@kolisachint/hoocode-agent 0.4.164 → 0.4.165

package/CHANGELOG.md

@@ -1,5 +1,7 @@
 # Changelog
 
+## [0.4.165] - 2026-08-08
+
 ## [0.4.164] - 2026-08-08
 
 ### Removed

package/dist/core/embsearch/client.d.ts

@@ -19,7 +19,11 @@ export interface EmbSearchBulkResult {
     inserted: number;
     updated: number;
 }
+/** Which daemon-side retriever answers a query (embsearch >= 0.2.0). */
+export type DaemonRetriever = "dense" | "lexical" | "hybrid";
 export interface EmbSearchClientOptions {
+    /** Open/create the store with a BM25 lexical index alongside the vectors. */
+    hybrid?: boolean;
     /** Path to the `embsearch` binary. */
     binaryPath: string;
     /** Store directory passed as `--path`. */
@@ -39,8 +43,20 @@ export declare class EmbSearchClient {
     get isClosed(): boolean;
     private onStdout;
     private send;
-    /** Search for the top-`k` matches for `text`. */
-    query(text: string, k?: number): Promise<EmbSearchResult[]>;
+    /**
+     * Search for the top-`k` matches for `text`.
+     *
+     * `retriever` selects which leg answers:
+     *  - `dense` (default) — vector search, the historical behaviour;
+     *  - `lexical` — BM25 only, raw scores, no embedding computed;
+     *  - `hybrid` — both, pre-fused by the daemon's own RRF constant.
+     *
+     * `lexical` and `hybrid` need a store created with `--hybrid`. Prefer
+     * `lexical` over `hybrid` when fusing here: `hybrid` collapses both legs
+     * into one RRF score, discarding the per-retriever ranks the trace records
+     * and preventing n-way fusion with the grep leg.
+     */
+    query(text: string, k?: number, retriever?: DaemonRetriever): Promise<EmbSearchResult[]>;
     /**
      * Batched insert-or-replace. One embedding inference for the whole batch —
      * the fast path for bulk indexing. Keep batches modest (e.g. 32–64) so a

