IAAA,CACzC;IAED,wEAAwE;IACxE,KAAK,CAAC,YAAY,CAAC,KAAa,EAAE,CAAC,GAAG,EAAE,EAAE,IAAa,EAA+B;QACrF,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YACxD,MAAM,IAAI,KAAK,CAAC,iCAAiC,CAAC,CAAC;QACpD,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;QAClD,MAAM,IAAI,GAAuB,EAAE,CAAC;QACpC,MAAM,SAAS,GAAG,CAAC,GAAW,EAAW,EAAE,CAAC;YAC3C,IAAI,CAAC,IAAI;gBAAE,OAAO,IAAI,CAAC;YACvB,OAAO,SAAS,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAAA,CAC3E,CAAC;QACF,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC9B,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACvC,IAAI,GAAG,KAAK,CAAC,CAAC;gBAAE,SAAS;YACzB,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;YACpC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC;gBAAE,SAAS;YAC9B,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACjE,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;YACvD,IAAI,CAAC,KAAK;gBAAE,SAAS;YACrB,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;QACtG,CAAC;QACD,OAAO,IAAI,CAAC;IAAA,CACZ;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAW,EACX,IAAY,EACmE;QAC/E,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,CAAC,IAAI;YAAE,OAAO,SAAS,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,CAAC,SAAS,EAAE,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC5C,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI,IAAI,OAAO,EAAE,CAAC;gBAC1C,OAAO,EAAE,EAAE,EAAE,GAAG,GAAG,IAAI,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC;YAC7D,CAAC;QACF,CAAC;QACD,OAAO,SAAS,CAAC;IAAA,CACjB;IAEO,KAAK,CAAC,WAAW,GAAkB;QAC1C,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;QAC3B,IAAI,CAAC,MAAM,GAAG,SAAS,CAAC;QACxB,IAAI,MAAM,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAChC,IAAI,CAAC;gBACJ,MAAM,MAAM,CAAC,KAAK,EAAE,CAAC;YACtB,CAAC;YAAC,MAAM,CAAC;gBACR,eAAe;YAChB,CAAC;QACF,CAAC;IAAA,CACD;IAED,kEAAkE;IAClE,KAAK,CAAC,OAAO,GAAkB;QAC9B,IAAI,IAAI,CAAC,QAAQ;YAAE,OAAO;QAC1B,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;QACrB,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAC1C,IAAI,CAAC;gBACJ,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;YAC1B,CAAC;YAAC,MAAM,CAAC;gBACR,0CAA0C;YAC3C,CAAC;QACF,CAAC;QACD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;IAAA,CACzB;CACD;AAED,mCAAmC;AACnC,EAAE;AACF,uEAAuE;AACvE,8EAA8E;AAC9E,8EAA8E;AAC9E,8DAA8D;AAE9D,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA4B,CAAC;AAErD,MAAM,UAAU,wBAAwB,CAAC,GAAW,EAAE,OAAyB,EAAQ;IACtF,MAAM,GAAG,GAAG,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAC9B,IAAI,GAAG,IAAI,GAAG,KAAK,OAAO,EAAE,CAAC;QAC5B,GAAG,CAAC,OAAO,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAC,CAAC,CAAC,CAAC;IAC/B,CAAC;IACD,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;AAAA,CAC3B;AAED,MAAM,UAAU,mBAAmB,CAAC,GAAW,EAAgC;IAC9E,OAAO,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;AAAA,CACzB;AAED,MAAM,UAAU,0BAA0B,CAAC,GAAW,EAAQ;IAC7D,QAAQ,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;AAAA,CACrB","sourcesContent":["/**\n * Orchestrates semantic indexing and search for a repository.\n *\n * Lifecycle (all behind --enable-embsearchtools):\n *  1. `start()` — resolve the embsearch binary, scan the repo (ignore-aware),\n *     apply the byte threshold. Under threshold → dormant. Over → spawn the\n *     daemon, verify the backend is not the mock embedder, then index changed\n *     files in the background in small batches, reporting progress.\n *  2. `search()` — top-k semantic query, mapping chunk ids back to\n *     `path:start-end` via the sidecar metadata.\n *  3. `dispose()` — save + close the daemon.\n *\n * Every failure degrades to `unavailable` with a reason; nothing here ever\n * blocks session startup or affects grep/find.\n */\n\nimport { readFileSync } from \"fs\";\nimport { minimatch } from \"minimatch\";\nimport { ensureTool } from \"../../utils/tools-manager.js\";\nimport { chunkFile } from \"./chunker.js\";\nimport { EmbSearchClient } from \"./client.js\";\nimport {\n\temptyIndexMeta,\n\ttype FileMeta,\n\tgetEmbsearchStoreDir,\n\tgetVectorStoreDir,\n\thashContent,\n\ttype IndexMeta,\n\tloadIndexMeta,\n\tsaveIndexMeta,\n} from \"./index-meta.js\";\nimport { type RepoScanFile, scanRepo } from \"./repo-scan.js\";\n\n/** Chunks per bulk request. Small enough that a concurrent query is never\n *  stuck long behind one padded batch inference. */\nconst BULK_BATCH_SIZE = 48;\n/** Yield between batches so background indexing doesn't starve the session. */\nconst BATCH_YIELD_MS = 15;\n/** The Rust mock backend's model id — semantically meaningless, never index with it. */\nconst MOCK_MODEL_ID = \"mock-hash-v1\";\n\nexport type EmbsearchState =\n\t| { phase: \"idle\" }\n\t| { phase: \"skipped\"; reason: string }\n\t| { phase: \"downloading\"; receivedBytes: number; totalBytes: number | null }\n\t| { phase: \"indexing\"; done: number; total: number }\n\t| { phase: \"ready\"; chunkCount: number }\n\t| { phase: \"unavailable\"; reason: string };\n\nexport interface EmbsearchServiceOptions {\n\tcwd: string;\n\t/** Explicit binary path (settings override). Default: \"embsearch\" from PATH. */\n\tbinaryPath?: string;\n\t/** Minimum indexable bytes before indexing kicks in. */\n\tthresholdBytes: number;\n\t/** Progress callback for UI (footer / stderr lines). */\n\tonProgress?: (state: EmbsearchState) => void;\n}\n\nexport interface SemanticHit {\n\tpath: string;\n\tstartLine: number;\n\tendLine: number;\n\tscore: number;\n}\n\nexport interface SemanticChunkHit extends SemanticHit {\n\t/** Per-build chunk id (`relpath#index`) — the fusion identity for hybrid search. */\n\tid: string;\n}\n\nexport class EmbsearchService {\n\tprivate readonly options: EmbsearchServiceOptions;\n\tprivate client: EmbSearchClient | undefined;\n\tprivate meta: IndexMeta | undefined;\n\tprivate state: EmbsearchState = { phase: \"idle\" };\n\tprivate disposed = false;\n\n\tconstructor(options: EmbsearchServiceOptions) {\n\t\tthis.options = options;\n\t}\n\n\tgetState(): EmbsearchState {\n\t\treturn this.state;\n\t}\n\n\t/** Semantic search is usable (index ready, or still building with partial data). */\n\tisAvailable(): boolean {\n\t\treturn this.state.phase === \"ready\" || this.state.phase === \"indexing\";\n\t}\n\n\tprivate setState(state: EmbsearchState): void {\n\t\tthis.state = state;\n\t\tthis.options.onProgress?.(state);\n\t}\n\n\tprivate async resolveBinary(): Promise<string | undefined> {\n\t\tif (this.options.binaryPath) {\n\t\t\treturn this.options.binaryPath;\n\t\t}\n\t\t// Surface the on-demand binary download through the same progress channel as\n\t\t// indexing, so the first-run fetch renders a progress bar instead of a stall.\n\t\t// A cached binary resolves without ever invoking this callback.\n\t\treturn await ensureTool(\"embsearch\", true, (receivedBytes, totalBytes) => {\n\t\t\tthis.setState({ phase: \"downloading\", receivedBytes, totalBytes });\n\t\t});\n\t}\n\n\t/**\n\t * Scan, threshold-check, and (when needed) index in the background.\n\t * Resolves when indexing completes or the feature settles dormant.\n\t */\n\tasync start(signal?: AbortSignal): Promise<void> {\n\t\ttry {\n\t\t\tawait this.run(signal);\n\t\t} catch (e) {\n\t\t\tconst reason = e instanceof Error ? e.message : String(e);\n\t\t\tthis.setState({ phase: \"unavailable\", reason });\n\t\t\tawait this.closeClient();\n\t\t}\n\t}\n\n\tprivate async run(signal?: AbortSignal): Promise<void> {\n\t\tconst binary = await this.resolveBinary();\n\t\tif (!binary) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"unavailable\",\n\t\t\t\treason: \"embsearch binary not found (PATH or embsearchBinaryPath setting)\",\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst scan = scanRepo(this.options.cwd, signal);\n\t\tif (scan.totalBytes < this.options.thresholdBytes) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"skipped\",\n\t\t\t\treason: `repo under threshold (${scan.totalBytes} < ${this.options.thresholdBytes} bytes)`,\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst storeDir = getEmbsearchStoreDir(this.options.cwd);\n\t\tthis.client = new EmbSearchClient({ binaryPath: binary, storePath: getVectorStoreDir(storeDir) });\n\t\tawait this.client.ready();\n\n\t\tconst info = await this.client.info();\n\t\tif (info.modelId === MOCK_MODEL_ID) {\n\t\t\tthrow new Error(\"embsearch binary uses the mock embedder (not semantic); install an onnx build\");\n\t\t}\n\n\t\t// Missing/stale sidecar (format, chunker, or model changed) → clean rebuild.\n\t\tthis.meta = loadIndexMeta(storeDir, info.modelId) ?? emptyIndexMeta(this.options.cwd, info.modelId);\n\t\tthis.meta.lastUsedMs = Date.now();\n\n\t\tawait this.indexChangedFiles(scan.files, storeDir, signal);\n\t}\n\n\tprivate async indexChangedFiles(files: RepoScanFile[], storeDir: string, signal?: AbortSignal): Promise<void> {\n\t\tconst meta = this.meta!;\n\t\tconst client = this.client!;\n\n\t\t// Diff scan vs sidecar: cheap mtime+size check first, hash only on delta.\n\t\tconst toIndex: Array<{ file: RepoScanFile; content: string; hash: string }> = [];\n\t\tconst seen = new Set<string>();\n\t\tfor (const file of files) {\n\t\t\tseen.add(file.rel);\n\t\t\tconst known = meta.files[file.rel];\n\t\t\tif (known && known.mtimeMs === file.mtimeMs && known.size === file.size) continue;\n\t\t\tlet content: string;\n\t\t\ttry {\n\t\t\t\tcontent = readFileSync(file.abs, \"utf-8\");\n\t\t\t} catch {\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tconst hash = hashContent(content);\n\t\t\tif (known && known.hash === hash) {\n\t\t\t\t// Touched but unchanged — refresh stat info only.\n\t\t\t\tknown.mtimeMs = file.mtimeMs;\n\t\t\t\tknown.size = file.size;\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\ttoIndex.push({ file, content, hash });\n\t\t}\n\t\tconst toRemove = Object.keys(meta.files).filter((rel) => !seen.has(rel));\n\n\t\tif (toIndex.length === 0 && toRemove.length === 0) {\n\t\t\tsaveIndexMeta(storeDir, meta);\n\t\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t\t\treturn;\n\t\t}\n\n\t\t// Chunk changed files; count total upserts for exact progress.\n\t\tconst work: Array<{ rel: string; fileMeta: FileMeta; chunks: Array<{ id: string; text: string }> }> = [];\n\t\tlet totalChunks = 0;\n\t\tfor (const { file, content, hash } of toIndex) {\n\t\t\tconst chunks = chunkFile(file.rel, content);\n\t\t\twork.push({\n\t\t\t\trel: file.rel,\n\t\t\t\tfileMeta: {\n\t\t\t\t\tmtimeMs: file.mtimeMs,\n\t\t\t\t\tsize: file.size,\n\t\t\t\t\thash,\n\t\t\t\t\tchunks: chunks.map((c) => [c.startLine, c.endLine]),\n\t\t\t\t},\n\t\t\t\tchunks: chunks.map((c) => ({ id: c.id, text: c.text })),\n\t\t\t});\n\t\t\ttotalChunks += chunks.length;\n\t\t}\n\n\t\tthis.setState({ phase: \"indexing\", done: 0, total: totalChunks });\n\n\t\t// Drop vectors of deleted files and superseded chunk tails.\n\t\tfor (const rel of toRemove) {\n\t\t\tfor (let i = 0; i < meta.files[rel].chunks.length; i++) await client.remove(`${rel}#${i}`);\n\t\t\tdelete meta.files[rel];\n\t\t}\n\n\t\tlet done = 0;\n\t\tfor (const item of work) {\n\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\tconst oldChunkCount = meta.files[item.rel]?.chunks.length ?? 0;\n\t\t\t// Remove old chunks beyond the new count (upsert covers the rest).\n\t\t\tfor (let i = item.chunks.length; i < oldChunkCount; i++) await client.remove(`${item.rel}#${i}`);\n\n\t\t\tfor (let offset = 0; offset < item.chunks.length; offset += BULK_BATCH_SIZE) {\n\t\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\t\tconst batch = item.chunks.slice(offset, offset + BULK_BATCH_SIZE);\n\t\t\t\tawait client.bulk(batch);\n\t\t\t\tdone += batch.length;\n\t\t\t\tthis.setState({ phase: \"indexing\", done, total: totalChunks });\n\t\t\t\t// Yield so queries and the event loop stay responsive.\n\t\t\t\tawait new Promise((resolve) => setTimeout(resolve, BATCH_YIELD_MS));\n\t\t\t}\n\t\t\tmeta.files[item.rel] = item.fileMeta;\n\t\t}\n\n\t\tawait client.compact();\n\t\tawait client.save();\n\t\tsaveIndexMeta(storeDir, meta);\n\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t}\n\n\tprivate countChunks(meta: IndexMeta): number {\n\t\tlet n = 0;\n\t\tfor (const rel of Object.keys(meta.files)) n += meta.files[rel].chunks.length;\n\t\treturn n;\n\t}\n\n\t/** Top-`k` semantic hits as `path` + line range + score. */\n\tasync search(query: string, k = 10): Promise<SemanticHit[]> {\n\t\treturn await this.searchChunks(query, k);\n\t}\n\n\t/** Top-`k` semantic hits including their chunk ids, for rank fusion. */\n\tasync searchChunks(query: string, k = 10, glob?: string): Promise<SemanticChunkHit[]> {\n\t\tif (!this.client || this.client.isClosed || !this.meta) {\n\t\t\tthrow new Error(\"semantic index is not available\");\n\t\t}\n\t\tconst results = await this.client.query(query, k);\n\t\tconst hits: SemanticChunkHit[] = [];\n\t\tconst matchGlob = (rel: string): boolean => {\n\t\t\tif (!glob) return true;\n\t\t\treturn minimatch(rel, glob, { dot: true, matchBase: !glob.includes(\"/\") });\n\t\t};\n\t\tfor (const result of results) {\n\t\t\tconst sep = result.id.lastIndexOf(\"#\");\n\t\t\tif (sep === -1) continue;\n\t\t\tconst rel = result.id.slice(0, sep);\n\t\t\tif (!matchGlob(rel)) continue;\n\t\t\tconst chunkIndex = Number.parseInt(result.id.slice(sep + 1), 10);\n\t\t\tconst range = this.meta.files[rel]?.chunks[chunkIndex];\n\t\t\tif (!range) continue;\n\t\t\thits.push({ id: result.id, path: rel, startLine: range[0], endLine: range[1], score: result.score });\n\t\t}\n\t\treturn hits;\n\t}\n\n\t/**\n\t * Resolve a repo-relative path + line to its enclosing indexed chunk, or\n\t * undefined when the file/line is not covered by the index. Chunks overlap\n\t * by a few lines; the first (lowest-index) containing chunk wins so the\n\t * mapping is deterministic.\n\t */\n\tfindEnclosingChunk(\n\t\trel: string,\n\t\tline: number,\n\t): { id: string; path: string; startLine: number; endLine: number } | undefined {\n\t\tconst file = this.meta?.files[rel];\n\t\tif (!file) return undefined;\n\t\tfor (let i = 0; i < file.chunks.length; i++) {\n\t\t\tconst [startLine, endLine] = file.chunks[i];\n\t\t\tif (line >= startLine && line <= endLine) {\n\t\t\t\treturn { id: `${rel}#${i}`, path: rel, startLine, endLine };\n\t\t\t}\n\t\t}\n\t\treturn undefined;\n\t}\n\n\tprivate async closeClient(): Promise<void> {\n\t\tconst client = this.client;\n\t\tthis.client = undefined;\n\t\tif (client && !client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait client.close();\n\t\t\t} catch {\n\t\t\t\t// already dead\n\t\t\t}\n\t\t}\n\t}\n\n\t/** Persist state and shut the daemon down. Safe to call twice. */\n\tasync dispose(): Promise<void> {\n\t\tif (this.disposed) return;\n\t\tthis.disposed = true;\n\t\tif (this.client && !this.client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait this.client.save();\n\t\t\t} catch {\n\t\t\t\t// daemon may have exited; nothing to save\n\t\t\t}\n\t\t}\n\t\tawait this.closeClient();\n\t}\n}\n\n// --- Per-cwd service registry ---\n//\n// The search tool is constructed by the generic tool factory table and\n// only receives `cwd`; the service is created later during session init (flag\n// gated). This registry connects the two without threading a service instance\n// through every layer between main.ts and the tool factories.\n\nconst services = new Map<string, EmbsearchService>();\n\nexport function registerEmbsearchService(cwd: string, service: EmbsearchService): void {\n\tconst old = services.get(cwd);\n\tif (old && old !== service) {\n\t\told.dispose().catch(() => {});\n\t}\n\tservices.set(cwd, service);\n}\n\nexport function getEmbsearchService(cwd: string): EmbsearchService | undefined {\n\treturn services.get(cwd);\n}\n\nexport function unregisterEmbsearchService(cwd: string): void {\n\tservices.delete(cwd);\n}\n"]}