package/dist/core/embsearch/client.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"client.d.ts","sourceRoot":"","sources":["../../../src/core/embsearch/client.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAIH,MAAM,WAAW,eAAe;IAC/B,EAAE,EAAE,MAAM,CAAC;IACX,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,mBAAmB;IACnC,OAAO,EAAE,MAAM,CAAC;IAChB,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,mBAAmB;IACnC,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,sBAAsB;IACtC,sCAAsC;IACtC,UAAU,EAAE,MAAM,CAAC;IACnB,0CAA0C;IAC1C,SAAS,EAAE,MAAM,CAAC;IAClB,6DAA6D;IAC7D,MAAM,CAAC,EAAE,QAAQ,GAAG,KAAK,GAAG,WAAW,CAAC;CACxC;AAoBD,qBAAa,eAAe;IAC3B,OAAO,CAAC,IAAI,CAAiC;IAC7C,OAAO,CAAC,KAAK,CAAiB;IAC9B,OAAO,CAAC,MAAM,CAAM;IACpB,OAAO,CAAC,MAAM,CAAS;IACvB,OAAO,CAAC,YAAY,CAAgB;IAEpC,YAAY,IAAI,EAAE,sBAAsB,EA2BvC;IAED,6DAA6D;IAC7D,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,CAErB;IAED,IAAI,QAAQ,IAAI,OAAO,CAEtB;IAED,OAAO,CAAC,QAAQ;IAqBhB,OAAO,CAAC,IAAI;IAQZ,iDAAiD;IAC3C,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC,SAAK,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC,CAG5D;IAED;;;;OAIG;IACG,IAAI,CAAC,KAAK,EAAE,KAAK,CAAC;QAAE,EAAE,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC,GAAG,OAAO,CAAC,mBAAmB,CAAC,CAGnF;IAED,qEAAqE;IAC/D,IAAI,IAAI,OAAO,CAAC,mBAAmB,CAAC,CAGzC;IAED,yDAAyD;IACnD,MAAM,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAGzC;IAED,uDAAuD;IACjD,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAE7B;IAED,gDAAgD;IAC1C,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,CAE1B;IAED,+DAA+D;IACzD,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,CAI3B;CACD","sourcesContent":["/**\n * Client for the `embsearch` stdio daemon (vendored from\n * github.com/kolisachint/embeddingsearchtools ts/client.ts).\n *\n * Spawns `embsearch serve` once and talks newline-delimited JSON over its\n * stdin/stdout. The process stays alive so the model and index load a single\n * time; every query after startup is hot.\n */\n\nimport { type ChildProcessWithoutNullStreams, spawn } from \"child_process\";\n\nexport interface EmbSearchResult {\n\tid: string;\n\tscore: number;\n}\n\nexport interface EmbSearchDaemonInfo {\n\tmodelId: string;\n\tdim: number;\n\tcount: number;\n}\n\nexport interface EmbSearchBulkResult {\n\tinserted: number;\n\tupdated: number;\n}\n\nexport interface EmbSearchClientOptions {\n\t/** Path to the `embsearch` binary. */\n\tbinaryPath: string;\n\t/** Store directory passed as `--path`. */\n\tstorePath: string;\n\t/** Metric for a freshly created store. Default: \"cosine\". */\n\tmetric?: \"cosine\" | \"dot\" | \"euclidean\";\n}\n\ninterface Pending {\n\tresolve: (value: EmbSearchRawResponse) => void;\n\treject: (err: Error) => void;\n}\n\ninterface EmbSearchRawResponse {\n\tok: boolean;\n\terror?: string;\n\tresults?: EmbSearchResult[];\n\tinserted?: boolean;\n\tremoved?: boolean;\n\tcount?: number;\n\tinserted_count?: number;\n\tupdated_count?: number;\n\tmodel_id?: string;\n\tdim?: number;\n}\n\nexport class EmbSearchClient {\n\tprivate proc: ChildProcessWithoutNullStreams;\n\tprivate queue: Pending[] = [];\n\tprivate buffer = \"\";\n\tprivate closed = false;\n\tprivate readyPromise: Promise<void>;\n\n\tconstructor(opts: EmbSearchClientOptions) {\n\t\tconst args = [\"serve\", \"--path\", opts.storePath];\n\t\tif (opts.metric) args.push(\"--metric\", opts.metric);\n\n\t\tthis.proc = spawn(opts.binaryPath, args, { stdio: [\"pipe\", \"pipe\", \"pipe\"] });\n\n\t\tthis.proc.stdout.setEncoding(\"utf8\");\n\t\tthis.proc.stdout.on(\"data\", (chunk: string) => this.onStdout(chunk));\n\t\t// Swallow the informational stderr banner; errors surface via responses/exit.\n\t\tthis.proc.stderr.resume();\n\n\t\t// Readiness = the daemon answering a ping (probes the actual request loop).\n\t\tthis.readyPromise = this.send({ op: \"ping\" }).then(() => undefined);\n\t\t// Spawn failures reject the pending ping via the exit handler; mark the\n\t\t// promise handled so a failed spawn doesn't raise an unhandled rejection\n\t\t// before ready() is awaited.