+{"version":3,"file":"embsearch-service.js","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EAAE,YAAY,EAAE,MAAM,IAAI,CAAC;AAClC,OAAO,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AACtC,OAAO,EAAE,UAAU,EAAE,MAAM,8BAA8B,CAAC;AAC1D,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAEN,eAAe,GAGf,MAAM,aAAa,CAAC;AACrB,OAAO,EACN,cAAc,EAEd,oBAAoB,EACpB,iBAAiB,EACjB,WAAW,EAEX,aAAa,EACb,aAAa,GACb,MAAM,iBAAiB,CAAC;AACzB,OAAO,EAAqB,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAE7D;oDACoD;AACpD,MAAM,eAAe,GAAG,EAAE,CAAC;AAC3B,+EAA+E;AAC/E,MAAM,cAAc,GAAG,EAAE,CAAC;AAC1B,0FAAwF;AACxF,MAAM,aAAa,GAAG,cAAc,CAAC;AACrC;;;;;;;GAOG;AACH,MAAM,6BAA6B,GAAG,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAU,CAAC;AACzD,uEAAuE;AACvE,MAAM,kBAAkB,GAAG,CAAC,CAAC,EAAE,CAAC,EAAE,CAAC,CAAU,CAAC;AAE9C,0EAA0E;AAC1E,SAAS,kBAAkB,CAAC,MAAc,EAAwB;IACjE,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,qBAAqB,CAAC,CAAC;IAClD,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AAAA,CAClF;AAED,SAAS,OAAO,CAAC,OAA0B,EAAE,OAA0B,EAAW;IACjF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACzC,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAC7B,IAAI,IAAI,KAAK,OAAO,CAAC,CAAC,CAAC;YAAE,OAAO,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;IACnD,CAAC;IACD,OAAO,IAAI,CAAC;AAAA,CACZ;AA2CD,MAAM,OAAO,gBAAgB;IACX,OAAO,CAA0B;IAC1C,MAAM,CAA8B;IACpC,IAAI,CAAwB;IAC5B,KAAK,GAAmB,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;IAC1C,QAAQ,GAAG,KAAK,CAAC;IACzB,iEAAiE;IACzD,gBAAgB,GAAG,KAAK,CAAC;IACjC,wEAAwE;IAChE,YAAY,GAAG,KAAK,CAAC;IAE7B,YAAY,OAAgC,EAAE;QAC7C,IAAI,CAAC,OAAO,GAAG,OAAO,CAAC;IAAA,CACvB;IAED,QAAQ,GAAmB;QAC1B,OAAO,IAAI,CAAC,KAAK,CAAC;IAAA,CAClB;IAED,oFAAoF;IACpF,WAAW,GAAY;QACtB,OAAO,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,OAAO,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,UAAU,CAAC;IAAA,CACvE;IAEO,QAAQ,CAAC,KAAqB,EAAQ;QAC7C,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC;QACnB,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC,KAAK,CAAC,CAAC;IAAA,CACjC;IAED,8DAA8D;IAC9D,wBAAwB,GAAY;QACnC,OAAO,IAAI,CAAC,gBAAgB,CAAC;IAAA,CAC7B;IAED,iEAAiE;IACjE,oBAAoB,GAAY;QAC/B,OAAO,IAAI,CAAC,YAAY,CAAC;IAAA,CACzB;IAED;;;;;;OAMG;IACH,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,QAAkC,EAAE,CAAS,EAAoC;QAC5G,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAC1C,MAAM,IAAI,KAAK,CAAC,iCAAiC,CAAC,CAAC;QACpD,CAAC;QACD,IAAI,CAAC,IAAI,CAAC,YAAY,EAAE,CAAC;YACxB,MAAM,IAAI,KAAK,CACd,4CAA4C,kBAAkB,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK;gBAC5E,8CAA8C,CAC/C,CAAC;QACH,CAAC;QACD,OAAO,MAAM,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC;IAAA,CACpD;IAEO,kBAAkB,CAAC,MAAc,EAAwB;QAChE,IAAI,CAAC;YACJ,OAAO,kBAAkB,CAAC,YAAY,CAAC,MAAM,EAAE,CAAC,WAAW,CAAC,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC;QACxG,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,SAAS,CAAC;QAClB,CAAC;IAAA,CACD;IAEO,KAAK,CAAC,aAAa,GAAgC;QAC1D,IAAI,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC;QAChC,CAAC;QACD,6EAA6E;QAC7E,8EAA8E;QAC9E,gEAAgE;QAChE,OAAO,MAAM,UAAU,CAAC,WAAW,EAAE,IAAI,EAAE,CAAC,aAAa,EAAE,UAAU,EAAE,EAAE,CAAC;YACzE,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,aAAa,EAAE,aAAa,EAAE,UAAU,EAAE,CAAC,CAAC;QAAA,CACnE,CAAC,CAAC;IAAA,CACH;IAED;;;OAGG;IACH,KAAK,CAAC,KAAK,CAAC,MAAoB,EAAiB;QAChD,IAAI,CAAC;YACJ,MAAM,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QACxB,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,MAAM,GAAG,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC1D,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,aAAa,EAAE,MAAM,EAAE,CAAC,CAAC;YAChD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;QAC1B,CAAC;IAAA,CACD;IAEO,KAAK,CAAC,GAAG,CAAC,MAAoB,EAAiB;QACtD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,aAAa,EAAE,CAAC;QAC1C,IAAI,CAAC,MAAM,EAAE,CAAC;YACb,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,aAAa;gBACpB,MAAM,EAAE,kEAAkE;aAC1E,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;QAChD,IAAI,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,OAAO,CAAC,cAAc,EAAE,CAAC;YACnD,IAAI,CAAC,QAAQ,CAAC;gBACb,KAAK,EAAE,SAAS;gBAChB,MAAM,EAAE,yBAAyB,IAAI,CAAC,UAAU,MAAM,IAAI,CAAC,OAAO,CAAC,cAAc,SAAS;aAC1F,CAAC,CAAC;YACH,OAAO;QACR,CAAC;QAED,MAAM,aAAa,GAAG,IAAI,CAAC,kBAAkB,CAAC,MAAM,CAAC,CAAC;QACtD,IAAI,CAAC,gBAAgB,GAAG,aAAa,KAAK,SAAS,IAAI,OAAO,CAAC,aAAa,EAAE,6BAA6B,CAAC,CAAC;QAC7G,IAAI,CAAC,YAAY,GAAG,aAAa,KAAK,SAAS,IAAI,OAAO,CAAC,aAAa,EAAE,kBAAkB,CAAC,CAAC;QAE9F,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,IAAI,oBAAoB,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QACjF,IAAI,CAAC,MAAM,GAAG,IAAI,eAAe,CAAC;YACjC,UAAU,EAAE,MAAM;YAClB,SAAS,EAAE,iBAAiB,CAAC,QAAQ,CAAC;YACtC,MAAM,EAAE,IAAI,CAAC,OAAO,CAAC,WAAW;SAChC,CAAC,CAAC;QACH,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;QAE1B,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QACtC,IAAI,IAAI,CAAC,OAAO,KAAK,aAAa,EAAE,CAAC;YACpC,MAAM,IAAI,KAAK,CAAC,+EAA+E,CAAC,CAAC;QAClG,CAAC;QAED,+EAA6E;QAC7E,IAAI,CAAC,IAAI,GAAG,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;QACpG,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAElC,MAAM,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC;IAAA,CAC3D;IAEO,KAAK,CAAC,iBAAiB,CAAC,KAAqB,EAAE,QAAgB,EAAE,MAAoB,EAAiB;QAC7G,MAAM,IAAI,GAAG,IAAI,CAAC,IAAK,CAAC;QACxB,MAAM,MAAM,GAAG,IAAI,CAAC,MAAO,CAAC;QAE5B,0EAA0E;QAC1E,MAAM,OAAO,GAAiE,EAAE,CAAC;QACjF,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;QAC/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YAC1B,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnB,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACnC,IAAI,KAAK,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI;gBAAE,SAAS;YAClF,IAAI,OAAe,CAAC;YACpB,IAAI,CAAC;gBACJ,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC3C,CAAC;YAAC,MAAM,CAAC;gBACR,SAAS;YACV,CAAC;YACD,MAAM,IAAI,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;YAClC,IAAI,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,IAAI,EAAE,CAAC;gBAClC,oDAAkD;gBAClD,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;gBAC7B,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;gBACvB,SAAS;YACV,CAAC;YACD,OAAO,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;QAEzE,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YACnD,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;YAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACtE,OAAO;QACR,CAAC;QAED,+DAA+D;QAC/D,MAAM,IAAI,GAA4F,EAAE,CAAC;QACzG,IAAI,WAAW,GAAG,CAAC,CAAC;QACpB,KAAK,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,IAAI,OAAO,EAAE,CAAC;YAC/C,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC5C,IAAI,CAAC,IAAI,CAAC;gBACT,GAAG,EAAE,IAAI,CAAC,GAAG;gBACb,QAAQ,EAAE;oBACT,OAAO,EAAE,IAAI,CAAC,OAAO;oBACrB,IAAI,EAAE,IAAI,CAAC,IAAI;oBACf,IAAI;oBACJ,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC;iBACnD;gBACD,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;aACvD,CAAC,CAAC;YACH,WAAW,IAAI,MAAM,CAAC,MAAM,CAAC;QAC9B,CAAC;QAED,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;QAElE,4DAA4D;QAC5D,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;YAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAC3F,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC;QAED,IAAI,IAAI,GAAG,CAAC,CAAC;QACb,KAAK,MAAM,IAAI,IAAI,IAAI,EAAE,CAAC;YACzB,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;gBAAE,OAAO;YAC7C,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,MAAM,IAAI,CAAC,CAAC;YAC/D,mEAAmE;YACnE,KAAK,IAAI,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,GAAG,aAAa,EAAE,CAAC,EAAE;gBAAE,MAAM,MAAM,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC;YAEjG,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,MAAM,IAAI,eAAe,EAAE,CAAC;gBAC7E,IAAI,MAAM,EAAE,OAAO,IAAI,IAAI,CAAC,QAAQ;oBAAE,OAAO;gBAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,GAAG,eAAe,CAAC,CAAC;gBAClE,MAAM,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBACzB,IAAI,IAAI,KAAK,CAAC,MAAM,CAAC;gBACrB,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,KAAK,EAAE,WAAW,EAAE,CAAC,CAAC;gBAC/D,uDAAuD;gBACvD,MAAM,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,cAAc,CAAC,CAAC,CAAC;YACrE,CAAC;YACD,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC;QACtC,CAAC;QAED,MAAM,MAAM,CAAC,OAAO,EAAE,CAAC;QACvB,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;QACpB,aAAa,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QAC9B,IAAI,CAAC,QAAQ,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IAAA,CACtE;IAEO,WAAW,CAAC,IAAe,EAAU;QAC5C,IAAI,CAAC,GAAG,CAAC,CAAC;QACV,KAAK,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC;QAC9E,OAAO,CAAC,CAAC;IAAA,CACT;IAED,4DAA4D;IAC5D,KAAK,CAAC,MAAM,CAAC,KAAa,EAAE,CAAC,GAAG,EAAE,EAA0B;QAC3D,OAAO,MAAM,IAAI,CAAC,YAAY,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;IAAA,CACzC;IAED,wEAAwE;IACxE,KAAK,CAAC,YAAY,CACjB,KAAa,EACb,CAAC,GAAG,EAAE,EACN,IAAa,EACb,SAAS,GAAoB,OAAO,EACN;QAC9B,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,MAAM,CAAC,QAAQ,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YACxD,MAAM,IAAI,KAAK,CAAC,iCAAiC,CAAC,CAAC;QACpD,CAAC;QACD,IAAI,SAAS,KAAK,SAAS,IAAI,CAAC,IAAI,CAAC,gBAAgB,EAAE,CAAC;YACvD,MAAM,IAAI,KAAK,CACd,0DAA0D,6BAA6B,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK;gBACrG,kEAAkE,CACnE,CAAC;QACH,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,EAAE,SAAS,CAAC,CAAC;QAC7D,MAAM,IAAI,GAAuB,EAAE,CAAC;QACpC,MAAM,SAAS,GAAG,CAAC,GAAW,EAAW,EAAE,CAAC;YAC3C,IAAI,CAAC,IAAI;gBAAE,OAAO,IAAI,CAAC;YACvB,OAAO,SAAS,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAAA,CAC3E,CAAC;QACF,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC9B,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACvC,IAAI,GAAG,KAAK,CAAC,CAAC;gBAAE,SAAS;YACzB,MAAM,GAAG,GAAG,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;YACpC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC;gBAAE,SAAS;YAC9B,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACjE,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;YACvD,IAAI,CAAC,KAAK;gBAAE,SAAS;YACrB,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;QACtG,CAAC;QACD,OAAO,IAAI,CAAC;IAAA,CACZ;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAW,EACX,IAAY,EACmE;QAC/E,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;QACnC,IAAI,CAAC,IAAI;YAAE,OAAO,SAAS,CAAC;QAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,CAAC,SAAS,EAAE,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC5C,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI,IAAI,OAAO,EAAE,CAAC;gBAC1C,OAAO,EAAE,EAAE,EAAE,GAAG,GAAG,IAAI,CAAC,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC;YAC7D,CAAC;QACF,CAAC;QACD,OAAO,SAAS,CAAC;IAAA,CACjB;IAEO,KAAK,CAAC,WAAW,GAAkB;QAC1C,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;QAC3B,IAAI,CAAC,MAAM,GAAG,SAAS,CAAC;QACxB,IAAI,MAAM,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAChC,IAAI,CAAC;gBACJ,MAAM,MAAM,CAAC,KAAK,EAAE,CAAC;YACtB,CAAC;YAAC,MAAM,CAAC;gBACR,eAAe;YAChB,CAAC;QACF,CAAC;IAAA,CACD;IAED,kEAAkE;IAClE,KAAK,CAAC,OAAO,GAAkB;QAC9B,IAAI,IAAI,CAAC,QAAQ;YAAE,OAAO;QAC1B,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC;QACrB,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,EAAE,CAAC;YAC1C,IAAI,CAAC;gBACJ,MAAM,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;YAC1B,CAAC;YAAC,MAAM,CAAC;gBACR,0CAA0C;YAC3C,CAAC;QACF,CAAC;QACD,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC;IAAA,CACzB;CACD;AAED,mCAAmC;AACnC,EAAE;AACF,uEAAuE;AACvE,8EAA8E;AAC9E,8EAA8E;AAC9E,8DAA8D;AAE9D,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA4B,CAAC;AAErD,MAAM,UAAU,wBAAwB,CAAC,GAAW,EAAE,OAAyB,EAAQ;IACtF,MAAM,GAAG,GAAG,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAC9B,IAAI,GAAG,IAAI,GAAG,KAAK,OAAO,EAAE,CAAC;QAC5B,GAAG,CAAC,OAAO,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAC,CAAC,CAAC,CAAC;IAC/B,CAAC;IACD,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;AAAA,CAC3B;AAED,MAAM,UAAU,mBAAmB,CAAC,GAAW,EAAgC;IAC9E,OAAO,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;AAAA,CACzB;AAED,MAAM,UAAU,0BAA0B,CAAC,GAAW,EAAQ;IAC7D,QAAQ,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;AAAA,CACrB","sourcesContent":["/**\n * Orchestrates semantic indexing and search for a repository.\n *\n * Lifecycle (all behind --enable-embsearchtools):\n *  1. `start()` — resolve the embsearch binary, scan the repo (ignore-aware),\n *     apply the byte threshold. Under threshold → dormant. Over → spawn the\n *     daemon, verify the backend is not the mock embedder, then index changed\n *     files in the background in small batches, reporting progress.\n *  2. `search()` — top-k semantic query, mapping chunk ids back to\n *     `path:start-end` via the sidecar metadata.\n *  3. `dispose()` — save + close the daemon.\n *\n * Every failure degrades to `unavailable` with a reason; nothing here ever\n * blocks session startup or affects grep/find.\n */\n\nimport { execFileSync } from \"child_process\";\nimport { readFileSync } from \"fs\";\nimport { minimatch } from \"minimatch\";\nimport { ensureTool } from \"../../utils/tools-manager.js\";\nimport { chunkFile } from \"./chunker.js\";\nimport {\n\ttype DaemonRetriever,\n\tEmbSearchClient,\n\ttype EmbSearchRerankPassage,\n\ttype EmbSearchRerankResult,\n} from \"./client.js\";\nimport {\n\temptyIndexMeta,\n\ttype FileMeta,\n\tgetEmbsearchStoreDir,\n\tgetVectorStoreDir,\n\thashContent,\n\ttype IndexMeta,\n\tloadIndexMeta,\n\tsaveIndexMeta,\n} from \"./index-meta.js\";\nimport { type RepoScanFile, scanRepo } from \"./repo-scan.js\";\n\n/** Chunks per bulk request. Small enough that a concurrent query is never\n *  stuck long behind one padded batch inference. */\nconst BULK_BATCH_SIZE = 48;\n/** Yield between batches so background indexing doesn't starve the session. */\nconst BATCH_YIELD_MS = 15;\n/** The Rust mock backend's model id — semantically meaningless, never index with it. */\nconst MOCK_MODEL_ID = \"mock-hash-v1\";\n/**\n * First embsearch release serving `retriever: \"lexical\"`.\n *\n * The guard matters because the daemon does not reject unknown request fields:\n * an older binary silently ignores `retriever` and answers with dense results.\n * Fusing that list a second time as a \"bm25\" leg would double-count it and\n * corrupt the ranking with no error anywhere — so refuse instead of degrading.\n */\nconst MIN_LEXICAL_RETRIEVER_VERSION = [0, 2, 0] as const;\n/** First embsearch release bundling cross-encoder reranker weights. */\nconst MIN_RERANK_VERSION = [0, 3, 0] as const;\n\n/** `embsearch 0.2.0` -> [0, 2, 0]; undefined when it cannot be parsed. */\nfunction parseBinaryVersion(output: string): number[] | undefined {\n\tconst match = output.match(/(\\d+)\\.(\\d+)\\.(\\d+)/);\n\treturn match ? [Number(match[1]), Number(match[2]), Number(match[3])] : undefined;\n}\n\nfunction atLeast(version: readonly number[], minimum: readonly number[]): boolean {\n\tfor (let i = 0; i < minimum.length; i++) {\n\t\tconst part = version[i] ?? 0;\n\t\tif (part !== minimum[i]) return part > minimum[i];\n\t}\n\treturn true;\n}\n\nexport type EmbsearchState =\n\t| { phase: \"idle\" }\n\t| { phase: \"skipped\"; reason: string }\n\t| { phase: \"downloading\"; receivedBytes: number; totalBytes: number | null }\n\t| { phase: \"indexing\"; done: number; total: number }\n\t| { phase: \"ready\"; chunkCount: number }\n\t| { phase: \"unavailable\"; reason: string };\n\nexport interface EmbsearchServiceOptions {\n\tcwd: string;\n\t/** Explicit binary path (settings override). Default: \"embsearch\" from PATH. */\n\tbinaryPath?: string;\n\t/** Minimum indexable bytes before indexing kicks in. */\n\tthresholdBytes: number;\n\t/**\n\t * Override the store location. Only the eval harness sets this, so a\n\t * second index (e.g. a BM25-hybrid store) can exist for the same repo\n\t * without colliding with the primary one.\n\t */\n\tstoreDir?: string;\n\t/**\n\t * Create the store with the daemon's BM25 lexical index. Fixed at store\n\t * creation, so this must pair with a distinct `storeDir`.\n\t */\n\thybridStore?: boolean;\n\t/** Progress callback for UI (footer / stderr lines). */\n\tonProgress?: (state: EmbsearchState) => void;\n}\n\nexport interface SemanticHit {\n\tpath: string;\n\tstartLine: number;\n\tendLine: number;\n\tscore: number;\n}\n\nexport interface SemanticChunkHit extends SemanticHit {\n\t/** Per-build chunk id (`relpath#index`) — the fusion identity for hybrid search. */\n\tid: string;\n}\n\nexport class EmbsearchService {\n\tprivate readonly options: EmbsearchServiceOptions;\n\tprivate client: EmbSearchClient | undefined;\n\tprivate meta: IndexMeta | undefined;\n\tprivate state: EmbsearchState = { phase: \"idle\" };\n\tprivate disposed = false;\n\t/** Whether the resolved binary serves `retriever: \"lexical\"`. */\n\tprivate lexicalRetriever = false;\n\t/** Whether the resolved binary serves the cross-encoder `rerank` op. */\n\tprivate crossEncoder = false;\n\n\tconstructor(options: EmbsearchServiceOptions) {\n\t\tthis.options = options;\n\t}\n\n\tgetState(): EmbsearchState {\n\t\treturn this.state;\n\t}\n\n\t/** Semantic search is usable (index ready, or still building with partial data). */\n\tisAvailable(): boolean {\n\t\treturn this.state.phase === \"ready\" || this.state.phase === \"indexing\";\n\t}\n\n\tprivate setState(state: EmbsearchState): void {\n\t\tthis.state = state;\n\t\tthis.options.onProgress?.(state);\n\t}\n\n\t/** Whether the running daemon can serve a BM25-only query. */\n\tsupportsLexicalRetriever(): boolean {\n\t\treturn this.lexicalRetriever;\n\t}\n\n\t/** Whether the running daemon can score with a cross-encoder. */\n\tsupportsCrossEncoder(): boolean {\n\t\treturn this.crossEncoder;\n\t}\n\n\t/**\n\t * Cross-encoder rerank of caller-supplied passages.\n\t *\n\t * Unlike the retrievers this does not consult the index at all — it scores\n\t * exactly the text passed in, which is why the caller sends its expanded\n\t * windows rather than chunk ids.\n\t */\n\tasync rerank(query: string, passages: EmbSearchRerankPassage[], k: number): Promise<EmbSearchRerankResult[]> {\n\t\tif (!this.client || this.client.isClosed) {\n\t\t\tthrow new Error(\"semantic index is not available\");\n\t\t}\n\t\tif (!this.crossEncoder) {\n\t\t\tthrow new Error(\n\t\t\t\t`embsearch is too old to rerank (needs >= ${MIN_RERANK_VERSION.join(\".\")}); ` +\n\t\t\t\t\t\"an older daemon has no cross-encoder weights\",\n\t\t\t);\n\t\t}\n\t\treturn await this.client.rerank(query, passages, k);\n\t}\n\n\tprivate probeBinaryVersion(binary: string): number[] | undefined {\n\t\ttry {\n\t\t\treturn parseBinaryVersion(execFileSync(binary, [\"--version\"], { encoding: \"utf-8\", timeout: 10_000 }));\n\t\t} catch {\n\t\t\treturn undefined;\n\t\t}\n\t}\n\n\tprivate async resolveBinary(): Promise<string | undefined> {\n\t\tif (this.options.binaryPath) {\n\t\t\treturn this.options.binaryPath;\n\t\t}\n\t\t// Surface the on-demand binary download through the same progress channel as\n\t\t// indexing, so the first-run fetch renders a progress bar instead of a stall.\n\t\t// A cached binary resolves without ever invoking this callback.\n\t\treturn await ensureTool(\"embsearch\", true, (receivedBytes, totalBytes) => {\n\t\t\tthis.setState({ phase: \"downloading\", receivedBytes, totalBytes });\n\t\t});\n\t}\n\n\t/**\n\t * Scan, threshold-check, and (when needed) index in the background.\n\t * Resolves when indexing completes or the feature settles dormant.\n\t */\n\tasync start(signal?: AbortSignal): Promise<void> {\n\t\ttry {\n\t\t\tawait this.run(signal);\n\t\t} catch (e) {\n\t\t\tconst reason = e instanceof Error ? e.message : String(e);\n\t\t\tthis.setState({ phase: \"unavailable\", reason });\n\t\t\tawait this.closeClient();\n\t\t}\n\t}\n\n\tprivate async run(signal?: AbortSignal): Promise<void> {\n\t\tconst binary = await this.resolveBinary();\n\t\tif (!binary) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"unavailable\",\n\t\t\t\treason: \"embsearch binary not found (PATH or embsearchBinaryPath setting)\",\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst scan = scanRepo(this.options.cwd, signal);\n\t\tif (scan.totalBytes < this.options.thresholdBytes) {\n\t\t\tthis.setState({\n\t\t\t\tphase: \"skipped\",\n\t\t\t\treason: `repo under threshold (${scan.totalBytes} < ${this.options.thresholdBytes} bytes)`,\n\t\t\t});\n\t\t\treturn;\n\t\t}\n\n\t\tconst binaryVersion = this.probeBinaryVersion(binary);\n\t\tthis.lexicalRetriever = binaryVersion !== undefined && atLeast(binaryVersion, MIN_LEXICAL_RETRIEVER_VERSION);\n\t\tthis.crossEncoder = binaryVersion !== undefined && atLeast(binaryVersion, MIN_RERANK_VERSION);\n\n\t\tconst storeDir = this.options.storeDir ?? getEmbsearchStoreDir(this.options.cwd);\n\t\tthis.client = new EmbSearchClient({\n\t\t\tbinaryPath: binary,\n\t\t\tstorePath: getVectorStoreDir(storeDir),\n\t\t\thybrid: this.options.hybridStore,\n\t\t});\n\t\tawait this.client.ready();\n\n\t\tconst info = await this.client.info();\n\t\tif (info.modelId === MOCK_MODEL_ID) {\n\t\t\tthrow new Error(\"embsearch binary uses the mock embedder (not semantic); install an onnx build\");\n\t\t}\n\n\t\t// Missing/stale sidecar (format, chunker, or model changed) → clean rebuild.\n\t\tthis.meta = loadIndexMeta(storeDir, info.modelId) ?? emptyIndexMeta(this.options.cwd, info.modelId);\n\t\tthis.meta.lastUsedMs = Date.now();\n\n\t\tawait this.indexChangedFiles(scan.files, storeDir, signal);\n\t}\n\n\tprivate async indexChangedFiles(files: RepoScanFile[], storeDir: string, signal?: AbortSignal): Promise<void> {\n\t\tconst meta = this.meta!;\n\t\tconst client = this.client!;\n\n\t\t// Diff scan vs sidecar: cheap mtime+size check first, hash only on delta.\n\t\tconst toIndex: Array<{ file: RepoScanFile; content: string; hash: string }> = [];\n\t\tconst seen = new Set<string>();\n\t\tfor (const file of files) {\n\t\t\tseen.add(file.rel);\n\t\t\tconst known = meta.files[file.rel];\n\t\t\tif (known && known.mtimeMs === file.mtimeMs && known.size === file.size) continue;\n\t\t\tlet content: string;\n\t\t\ttry {\n\t\t\t\tcontent = readFileSync(file.abs, \"utf-8\");\n\t\t\t} catch {\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tconst hash = hashContent(content);\n\t\t\tif (known && known.hash === hash) {\n\t\t\t\t// Touched but unchanged — refresh stat info only.\n\t\t\t\tknown.mtimeMs = file.mtimeMs;\n\t\t\t\tknown.size = file.size;\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\ttoIndex.push({ file, content, hash });\n\t\t}\n\t\tconst toRemove = Object.keys(meta.files).filter((rel) => !seen.has(rel));\n\n\t\tif (toIndex.length === 0 && toRemove.length === 0) {\n\t\t\tsaveIndexMeta(storeDir, meta);\n\t\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t\t\treturn;\n\t\t}\n\n\t\t// Chunk changed files; count total upserts for exact progress.\n\t\tconst work: Array<{ rel: string; fileMeta: FileMeta; chunks: Array<{ id: string; text: string }> }> = [];\n\t\tlet totalChunks = 0;\n\t\tfor (const { file, content, hash } of toIndex) {\n\t\t\tconst chunks = chunkFile(file.rel, content);\n\t\t\twork.push({\n\t\t\t\trel: file.rel,\n\t\t\t\tfileMeta: {\n\t\t\t\t\tmtimeMs: file.mtimeMs,\n\t\t\t\t\tsize: file.size,\n\t\t\t\t\thash,\n\t\t\t\t\tchunks: chunks.map((c) => [c.startLine, c.endLine]),\n\t\t\t\t},\n\t\t\t\tchunks: chunks.map((c) => ({ id: c.id, text: c.text })),\n\t\t\t});\n\t\t\ttotalChunks += chunks.length;\n\t\t}\n\n\t\tthis.setState({ phase: \"indexing\", done: 0, total: totalChunks });\n\n\t\t// Drop vectors of deleted files and superseded chunk tails.\n\t\tfor (const rel of toRemove) {\n\t\t\tfor (let i = 0; i < meta.files[rel].chunks.length; i++) await client.remove(`${rel}#${i}`);\n\t\t\tdelete meta.files[rel];\n\t\t}\n\n\t\tlet done = 0;\n\t\tfor (const item of work) {\n\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\tconst oldChunkCount = meta.files[item.rel]?.chunks.length ?? 0;\n\t\t\t// Remove old chunks beyond the new count (upsert covers the rest).\n\t\t\tfor (let i = item.chunks.length; i < oldChunkCount; i++) await client.remove(`${item.rel}#${i}`);\n\n\t\t\tfor (let offset = 0; offset < item.chunks.length; offset += BULK_BATCH_SIZE) {\n\t\t\t\tif (signal?.aborted || this.disposed) return;\n\t\t\t\tconst batch = item.chunks.slice(offset, offset + BULK_BATCH_SIZE);\n\t\t\t\tawait client.bulk(batch);\n\t\t\t\tdone += batch.length;\n\t\t\t\tthis.setState({ phase: \"indexing\", done, total: totalChunks });\n\t\t\t\t// Yield so queries and the event loop stay responsive.\n\t\t\t\tawait new Promise((resolve) => setTimeout(resolve, BATCH_YIELD_MS));\n\t\t\t}\n\t\t\tmeta.files[item.rel] = item.fileMeta;\n\t\t}\n\n\t\tawait client.compact();\n\t\tawait client.save();\n\t\tsaveIndexMeta(storeDir, meta);\n\t\tthis.setState({ phase: \"ready\", chunkCount: this.countChunks(meta) });\n\t}\n\n\tprivate countChunks(meta: IndexMeta): number {\n\t\tlet n = 0;\n\t\tfor (const rel of Object.keys(meta.files)) n += meta.files[rel].chunks.length;\n\t\treturn n;\n\t}\n\n\t/** Top-`k` semantic hits as `path` + line range + score. */\n\tasync search(query: string, k = 10): Promise<SemanticHit[]> {\n\t\treturn await this.searchChunks(query, k);\n\t}\n\n\t/** Top-`k` semantic hits including their chunk ids, for rank fusion. */\n\tasync searchChunks(\n\t\tquery: string,\n\t\tk = 10,\n\t\tglob?: string,\n\t\tretriever: DaemonRetriever = \"dense\",\n\t): Promise<SemanticChunkHit[]> {\n\t\tif (!this.client || this.client.isClosed || !this.meta) {\n\t\t\tthrow new Error(\"semantic index is not available\");\n\t\t}\n\t\tif (retriever === \"lexical\" && !this.lexicalRetriever) {\n\t\t\tthrow new Error(\n\t\t\t\t`embsearch is too old for retriever \"lexical\" (needs >= ${MIN_LEXICAL_RETRIEVER_VERSION.join(\".\")}); ` +\n\t\t\t\t\t\"an older daemon ignores the field and answers with dense results\",\n\t\t\t);\n\t\t}\n\t\tconst results = await this.client.query(query, k, retriever);\n\t\tconst hits: SemanticChunkHit[] = [];\n\t\tconst matchGlob = (rel: string): boolean => {\n\t\t\tif (!glob) return true;\n\t\t\treturn minimatch(rel, glob, { dot: true, matchBase: !glob.includes(\"/\") });\n\t\t};\n\t\tfor (const result of results) {\n\t\t\tconst sep = result.id.lastIndexOf(\"#\");\n\t\t\tif (sep === -1) continue;\n\t\t\tconst rel = result.id.slice(0, sep);\n\t\t\tif (!matchGlob(rel)) continue;\n\t\t\tconst chunkIndex = Number.parseInt(result.id.slice(sep + 1), 10);\n\t\t\tconst range = this.meta.files[rel]?.chunks[chunkIndex];\n\t\t\tif (!range) continue;\n\t\t\thits.push({ id: result.id, path: rel, startLine: range[0], endLine: range[1], score: result.score });\n\t\t}\n\t\treturn hits;\n\t}\n\n\t/**\n\t * Resolve a repo-relative path + line to its enclosing indexed chunk, or\n\t * undefined when the file/line is not covered by the index. Chunks overlap\n\t * by a few lines; the first (lowest-index) containing chunk wins so the\n\t * mapping is deterministic.\n\t */\n\tfindEnclosingChunk(\n\t\trel: string,\n\t\tline: number,\n\t): { id: string; path: string; startLine: number; endLine: number } | undefined {\n\t\tconst file = this.meta?.files[rel];\n\t\tif (!file) return undefined;\n\t\tfor (let i = 0; i < file.chunks.length; i++) {\n\t\t\tconst [startLine, endLine] = file.chunks[i];\n\t\t\tif (line >= startLine && line <= endLine) {\n\t\t\t\treturn { id: `${rel}#${i}`, path: rel, startLine, endLine };\n\t\t\t}\n\t\t}\n\t\treturn undefined;\n\t}\n\n\tprivate async closeClient(): Promise<void> {\n\t\tconst client = this.client;\n\t\tthis.client = undefined;\n\t\tif (client && !client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait client.close();\n\t\t\t} catch {\n\t\t\t\t// already dead\n\t\t\t}\n\t\t}\n\t}\n\n\t/** Persist state and shut the daemon down. Safe to call twice. */\n\tasync dispose(): Promise<void> {\n\t\tif (this.disposed) return;\n\t\tthis.disposed = true;\n\t\tif (this.client && !this.client.isClosed) {\n\t\t\ttry {\n\t\t\t\tawait this.client.save();\n\t\t\t} catch {\n\t\t\t\t// daemon may have exited; nothing to save\n\t\t\t}\n\t\t}\n\t\tawait this.closeClient();\n\t}\n}\n\n// --- Per-cwd service registry ---\n//\n// The search tool is constructed by the generic tool factory table and\n// only receives `cwd`; the service is created later during session init (flag\n// gated). This registry connects the two without threading a service instance\n// through every layer between main.ts and the tool factories.\n\nconst services = new Map<string, EmbsearchService>();\n\nexport function registerEmbsearchService(cwd: string, service: EmbsearchService): void {\n\tconst old = services.get(cwd);\n\tif (old && old !== service) {\n\t\told.dispose().catch(() => {});\n\t}\n\tservices.set(cwd, service);\n}\n\nexport function getEmbsearchService(cwd: string): EmbsearchService | undefined {\n\treturn services.get(cwd);\n}\n\nexport function unregisterEmbsearchService(cwd: string): void {\n\tservices.delete(cwd);\n}\n"]}

package/dist/core/search/cross-rerank.d.ts

@@ -0,0 +1,44 @@
+/**
+ * Cross-encoder reranking via the embsearch daemon.
+ *
+ * The deterministic reranker in `rerank.ts` scores candidates with lexical
+ * evidence — term coverage, path affinity, whether the window declares a query
+ * term. That took Recall@1 from 0.10 to 0.24 and then stopped paying: a
+ * term-proximity signal aimed at the classes it still handles worst moved
+ * nothing (p = 1.00 on every metric).
+ *
+ * What is left is a ranking problem the lexical view cannot see. Across the
+ * 62-query gold set the right span reaches the fused top-50 far more often
+ * than the top-10, so the candidates are in hand and merely ordered badly. A
+ * cross-encoder reads the query and the candidate *together*, which is exactly
+ * the evidence term counting lacks.
+ *
+ * It is also far more expensive — one model pass per candidate, with no
+ * precomputation possible — so it runs over a shortlist and its depth is
+ * capped separately from the fused window.
+ */
+import type { EmbsearchService } from "../embsearch/embsearch-service.js";
+import type { FusedCandidate } from "./types.js";
+/**
+ * Candidates sent to the cross-encoder per query.
+ *
+ * Each one is a model pass, so this is a latency dial, not a quality dial:
+ * every candidate past here keeps its fused order rather than being scored.
+ */
+export declare const CROSS_ENCODER_DEPTH = 30;
+export interface CrossRerankResult {
+    candidates: FusedCandidate[];
+    latencyMs: number;
+    /** How many candidates the model actually scored. */
+    scored: number;
+}
+/**
+ * Reorder `candidates` by cross-encoder relevance.
+ *
+ * Only the first {@link CROSS_ENCODER_DEPTH} are scored; the remainder keep
+ * their incoming order and follow. Candidates whose window cannot be read are
+ * left unscored for the same reason — there is no text to give the model, and
+ * inventing one would score a fiction.
+ */
+export declare function crossEncoderRerank(query: string, candidates: readonly FusedCandidate[], cwd: string, service: EmbsearchService): Promise<CrossRerankResult>;
+//# sourceMappingURL=cross-rerank.d.ts.map

package/dist/core/search/cross-rerank.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-rerank.d.ts","sourceRoot":"","sources":["../../../src/core/search/cross-rerank.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAEH,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,mCAAmC,CAAC;AAE1E,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAEjD;;;;;GAKG;AACH,eAAO,MAAM,mBAAmB,KAAK,CAAC;AAEtC,MAAM,WAAW,iBAAiB;IACjC,UAAU,EAAE,cAAc,EAAE,CAAC;IAC7B,SAAS,EAAE,MAAM,CAAC;IAClB,qDAAqD;IACrD,MAAM,EAAE,MAAM,CAAC;CACf;AAED;;;;;;;GAOG;AACH,wBAAsB,kBAAkB,CACvC,KAAK,EAAE,MAAM,EACb,UAAU,EAAE,SAAS,cAAc,EAAE,EACrC,GAAG,EAAE,MAAM,EACX,OAAO,EAAE,gBAAgB,GACvB,OAAO,CAAC,iBAAiB,CAAC,CA4C5B","sourcesContent":["/**\n * Cross-encoder reranking via the embsearch daemon.\n *\n * The deterministic reranker in `rerank.ts` scores candidates with lexical\n * evidence — term coverage, path affinity, whether the window declares a query\n * term. That took Recall@1 from 0.10 to 0.24 and then stopped paying: a\n * term-proximity signal aimed at the classes it still handles worst moved\n * nothing (p = 1.00 on every metric).\n *\n * What is left is a ranking problem the lexical view cannot see. Across the\n * 62-query gold set the right span reaches the fused top-50 far more often\n * than the top-10, so the candidates are in hand and merely ordered badly. A\n * cross-encoder reads the query and the candidate *together*, which is exactly\n * the evidence term counting lacks.\n *\n * It is also far more expensive — one model pass per candidate, with no\n * precomputation possible — so it runs over a shortlist and its depth is\n * capped separately from the fused window.\n */\n\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport { readCandidateWindows } from \"./rerank.js\";\nimport type { FusedCandidate } from \"./types.js\";\n\n/**\n * Candidates sent to the cross-encoder per query.\n *\n * Each one is a model pass, so this is a latency dial, not a quality dial:\n * every candidate past here keeps its fused order rather than being scored.\n */\nexport const CROSS_ENCODER_DEPTH = 30;\n\nexport interface CrossRerankResult {\n\tcandidates: FusedCandidate[];\n\tlatencyMs: number;\n\t/** How many candidates the model actually scored. */\n\tscored: number;\n}\n\n/**\n * Reorder `candidates` by cross-encoder relevance.\n *\n * Only the first {@link CROSS_ENCODER_DEPTH} are scored; the remainder keep\n * their incoming order and follow. Candidates whose window cannot be read are\n * left unscored for the same reason — there is no text to give the model, and\n * inventing one would score a fiction.\n */\nexport async function crossEncoderRerank(\n\tquery: string,\n\tcandidates: readonly FusedCandidate[],\n\tcwd: string,\n\tservice: EmbsearchService,\n): Promise<CrossRerankResult> {\n\tconst startedMs = Date.now();\n\tif (candidates.length < 2) {\n\t\treturn { candidates: [...candidates], latencyMs: Date.now() - startedMs, scored: 0 };\n\t}\n\n\tconst head = candidates.slice(0, CROSS_ENCODER_DEPTH);\n\tconst tail = candidates.slice(CROSS_ENCODER_DEPTH);\n\tconst windows = readCandidateWindows(head, cwd);\n\n\tconst passages: Array<{ id: string; text: string }> = [];\n\tconst byId = new Map<string, FusedCandidate>();\n\thead.forEach((candidate, i) => {\n\t\tconst text = windows[i];\n\t\tif (text === undefined) return;\n\t\t// Fused ids are unique per query, so they round-trip as passage ids.\n\t\tpassages.push({ id: candidate.id, text });\n\t\tbyId.set(candidate.id, candidate);\n\t});\n\n\tif (passages.length === 0) {\n\t\treturn { candidates: [...candidates], latencyMs: Date.now() - startedMs, scored: 0 };\n\t}\n\n\tconst scored = await service.rerank(query, passages, passages.length);\n\n\t// Reassemble: scored candidates in model order, then anything the model did\n\t// not see, in fused order. An id the daemon did not return would otherwise\n\t// vanish from the results entirely.\n\tconst ordered: FusedCandidate[] = [];\n\tconst seen = new Set<string>();\n\tfor (const result of scored) {\n\t\tconst candidate = byId.get(result.id);\n\t\tif (candidate && !seen.has(result.id)) {\n\t\t\tordered.push(candidate);\n\t\t\tseen.add(result.id);\n\t\t}\n\t}\n\tfor (const candidate of head) {\n\t\tif (!seen.has(candidate.id)) ordered.push(candidate);\n\t}\n\tordered.push(...tail);\n\n\treturn { candidates: ordered, latencyMs: Date.now() - startedMs, scored: passages.length };\n}\n"]}

package/dist/core/search/cross-rerank.js

@@ -0,0 +1,77 @@
+/**
+ * Cross-encoder reranking via the embsearch daemon.
+ *
+ * The deterministic reranker in `rerank.ts` scores candidates with lexical
+ * evidence — term coverage, path affinity, whether the window declares a query
+ * term. That took Recall@1 from 0.10 to 0.24 and then stopped paying: a
+ * term-proximity signal aimed at the classes it still handles worst moved
+ * nothing (p = 1.00 on every metric).
+ *
+ * What is left is a ranking problem the lexical view cannot see. Across the
+ * 62-query gold set the right span reaches the fused top-50 far more often
+ * than the top-10, so the candidates are in hand and merely ordered badly. A
+ * cross-encoder reads the query and the candidate *together*, which is exactly
+ * the evidence term counting lacks.
+ *
+ * It is also far more expensive — one model pass per candidate, with no
+ * precomputation possible — so it runs over a shortlist and its depth is
+ * capped separately from the fused window.
+ */
+import { readCandidateWindows } from "./rerank.js";
+/**
+ * Candidates sent to the cross-encoder per query.
+ *
+ * Each one is a model pass, so this is a latency dial, not a quality dial:
+ * every candidate past here keeps its fused order rather than being scored.
+ */
+export const CROSS_ENCODER_DEPTH = 30;
+/**
+ * Reorder `candidates` by cross-encoder relevance.
+ *
+ * Only the first {@link CROSS_ENCODER_DEPTH} are scored; the remainder keep
+ * their incoming order and follow. Candidates whose window cannot be read are
+ * left unscored for the same reason — there is no text to give the model, and
+ * inventing one would score a fiction.
+ */
+export async function crossEncoderRerank(query, candidates, cwd, service) {
+    const startedMs = Date.now();
+    if (candidates.length < 2) {
+        return { candidates: [...candidates], latencyMs: Date.now() - startedMs, scored: 0 };
+    }
+    const head = candidates.slice(0, CROSS_ENCODER_DEPTH);
+    const tail = candidates.slice(CROSS_ENCODER_DEPTH);
+    const windows = readCandidateWindows(head, cwd);
+    const passages = [];
+    const byId = new Map();
+    head.forEach((candidate, i) => {
+        const text = windows[i];
+        if (text === undefined)
+            return;
+        // Fused ids are unique per query, so they round-trip as passage ids.
+        passages.push({ id: candidate.id, text });
+        byId.set(candidate.id, candidate);
+    });
+    if (passages.length === 0) {
+        return { candidates: [...candidates], latencyMs: Date.now() - startedMs, scored: 0 };
+    }
+    const scored = await service.rerank(query, passages, passages.length);
+    // Reassemble: scored candidates in model order, then anything the model did
+    // not see, in fused order. An id the daemon did not return would otherwise
+    // vanish from the results entirely.
+    const ordered = [];
+    const seen = new Set();
+    for (const result of scored) {
+        const candidate = byId.get(result.id);
+        if (candidate && !seen.has(result.id)) {
+            ordered.push(candidate);
+            seen.add(result.id);
+        }
+    }
+    for (const candidate of head) {
+        if (!seen.has(candidate.id))
+            ordered.push(candidate);
+    }
+    ordered.push(...tail);
+    return { candidates: ordered, latencyMs: Date.now() - startedMs, scored: passages.length };
+}
+//# sourceMappingURL=cross-rerank.js.map

package/dist/core/search/cross-rerank.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"cross-rerank.js","sourceRoot":"","sources":["../../../src/core/search/cross-rerank.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG;AAGH,OAAO,EAAE,oBAAoB,EAAE,MAAM,aAAa,CAAC;AAGnD;;;;;GAKG;AACH,MAAM,CAAC,MAAM,mBAAmB,GAAG,EAAE,CAAC;AAStC;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,kBAAkB,CACvC,KAAa,EACb,UAAqC,EACrC,GAAW,EACX,OAAyB,EACI;IAC7B,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAC7B,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC3B,OAAO,EAAE,UAAU,EAAE,CAAC,GAAG,UAAU,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IACtF,CAAC;IAED,MAAM,IAAI,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,mBAAmB,CAAC,CAAC;IACtD,MAAM,IAAI,GAAG,UAAU,CAAC,KAAK,CAAC,mBAAmB,CAAC,CAAC;IACnD,MAAM,OAAO,GAAG,oBAAoB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAEhD,MAAM,QAAQ,GAAwC,EAAE,CAAC;IACzD,MAAM,IAAI,GAAG,IAAI,GAAG,EAA0B,CAAC;IAC/C,IAAI,CAAC,OAAO,CAAC,CAAC,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9B,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACxB,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO;QAC/B,qEAAqE;QACrE,QAAQ,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,SAAS,CAAC,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1C,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,EAAE,SAAS,CAAC,CAAC;IAAA,CAClC,CAAC,CAAC;IAEH,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,OAAO,EAAE,UAAU,EAAE,CAAC,GAAG,UAAU,CAAC,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IACtF,CAAC;IAED,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,MAAM,CAAC,KAAK,EAAE,QAAQ,EAAE,QAAQ,CAAC,MAAM,CAAC,CAAC;IAEtE,4EAA4E;IAC5E,2EAA2E;IAC3E,oCAAoC;IACpC,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,KAAK,MAAM,MAAM,IAAI,MAAM,EAAE,CAAC;QAC7B,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;QACtC,IAAI,SAAS,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,EAAE,CAAC;YACvC,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YACxB,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;QACrB,CAAC;IACF,CAAC;IACD,KAAK,MAAM,SAAS,IAAI,IAAI,EAAE,CAAC;QAC9B,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC;YAAE,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IACtD,CAAC;IACD,OAAO,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC;IAEtB,OAAO,EAAE,UAAU,EAAE,OAAO,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,MAAM,EAAE,QAAQ,CAAC,MAAM,EAAE,CAAC;AAAA,CAC3F","sourcesContent":["/**\n * Cross-encoder reranking via the embsearch daemon.\n *\n * The deterministic reranker in `rerank.ts` scores candidates with lexical\n * evidence — term coverage, path affinity, whether the window declares a query\n * term. That took Recall@1 from 0.10 to 0.24 and then stopped paying: a\n * term-proximity signal aimed at the classes it still handles worst moved\n * nothing (p = 1.00 on every metric).\n *\n * What is left is a ranking problem the lexical view cannot see. Across the\n * 62-query gold set the right span reaches the fused top-50 far more often\n * than the top-10, so the candidates are in hand and merely ordered badly. A\n * cross-encoder reads the query and the candidate *together*, which is exactly\n * the evidence term counting lacks.\n *\n * It is also far more expensive — one model pass per candidate, with no\n * precomputation possible — so it runs over a shortlist and its depth is\n * capped separately from the fused window.\n */\n\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport { readCandidateWindows } from \"./rerank.js\";\nimport type { FusedCandidate } from \"./types.js\";\n\n/**\n * Candidates sent to the cross-encoder per query.\n *\n * Each one is a model pass, so this is a latency dial, not a quality dial:\n * every candidate past here keeps its fused order rather than being scored.\n */\nexport const CROSS_ENCODER_DEPTH = 30;\n\nexport interface CrossRerankResult {\n\tcandidates: FusedCandidate[];\n\tlatencyMs: number;\n\t/** How many candidates the model actually scored. */\n\tscored: number;\n}\n\n/**\n * Reorder `candidates` by cross-encoder relevance.\n *\n * Only the first {@link CROSS_ENCODER_DEPTH} are scored; the remainder keep\n * their incoming order and follow. Candidates whose window cannot be read are\n * left unscored for the same reason — there is no text to give the model, and\n * inventing one would score a fiction.\n */\nexport async function crossEncoderRerank(\n\tquery: string,\n\tcandidates: readonly FusedCandidate[],\n\tcwd: string,\n\tservice: EmbsearchService,\n): Promise<CrossRerankResult> {\n\tconst startedMs = Date.now();\n\tif (candidates.length < 2) {\n\t\treturn { candidates: [...candidates], latencyMs: Date.now() - startedMs, scored: 0 };\n\t}\n\n\tconst head = candidates.slice(0, CROSS_ENCODER_DEPTH);\n\tconst tail = candidates.slice(CROSS_ENCODER_DEPTH);\n\tconst windows = readCandidateWindows(head, cwd);\n\n\tconst passages: Array<{ id: string; text: string }> = [];\n\tconst byId = new Map<string, FusedCandidate>();\n\thead.forEach((candidate, i) => {\n\t\tconst text = windows[i];\n\t\tif (text === undefined) return;\n\t\t// Fused ids are unique per query, so they round-trip as passage ids.\n\t\tpassages.push({ id: candidate.id, text });\n\t\tbyId.set(candidate.id, candidate);\n\t});\n\n\tif (passages.length === 0) {\n\t\treturn { candidates: [...candidates], latencyMs: Date.now() - startedMs, scored: 0 };\n\t}\n\n\tconst scored = await service.rerank(query, passages, passages.length);\n\n\t// Reassemble: scored candidates in model order, then anything the model did\n\t// not see, in fused order. An id the daemon did not return would otherwise\n\t// vanish from the results entirely.\n\tconst ordered: FusedCandidate[] = [];\n\tconst seen = new Set<string>();\n\tfor (const result of scored) {\n\t\tconst candidate = byId.get(result.id);\n\t\tif (candidate && !seen.has(result.id)) {\n\t\t\tordered.push(candidate);\n\t\t\tseen.add(result.id);\n\t\t}\n\t}\n\tfor (const candidate of head) {\n\t\tif (!seen.has(candidate.id)) ordered.push(candidate);\n\t}\n\tordered.push(...tail);\n\n\treturn { candidates: ordered, latencyMs: Date.now() - startedMs, scored: passages.length };\n}\n"]}

package/dist/core/search/eval-compare.d.ts

@@ -0,0 +1,57 @@
+/**
+ * Paired comparison between two configs in one run record.
+ *
+ * Aggregate means are the wrong instrument for "should we change this
+ * default". Both configs score the *same* queries, so the informative
+ * question is per-query: how many did B improve, how many did it hurt, and
+ * could that split have come from chance? A 3pp mean gap can be two queries
+ * out of sixty moving, which is not a reason to change anything.
+ *
+ * Reports a two-sided sign test over the queries where the two configs
+ * differ. Ties carry no information about direction and are excluded, which
+ * is what makes the test paired.
+ */
+import type { EvalQueryResult } from "./eval.js";
+/** Per-query metrics that can be compared. */
+export type ComparableMetric = "recallAt1" | "recallAt5" | "recallAt10" | "recallAt50" | "mrr";
+export interface PairedQueryDelta {
+    id: string;
+    class: string;
+    before: number;
+    after: number;
+}
+export interface PairedComparison {
+    metric: ComparableMetric;
+    before: string;
+    after: string;
+    better: number;
+    worse: number;
+    tied: number;
+    meanBefore: number;
+    meanAfter: number;
+    /** Two-sided sign-test p-value over non-tied queries. 1 when all tie. */
+    pValue: number;
+    deltas: PairedQueryDelta[];
+}
+export interface ComparableRun {
+    perQuery: Array<{
+        id: string;
+        class: string;
+        results: EvalQueryResult[];
+    }>;
+}
+/** Two-sided sign test: probability of a split at least this lopsided under
+ *  a fair coin. */
+export declare function signTestPValue(better: number, worse: number): number;
+/**
+ * Pair the *same* config across two run records, so "did my change help?" is
+ * answerable at all. `comparePaired` compares two configs inside one run;
+ * this compares one config before and after a retrieval change.
+ *
+ * Only sound when both runs scored the same corpus and the same gold set —
+ * otherwise the delta includes corpus drift. The caller checks provenance.
+ */
+export declare function mergeRunsForComparison(before: ComparableRun, after: ComparableRun, label: string): ComparableRun;
+export declare function comparePaired(run: ComparableRun, before: string, after: string, metric: ComparableMetric): PairedComparison;
+export declare function formatComparison(comparison: PairedComparison): string;
+//# sourceMappingURL=eval-compare.d.ts.map

package/dist/core/search/eval-compare.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-compare.d.ts","sourceRoot":"","sources":["../../../src/core/search/eval-compare.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,WAAW,CAAC;AAEjD,8CAA8C;AAC9C,MAAM,MAAM,gBAAgB,GAAG,WAAW,GAAG,WAAW,GAAG,YAAY,GAAG,YAAY,GAAG,KAAK,CAAC;AAE/F,MAAM,WAAW,gBAAgB;IAChC,EAAE,EAAE,MAAM,CAAC;IACX,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,gBAAgB;IAChC,MAAM,EAAE,gBAAgB,CAAC;IACzB,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,MAAM,CAAC;IACf,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;IAClB,yEAAyE;IACzE,MAAM,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,gBAAgB,EAAE,CAAC;CAC3B;AAED,MAAM,WAAW,aAAa;IAC7B,QAAQ,EAAE,KAAK,CAAC;QAAE,EAAE,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,eAAe,EAAE,CAAA;KAAE,CAAC,CAAC;CAC3E;AAQD;mBACmB;AACnB,wBAAgB,cAAc,CAAC,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,GAAG,MAAM,CAOpE;AAED;;;;;;;GAOG;AACH,wBAAgB,sBAAsB,CAAC,MAAM,EAAE,aAAa,EAAE,KAAK,EAAE,aAAa,EAAE,KAAK,EAAE,MAAM,GAAG,aAAa,CAkBhH;AAED,wBAAgB,aAAa,CAC5B,GAAG,EAAE,aAAa,EAClB,MAAM,EAAE,MAAM,EACd,KAAK,EAAE,MAAM,EACb,MAAM,EAAE,gBAAgB,GACtB,gBAAgB,CA0ClB;AAED,wBAAgB,gBAAgB,CAAC,UAAU,EAAE,gBAAgB,GAAG,MAAM,CAcrE","sourcesContent":["/**\n * Paired comparison between two configs in one run record.\n *\n * Aggregate means are the wrong instrument for \"should we change this\n * default\". Both configs score the *same* queries, so the informative\n * question is per-query: how many did B improve, how many did it hurt, and\n * could that split have come from chance? A 3pp mean gap can be two queries\n * out of sixty moving, which is not a reason to change anything.\n *\n * Reports a two-sided sign test over the queries where the two configs\n * differ. Ties carry no information about direction and are excluded, which\n * is what makes the test paired.\n */\n\nimport type { EvalQueryResult } from \"./eval.js\";\n\n/** Per-query metrics that can be compared. */\nexport type ComparableMetric = \"recallAt1\" | \"recallAt5\" | \"recallAt10\" | \"recallAt50\" | \"mrr\";\n\nexport interface PairedQueryDelta {\n\tid: string;\n\tclass: string;\n\tbefore: number;\n\tafter: number;\n}\n\nexport interface PairedComparison {\n\tmetric: ComparableMetric;\n\tbefore: string;\n\tafter: string;\n\tbetter: number;\n\tworse: number;\n\ttied: number;\n\tmeanBefore: number;\n\tmeanAfter: number;\n\t/** Two-sided sign-test p-value over non-tied queries. 1 when all tie. */\n\tpValue: number;\n\tdeltas: PairedQueryDelta[];\n}\n\nexport interface ComparableRun {\n\tperQuery: Array<{ id: string; class: string; results: EvalQueryResult[] }>;\n}\n\nfunction binomial(n: number, k: number): number {\n\tlet result = 1;\n\tfor (let i = 0; i < k; i++) result = (result * (n - i)) / (i + 1);\n\treturn result;\n}\n\n/** Two-sided sign test: probability of a split at least this lopsided under\n *  a fair coin. */\nexport function signTestPValue(better: number, worse: number): number {\n\tconst n = better + worse;\n\tif (n === 0) return 1;\n\tconst tail = Math.min(better, worse);\n\tlet cumulative = 0;\n\tfor (let i = 0; i <= tail; i++) cumulative += binomial(n, i);\n\treturn Math.min(1, (cumulative / 2 ** n) * 2);\n}\n\n/**\n * Pair the *same* config across two run records, so \"did my change help?\" is\n * answerable at all. `comparePaired` compares two configs inside one run;\n * this compares one config before and after a retrieval change.\n *\n * Only sound when both runs scored the same corpus and the same gold set —\n * otherwise the delta includes corpus drift. The caller checks provenance.\n */\nexport function mergeRunsForComparison(before: ComparableRun, after: ComparableRun, label: string): ComparableRun {\n\tconst afterById = new Map(after.perQuery.map((q) => [q.id, q]));\n\tconst perQuery: ComparableRun[\"perQuery\"] = [];\n\tfor (const q of before.perQuery) {\n\t\tconst other = afterById.get(q.id);\n\t\tconst a = q.results.find((r) => r.label === label);\n\t\tconst b = other?.results.find((r) => r.label === label);\n\t\tif (!a || !b) continue;\n\t\tperQuery.push({\n\t\t\tid: q.id,\n\t\t\tclass: q.class,\n\t\t\tresults: [\n\t\t\t\t{ ...a, label: \"before\" },\n\t\t\t\t{ ...b, label: \"after\" },\n\t\t\t],\n\t\t});\n\t}\n\treturn { perQuery };\n}\n\nexport function comparePaired(\n\trun: ComparableRun,\n\tbefore: string,\n\tafter: string,\n\tmetric: ComparableMetric,\n): PairedComparison {\n\tconst pick = (results: readonly EvalQueryResult[], label: string): EvalQueryResult | undefined =>\n\t\tresults.find((r) => r.label === label);\n\n\tlet better = 0;\n\tlet worse = 0;\n\tlet tied = 0;\n\tlet sumBefore = 0;\n\tlet sumAfter = 0;\n\tlet n = 0;\n\tconst deltas: PairedQueryDelta[] = [];\n\n\tfor (const query of run.perQuery) {\n\t\tconst a = pick(query.results, before);\n\t\tconst b = pick(query.results, after);\n\t\tif (!a || !b) continue;\n\t\tn++;\n\t\tsumBefore += a[metric];\n\t\tsumAfter += b[metric];\n\t\tif (b[metric] > a[metric]) {\n\t\t\tbetter++;\n\t\t\tdeltas.push({ id: query.id, class: query.class, before: a[metric], after: b[metric] });\n\t\t} else if (b[metric] < a[metric]) {\n\t\t\tworse++;\n\t\t\tdeltas.push({ id: query.id, class: query.class, before: a[metric], after: b[metric] });\n\t\t} else {\n\t\t\ttied++;\n\t\t}\n\t}\n\n\treturn {\n\t\tmetric,\n\t\tbefore,\n\t\tafter,\n\t\tbetter,\n\t\tworse,\n\t\ttied,\n\t\tmeanBefore: n > 0 ? sumBefore / n : 0,\n\t\tmeanAfter: n > 0 ? sumAfter / n : 0,\n\t\tpValue: signTestPValue(better, worse),\n\t\tdeltas,\n\t};\n}\n\nexport function formatComparison(comparison: PairedComparison): string {\n\tconst { metric, before, after, better, worse, tied, pValue } = comparison;\n\tconst delta = comparison.meanAfter - comparison.meanBefore;\n\tconst verdict =\n\t\tpValue <= 0.05\n\t\t\t? \"significant at p<=0.05\"\n\t\t\t: `NOT significant (p=${pValue.toFixed(3)}) — mean gap is not evidence of a real difference`;\n\treturn (\n\t\t`${metric}: \"${after}\" vs \"${before}\"\\n` +\n\t\t`  mean ${comparison.meanBefore.toFixed(3)} -> ${comparison.meanAfter.toFixed(3)} ` +\n\t\t`(${delta >= 0 ? \"+\" : \"\"}${delta.toFixed(3)})\\n` +\n\t\t`  ${better} better, ${worse} worse, ${tied} tied\\n` +\n\t\t`  ${verdict}`\n\t);\n}\n"]}

package/dist/core/search/eval-compare.js

@@ -0,0 +1,114 @@
+/**
+ * Paired comparison between two configs in one run record.
+ *
+ * Aggregate means are the wrong instrument for "should we change this
+ * default". Both configs score the *same* queries, so the informative
+ * question is per-query: how many did B improve, how many did it hurt, and
+ * could that split have come from chance? A 3pp mean gap can be two queries
+ * out of sixty moving, which is not a reason to change anything.
+ *
+ * Reports a two-sided sign test over the queries where the two configs
+ * differ. Ties carry no information about direction and are excluded, which
+ * is what makes the test paired.
+ */
+function binomial(n, k) {
+    let result = 1;
+    for (let i = 0; i < k; i++)
+        result = (result * (n - i)) / (i + 1);
+    return result;
+}
+/** Two-sided sign test: probability of a split at least this lopsided under
+ *  a fair coin. */
+export function signTestPValue(better, worse) {
+    const n = better + worse;
+    if (n === 0)
+        return 1;
+    const tail = Math.min(better, worse);
+    let cumulative = 0;
+    for (let i = 0; i <= tail; i++)
+        cumulative += binomial(n, i);
+    return Math.min(1, (cumulative / 2 ** n) * 2);
+}
+/**
+ * Pair the *same* config across two run records, so "did my change help?" is
+ * answerable at all. `comparePaired` compares two configs inside one run;
+ * this compares one config before and after a retrieval change.
+ *
+ * Only sound when both runs scored the same corpus and the same gold set —
+ * otherwise the delta includes corpus drift. The caller checks provenance.
+ */
+export function mergeRunsForComparison(before, after, label) {
+    const afterById = new Map(after.perQuery.map((q) => [q.id, q]));
+    const perQuery = [];
+    for (const q of before.perQuery) {
+        const other = afterById.get(q.id);
+        const a = q.results.find((r) => r.label === label);
+        const b = other?.results.find((r) => r.label === label);
+        if (!a || !b)
+            continue;
+        perQuery.push({
+            id: q.id,
+            class: q.class,
+            results: [
+                { ...a, label: "before" },
+                { ...b, label: "after" },
+            ],
+        });
+    }
+    return { perQuery };
+}
+export function comparePaired(run, before, after, metric) {
+    const pick = (results, label) => results.find((r) => r.label === label);
+    let better = 0;
+    let worse = 0;
+    let tied = 0;
+    let sumBefore = 0;
+    let sumAfter = 0;
+    let n = 0;
+    const deltas = [];
+    for (const query of run.perQuery) {
+        const a = pick(query.results, before);
+        const b = pick(query.results, after);
+        if (!a || !b)
+            continue;
+        n++;
+        sumBefore += a[metric];
+        sumAfter += b[metric];
+        if (b[metric] > a[metric]) {
+            better++;
+            deltas.push({ id: query.id, class: query.class, before: a[metric], after: b[metric] });
+        }
+        else if (b[metric] < a[metric]) {
+            worse++;
+            deltas.push({ id: query.id, class: query.class, before: a[metric], after: b[metric] });
+        }
+        else {
+            tied++;
+        }
+    }
+    return {
+        metric,
+        before,
+        after,
+        better,
+        worse,
+        tied,
+        meanBefore: n > 0 ? sumBefore / n : 0,
+        meanAfter: n > 0 ? sumAfter / n : 0,
+        pValue: signTestPValue(better, worse),
+        deltas,
+    };
+}
+export function formatComparison(comparison) {
+    const { metric, before, after, better, worse, tied, pValue } = comparison;
+    const delta = comparison.meanAfter - comparison.meanBefore;
+    const verdict = pValue <= 0.05
+        ? "significant at p<=0.05"
+        : `NOT significant (p=${pValue.toFixed(3)}) — mean gap is not evidence of a real difference`;
+    return (`${metric}: "${after}" vs "${before}"\n` +
+        `  mean ${comparison.meanBefore.toFixed(3)} -> ${comparison.meanAfter.toFixed(3)} ` +
+        `(${delta >= 0 ? "+" : ""}${delta.toFixed(3)})\n` +
+        `  ${better} better, ${worse} worse, ${tied} tied\n` +
+        `  ${verdict}`);
+}
+//# sourceMappingURL=eval-compare.js.map

package/dist/core/search/eval-compare.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-compare.js","sourceRoot":"","sources":["../../../src/core/search/eval-compare.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAgCH,SAAS,QAAQ,CAAC,CAAS,EAAE,CAAS,EAAU;IAC/C,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE;QAAE,MAAM,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAClE,OAAO,MAAM,CAAC;AAAA,CACd;AAED;mBACmB;AACnB,MAAM,UAAU,cAAc,CAAC,MAAc,EAAE,KAAa,EAAU;IACrE,MAAM,CAAC,GAAG,MAAM,GAAG,KAAK,CAAC;IACzB,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IACtB,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC;IACrC,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,IAAI,EAAE,CAAC,EAAE;QAAE,UAAU,IAAI,QAAQ,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IAC7D,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,UAAU,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;AAAA,CAC9C;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,sBAAsB,CAAC,MAAqB,EAAE,KAAoB,EAAE,KAAa,EAAiB;IACjH,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAChE,MAAM,QAAQ,GAA8B,EAAE,CAAC;IAC/C,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;QACjC,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QAClC,MAAM,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC;QACnD,MAAM,CAAC,GAAG,KAAK,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC;QACxD,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;YAAE,SAAS;QACvB,QAAQ,CAAC,IAAI,CAAC;YACb,EAAE,EAAE,CAAC,CAAC,EAAE;YACR,KAAK,EAAE,CAAC,CAAC,KAAK;YACd,OAAO,EAAE;gBACR,EAAE,GAAG,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE;gBACzB,EAAE,GAAG,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;aACxB;SACD,CAAC,CAAC;IACJ,CAAC;IACD,OAAO,EAAE,QAAQ,EAAE,CAAC;AAAA,CACpB;AAED,MAAM,UAAU,aAAa,CAC5B,GAAkB,EAClB,MAAc,EACd,KAAa,EACb,MAAwB,EACL;IACnB,MAAM,IAAI,GAAG,CAAC,OAAmC,EAAE,KAAa,EAA+B,EAAE,CAChG,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC;IAExC,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,IAAI,SAAS,GAAG,CAAC,CAAC;IAClB,IAAI,QAAQ,GAAG,CAAC,CAAC;IACjB,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,MAAM,MAAM,GAAuB,EAAE,CAAC;IAEtC,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC;QAClC,MAAM,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;QACtC,MAAM,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;QACrC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;YAAE,SAAS;QACvB,CAAC,EAAE,CAAC;QACJ,SAAS,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC;QACvB,QAAQ,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC;QACtB,IAAI,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC;YAC3B,MAAM,EAAE,CAAC;YACT,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,KAAK,CAAC,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;QACxF,CAAC;aAAM,IAAI,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC;YAClC,KAAK,EAAE,CAAC;YACR,MAAM,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,KAAK,CAAC,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;QACxF,CAAC;aAAM,CAAC;YACP,IAAI,EAAE,CAAC;QACR,CAAC;IACF,CAAC;IAED,OAAO;QACN,MAAM;QACN,MAAM;QACN,KAAK;QACL,MAAM;QACN,KAAK;QACL,IAAI;QACJ,UAAU,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACrC,SAAS,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QACnC,MAAM,EAAE,cAAc,CAAC,MAAM,EAAE,KAAK,CAAC;QACrC,MAAM;KACN,CAAC;AAAA,CACF;AAED,MAAM,UAAU,gBAAgB,CAAC,UAA4B,EAAU;IACtE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,UAAU,CAAC;IAC1E,MAAM,KAAK,GAAG,UAAU,CAAC,SAAS,GAAG,UAAU,CAAC,UAAU,CAAC;IAC3D,MAAM,OAAO,GACZ,MAAM,IAAI,IAAI;QACb,CAAC,CAAC,wBAAwB;QAC1B,CAAC,CAAC,sBAAsB,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,qDAAmD,CAAC;IAC/F,OAAO,CACN,GAAG,MAAM,MAAM,KAAK,SAAS,MAAM,KAAK;QACxC,UAAU,UAAU,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,UAAU,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG;QACnF,IAAI,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK;QACjD,KAAK,MAAM,YAAY,KAAK,WAAW,IAAI,SAAS;QACpD,KAAK,OAAO,EAAE,CACd,CAAC;AAAA,CACF","sourcesContent":["/**\n * Paired comparison between two configs in one run record.\n *\n * Aggregate means are the wrong instrument for \"should we change this\n * default\". Both configs score the *same* queries, so the informative\n * question is per-query: how many did B improve, how many did it hurt, and\n * could that split have come from chance? A 3pp mean gap can be two queries\n * out of sixty moving, which is not a reason to change anything.\n *\n * Reports a two-sided sign test over the queries where the two configs\n * differ. Ties carry no information about direction and are excluded, which\n * is what makes the test paired.\n */\n\nimport type { EvalQueryResult } from \"./eval.js\";\n\n/** Per-query metrics that can be compared. */\nexport type ComparableMetric = \"recallAt1\" | \"recallAt5\" | \"recallAt10\" | \"recallAt50\" | \"mrr\";\n\nexport interface PairedQueryDelta {\n\tid: string;\n\tclass: string;\n\tbefore: number;\n\tafter: number;\n}\n\nexport interface PairedComparison {\n\tmetric: ComparableMetric;\n\tbefore: string;\n\tafter: string;\n\tbetter: number;\n\tworse: number;\n\ttied: number;\n\tmeanBefore: number;\n\tmeanAfter: number;\n\t/** Two-sided sign-test p-value over non-tied queries. 1 when all tie. */\n\tpValue: number;\n\tdeltas: PairedQueryDelta[];\n}\n\nexport interface ComparableRun {\n\tperQuery: Array<{ id: string; class: string; results: EvalQueryResult[] }>;\n}\n\nfunction binomial(n: number, k: number): number {\n\tlet result = 1;\n\tfor (let i = 0; i < k; i++) result = (result * (n - i)) / (i + 1);\n\treturn result;\n}\n\n/** Two-sided sign test: probability of a split at least this lopsided under\n *  a fair coin. */\nexport function signTestPValue(better: number, worse: number): number {\n\tconst n = better + worse;\n\tif (n === 0) return 1;\n\tconst tail = Math.min(better, worse);\n\tlet cumulative = 0;\n\tfor (let i = 0; i <= tail; i++) cumulative += binomial(n, i);\n\treturn Math.min(1, (cumulative / 2 ** n) * 2);\n}\n\n/**\n * Pair the *same* config across two run records, so \"did my change help?\" is\n * answerable at all. `comparePaired` compares two configs inside one run;\n * this compares one config before and after a retrieval change.\n *\n * Only sound when both runs scored the same corpus and the same gold set —\n * otherwise the delta includes corpus drift. The caller checks provenance.\n */\nexport function mergeRunsForComparison(before: ComparableRun, after: ComparableRun, label: string): ComparableRun {\n\tconst afterById = new Map(after.perQuery.map((q) => [q.id, q]));\n\tconst perQuery: ComparableRun[\"perQuery\"] = [];\n\tfor (const q of before.perQuery) {\n\t\tconst other = afterById.get(q.id);\n\t\tconst a = q.results.find((r) => r.label === label);\n\t\tconst b = other?.results.find((r) => r.label === label);\n\t\tif (!a || !b) continue;\n\t\tperQuery.push({\n\t\t\tid: q.id,\n\t\t\tclass: q.class,\n\t\t\tresults: [\n\t\t\t\t{ ...a, label: \"before\" },\n\t\t\t\t{ ...b, label: \"after\" },\n\t\t\t],\n\t\t});\n\t}\n\treturn { perQuery };\n}\n\nexport function comparePaired(\n\trun: ComparableRun,\n\tbefore: string,\n\tafter: string,\n\tmetric: ComparableMetric,\n): PairedComparison {\n\tconst pick = (results: readonly EvalQueryResult[], label: string): EvalQueryResult | undefined =>\n\t\tresults.find((r) => r.label === label);\n\n\tlet better = 0;\n\tlet worse = 0;\n\tlet tied = 0;\n\tlet sumBefore = 0;\n\tlet sumAfter = 0;\n\tlet n = 0;\n\tconst deltas: PairedQueryDelta[] = [];\n\n\tfor (const query of run.perQuery) {\n\t\tconst a = pick(query.results, before);\n\t\tconst b = pick(query.results, after);\n\t\tif (!a || !b) continue;\n\t\tn++;\n\t\tsumBefore += a[metric];\n\t\tsumAfter += b[metric];\n\t\tif (b[metric] > a[metric]) {\n\t\t\tbetter++;\n\t\t\tdeltas.push({ id: query.id, class: query.class, before: a[metric], after: b[metric] });\n\t\t} else if (b[metric] < a[metric]) {\n\t\t\tworse++;\n\t\t\tdeltas.push({ id: query.id, class: query.class, before: a[metric], after: b[metric] });\n\t\t} else {\n\t\t\ttied++;\n\t\t}\n\t}\n\n\treturn {\n\t\tmetric,\n\t\tbefore,\n\t\tafter,\n\t\tbetter,\n\t\tworse,\n\t\ttied,\n\t\tmeanBefore: n > 0 ? sumBefore / n : 0,\n\t\tmeanAfter: n > 0 ? sumAfter / n : 0,\n\t\tpValue: signTestPValue(better, worse),\n\t\tdeltas,\n\t};\n}\n\nexport function formatComparison(comparison: PairedComparison): string {\n\tconst { metric, before, after, better, worse, tied, pValue } = comparison;\n\tconst delta = comparison.meanAfter - comparison.meanBefore;\n\tconst verdict =\n\t\tpValue <= 0.05\n\t\t\t? \"significant at p<=0.05\"\n\t\t\t: `NOT significant (p=${pValue.toFixed(3)}) — mean gap is not evidence of a real difference`;\n\treturn (\n\t\t`${metric}: \"${after}\" vs \"${before}\"\\n` +\n\t\t`  mean ${comparison.meanBefore.toFixed(3)} -> ${comparison.meanAfter.toFixed(3)} ` +\n\t\t`(${delta >= 0 ? \"+\" : \"\"}${delta.toFixed(3)})\\n` +\n\t\t`  ${better} better, ${worse} worse, ${tied} tied\\n` +\n\t\t`  ${verdict}`\n\t);\n}\n"]}

package/dist/core/search/eval-gold.d.ts

@@ -0,0 +1,47 @@
+/**
+ * Gold-set loading and anchor resolution.
+ *
+ * A gold span recorded as bare line numbers rots on the next refactor, and a
+ * rotted gold set fails *silently* — it just scores lower. So each span also
+ * carries an `anchor`: a literal snippet that must sit inside the range. The
+ * anchor is the source of truth; the line numbers are a cache of where it was
+ * when the fixture was last regenerated.
+ *
+ * That gives three properties the first gold set lacked:
+ *  - `resolveGoldSet` recomputes ranges from anchors, so `--fix` re-pins the
+ *    whole set after a refactor instead of someone hand-editing 60 numbers;
+ *  - `validateGoldSet` fails loudly when an anchor has moved out of its
+ *    recorded range, drifted, or become ambiguous;
+ *  - anchors are never used for scoring, so this cannot flatter retrieval —
+ *    `spanMatchesGold` still sees only a path and a line range.
+ */
+import type { EvalGoldSpan, EvalQuery } from "./eval.js";
+export interface GoldValidationIssue {
+    queryId: string;
+    path: string;
+    problem: string;
+}
+/** Re-pin one gold span's line range from its anchor. Returns the span
+ *  unchanged when it is file-scoped or has no anchor to resolve. */
+export declare function resolveGoldSpan(corpusRoot: string, span: EvalGoldSpan): {
+    span: EvalGoldSpan;
+    problem?: string;
+};
+/** Re-pin every gold span in the set. Problems are collected, not thrown, so
+ *  a regeneration run reports all drift at once. */
+export declare function resolveGoldSet(corpusRoot: string, dataset: readonly EvalQuery[]): {
+    dataset: EvalQuery[];
+    issues: GoldValidationIssue[];
+};
+/**
+ * Check the committed fixture against the corpus without rewriting it.
+ *
+ * Catches the two ways a gold set goes wrong: the anchor no longer exists (the
+ * code was deleted or renamed), or it exists but has moved outside the
+ * recorded range (the fixture is stale and every score computed from it is
+ * wrong).
+ */
+export declare function validateGoldSet(corpusRoot: string, dataset: readonly EvalQuery[]): GoldValidationIssue[];
+/** Load the gold set from a fixture file. */
+export declare function loadGoldSet(fixturePath: string): EvalQuery[];
+//# sourceMappingURL=eval-gold.d.ts.map

package/dist/core/search/eval-gold.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"eval-gold.d.ts","sourceRoot":"","sources":["../../../src/core/search/eval-gold.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAIH,OAAO,KAAK,EAAE,YAAY,EAAE,SAAS,EAAE,MAAM,WAAW,CAAC;AAYzD,MAAM,WAAW,mBAAmB;IACnC,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,EAAE,MAAM,CAAC;IACb,OAAO,EAAE,MAAM,CAAC;CAChB;AA2CD;oEACoE;AACpE,wBAAgB,eAAe,CAAC,UAAU,EAAE,MAAM,EAAE,IAAI,EAAE,YAAY,GAAG;IAAE,IAAI,EAAE,YAAY,CAAC;IAAC,OAAO,CAAC,EAAE,MAAM,CAAA;CAAE,CAiBhH;AAED;oDACoD;AACpD,wBAAgB,cAAc,CAC7B,UAAU,EAAE,MAAM,EAClB,OAAO,EAAE,SAAS,SAAS,EAAE,GAC3B;IAAE,OAAO,EAAE,SAAS,EAAE,CAAC;IAAC,MAAM,EAAE,mBAAmB,EAAE,CAAA;CAAE,CAWzD;AAED;;;;;;;GAOG;AACH,wBAAgB,eAAe,CAAC,UAAU,EAAE,MAAM,EAAE,OAAO,EAAE,SAAS,SAAS,EAAE,GAAG,mBAAmB,EAAE,CAmDxG;AAED,6CAA6C;AAC7C,wBAAgB,WAAW,CAAC,WAAW,EAAE,MAAM,GAAG,SAAS,EAAE,CAE5D","sourcesContent":["/**\n * Gold-set loading and anchor resolution.\n *\n * A gold span recorded as bare line numbers rots on the next refactor, and a\n * rotted gold set fails *silently* — it just scores lower. So each span also\n * carries an `anchor`: a literal snippet that must sit inside the range. The\n * anchor is the source of truth; the line numbers are a cache of where it was\n * when the fixture was last regenerated.\n *\n * That gives three properties the first gold set lacked:\n *  - `resolveGoldSet` recomputes ranges from anchors, so `--fix` re-pins the\n *    whole set after a refactor instead of someone hand-editing 60 numbers;\n *  - `validateGoldSet` fails loudly when an anchor has moved out of its\n *    recorded range, drifted, or become ambiguous;\n *  - anchors are never used for scoring, so this cannot flatter retrieval —\n *    `spanMatchesGold` still sees only a path and a line range.\n */\n\nimport { readFileSync } from \"node:fs\";\nimport path from \"node:path\";\nimport type { EvalGoldSpan, EvalQuery } from \"./eval.js\";\n\n/** Lines of context kept either side of a non-block anchor (an error literal,\n *  a single statement) — enough to be a real target, tight enough that a\n *  60-line chunk elsewhere in the file does not count as a hit. */\nconst LINE_ANCHOR_PAD = 2;\n/** Upper bound on a resolved block, so an anchor that fails to find its\n *  closing brace cannot silently swallow an entire file. */\nconst MAX_BLOCK_LINES = 120;\n/** Fallback extent when a block anchor never finds its closing brace. */\nconst UNCLOSED_BLOCK_LINES = 15;\n\nexport interface GoldValidationIssue {\n\tqueryId: string;\n\tpath: string;\n\tproblem: string;\n}\n\nfunction readLines(corpusRoot: string, rel: string): string[] | undefined {\n\ttry {\n\t\treturn readFileSync(path.resolve(corpusRoot, rel), \"utf-8\").split(\"\\n\");\n\t} catch {\n\t\treturn undefined;\n\t}\n}\n\n/** Indices (0-based) of every line containing `anchor`. */\nfunction findAnchorLines(lines: readonly string[], anchor: string): number[] {\n\tconst found: number[] = [];\n\tfor (let i = 0; i < lines.length; i++) {\n\t\tif (lines[i].includes(anchor)) found.push(i);\n\t}\n\treturn found;\n}\n\n/**\n * Extent of the declaration an anchor names.\n *\n * A line ending in an opener (`{`, `(`, `[`) starts a block, which runs to the\n * first line closing at column 0 — the shape every top-level declaration in\n * this codebase has. Anything else is a statement, scored with a small pad.\n */\nfunction resolveExtent(lines: readonly string[], anchorIndex: number): { startLine: number; endLine: number } {\n\tconst head = lines[anchorIndex];\n\tconst opensBlock = /[{([]\\s*$/.test(head);\n\tif (!opensBlock) {\n\t\treturn {\n\t\t\tstartLine: Math.max(1, anchorIndex + 1 - LINE_ANCHOR_PAD),\n\t\t\tendLine: Math.min(lines.length, anchorIndex + 1 + LINE_ANCHOR_PAD),\n\t\t};\n\t}\n\tconst limit = Math.min(lines.length, anchorIndex + 1 + MAX_BLOCK_LINES);\n\tfor (let i = anchorIndex + 1; i < limit; i++) {\n\t\t// Column-0 close: `}`, `};`, `});`, `];` — the end of a top-level block.\n\t\tif (/^[}\\])]/.test(lines[i])) return { startLine: anchorIndex + 1, endLine: i + 1 };\n\t}\n\treturn { startLine: anchorIndex + 1, endLine: Math.min(lines.length, anchorIndex + 1 + UNCLOSED_BLOCK_LINES) };\n}\n\n/** Re-pin one gold span's line range from its anchor. Returns the span\n *  unchanged when it is file-scoped or has no anchor to resolve. */\nexport function resolveGoldSpan(corpusRoot: string, span: EvalGoldSpan): { span: EvalGoldSpan; problem?: string } {\n\tconst lines = readLines(corpusRoot, span.path);\n\tif (!lines) return { span, problem: `file not found: ${span.path}` };\n\n\tif (span.scope === \"file\") {\n\t\treturn { span: { ...span, startLine: 1, endLine: lines.length } };\n\t}\n\tif (!span.anchor) return { span, problem: `span-scoped gold needs an anchor: ${span.path}` };\n\n\tconst matches = findAnchorLines(lines, span.anchor);\n\tif (matches.length === 0) return { span, problem: `anchor not found in ${span.path}: ${span.anchor}` };\n\tif (matches.length > 1) {\n\t\treturn { span, problem: `anchor is ambiguous (${matches.length} matches) in ${span.path}: ${span.anchor}` };\n\t}\n\n\tconst { startLine, endLine } = resolveExtent(lines, matches[0]);\n\treturn { span: { ...span, startLine, endLine } };\n}\n\n/** Re-pin every gold span in the set. Problems are collected, not thrown, so\n *  a regeneration run reports all drift at once. */\nexport function resolveGoldSet(\n\tcorpusRoot: string,\n\tdataset: readonly EvalQuery[],\n): { dataset: EvalQuery[]; issues: GoldValidationIssue[] } {\n\tconst issues: GoldValidationIssue[] = [];\n\tconst resolved = dataset.map((query) => ({\n\t\t...query,\n\t\tgold: query.gold.map((span) => {\n\t\t\tconst result = resolveGoldSpan(corpusRoot, span);\n\t\t\tif (result.problem) issues.push({ queryId: query.id, path: span.path, problem: result.problem });\n\t\t\treturn result.span;\n\t\t}),\n\t}));\n\treturn { dataset: resolved, issues };\n}\n\n/**\n * Check the committed fixture against the corpus without rewriting it.\n *\n * Catches the two ways a gold set goes wrong: the anchor no longer exists (the\n * code was deleted or renamed), or it exists but has moved outside the\n * recorded range (the fixture is stale and every score computed from it is\n * wrong).\n */\nexport function validateGoldSet(corpusRoot: string, dataset: readonly EvalQuery[]): GoldValidationIssue[] {\n\tconst issues: GoldValidationIssue[] = [];\n\tconst seenIds = new Set<string>();\n\n\tfor (const query of dataset) {\n\t\tif (seenIds.has(query.id)) issues.push({ queryId: query.id, path: \"\", problem: \"duplicate query id\" });\n\t\tseenIds.add(query.id);\n\t\tif (query.gold.length === 0) issues.push({ queryId: query.id, path: \"\", problem: \"query has no gold spans\" });\n\n\t\tfor (const span of query.gold) {\n\t\t\tconst lines = readLines(corpusRoot, span.path);\n\t\t\tif (!lines) {\n\t\t\t\tissues.push({ queryId: query.id, path: span.path, problem: \"file not found\" });\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tif (span.startLine === undefined || span.endLine === undefined) {\n\t\t\t\tissues.push({ queryId: query.id, path: span.path, problem: \"gold span is missing line numbers\" });\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tif (span.startLine < 1 || span.endLine < span.startLine || span.endLine > lines.length) {\n\t\t\t\tissues.push({\n\t\t\t\t\tqueryId: query.id,\n\t\t\t\t\tpath: span.path,\n\t\t\t\t\tproblem: `line range ${span.startLine}-${span.endLine} out of bounds (file has ${lines.length} lines)`,\n\t\t\t\t});\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tif (span.scope === \"file\") continue;\n\t\t\tif (!span.anchor) {\n\t\t\t\tissues.push({ queryId: query.id, path: span.path, problem: \"span-scoped gold needs an anchor\" });\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tconst matches = findAnchorLines(lines, span.anchor);\n\t\t\tif (matches.length === 0) {\n\t\t\t\tissues.push({ queryId: query.id, path: span.path, problem: `anchor not found: ${span.anchor}` });\n\t\t\t} else if (matches.length > 1) {\n\t\t\t\tissues.push({\n\t\t\t\t\tqueryId: query.id,\n\t\t\t\t\tpath: span.path,\n\t\t\t\t\tproblem: `anchor is ambiguous (${matches.length} matches): ${span.anchor}`,\n\t\t\t\t});\n\t\t\t} else if (matches[0] + 1 < span.startLine || matches[0] + 1 > span.endLine) {\n\t\t\t\tissues.push({\n\t\t\t\t\tqueryId: query.id,\n\t\t\t\t\tpath: span.path,\n\t\t\t\t\tproblem: `anchor moved to line ${matches[0] + 1}, outside recorded range ${span.startLine}-${span.endLine}`,\n\t\t\t\t});\n\t\t\t}\n\t\t}\n\t}\n\treturn issues;\n}\n\n/** Load the gold set from a fixture file. */\nexport function loadGoldSet(fixturePath: string): EvalQuery[] {\n\treturn JSON.parse(readFileSync(fixturePath, \"utf-8\")) as EvalQuery[];\n}\n"]}