\n\t\tthis.readyPromise.catch(() => {});\n\n\t\tthis.proc.on(\"error\", (err) => {\n\t\t\tthis.closed = true;\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t\tthis.proc.on(\"exit\", (code) => {\n\t\t\tthis.closed = true;\n\t\t\tconst err = new Error(`embsearch daemon exited (code ${code})`);\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t}\n\n\t/** Resolves once the daemon has loaded the model + index. */\n\tready(): Promise<void> {\n\t\treturn this.readyPromise;\n\t}\n\n\tget isClosed(): boolean {\n\t\treturn this.closed;\n\t}\n\n\tprivate onStdout(chunk: string): void {\n\t\tthis.buffer += chunk;\n\t\t// Each response is one line; dispatch FIFO against the pending queue.\n\t\tfor (let nl = this.buffer.indexOf(\"\\n\"); nl !== -1; nl = this.buffer.indexOf(\"\\n\")) {\n\t\t\tconst line = this.buffer.slice(0, nl).trim();\n\t\t\tthis.buffer = this.buffer.slice(nl + 1);\n\t\t\tif (!line) continue;\n\t\t\tconst pending = this.queue.shift();\n\t\t\tif (!pending) continue;\n\t\t\tlet msg: EmbSearchRawResponse;\n\t\t\ttry {\n\t\t\t\tmsg = JSON.parse(line) as EmbSearchRawResponse;\n\t\t\t} catch {\n\t\t\t\tpending.reject(new Error(`bad response: ${line}`));\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tif (msg.ok) pending.resolve(msg);\n\t\t\telse pending.reject(new Error(msg.error ?? \"unknown error\"));\n\t\t}\n\t}\n\n\tprivate send(req: Record<string, unknown>): Promise<EmbSearchRawResponse> {\n\t\tif (this.closed) return Promise.reject(new Error(\"embsearch client is closed\"));\n\t\treturn new Promise<EmbSearchRawResponse>((resolve, reject) => {\n\t\t\tthis.queue.push({ resolve, reject });\n\t\t\tthis.proc.stdin.write(`${JSON.stringify(req)}\\n`);\n\t\t});\n\t}\n\n\t/** Search for the top-`k` matches for `text`. */\n\tasync query(text: string, k = 10): Promise<EmbSearchResult[]> {\n\t\tconst res = await this.send({ op: \"query\", text, k });\n\t\treturn res.results ?? [];\n\t}\n\n\t/**\n\t * Batched insert-or-replace. One embedding inference for the whole batch —\n\t * the fast path for bulk indexing. Keep batches modest (e.g. 32–64) so a\n\t * concurrent query is not stuck behind a huge inference.\n\t */\n\tasync bulk(items: Array<{ id: string; text: string }>): Promise<EmbSearchBulkResult> {\n\t\tconst res = await this.send({ op: \"bulk\", items });\n\t\treturn { inserted: res.inserted_count ?? 0, updated: res.updated_count ?? 0 };\n\t}\n\n\t/** Model id, dimensionality, and live vector count of the daemon. */\n\tasync info(): Promise<EmbSearchDaemonInfo> {\n\t\tconst res = await this.send({ op: \"info\" });\n\t\treturn { modelId: res.model_id ?? \"\", dim: res.dim ?? 0, count: res.count ?? 0 };\n\t}\n\n\t/** Remove a record. Resolves to `true` if it existed. */\n\tasync remove(id: string): Promise<boolean> {\n\t\tconst res = await this.send({ op: \"remove\", id });\n\t\treturn res.removed === true;\n\t}\n\n\t/** Reclaim tombstoned rows left behind by `remove`. */\n\tasync compact(): Promise<void> {\n\t\tawait this.send({ op: \"compact\" });\n\t}\n\n\t/** Persist the index to the store directory. */\n\tasync save(): Promise<void> {\n\t\tawait this.send({ op: \"save\" });\n\t}\n\n\t/** Shut the daemon down, closing stdin so it exits cleanly. */\n\tasync close(): Promise<void> {\n\t\tif (this.closed) return;\n\t\tthis.proc.stdin.end();\n\t\tawait new Promise<void>((resolve) => this.proc.on(\"exit\", () => resolve()));\n\t}\n}\n"]}
+{"version":3,"file":"client.d.ts","sourceRoot":"","sources":["../../../src/core/embsearch/client.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAIH,MAAM,WAAW,eAAe;IAC/B,EAAE,EAAE,MAAM,CAAC;IACX,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,mBAAmB;IACnC,OAAO,EAAE,MAAM,CAAC;IAChB,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,mBAAmB;IACnC,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,EAAE,MAAM,CAAC;CAChB;AAED,wEAAwE;AACxE,MAAM,MAAM,eAAe,GAAG,OAAO,GAAG,SAAS,GAAG,QAAQ,CAAC;AAE7D,MAAM,WAAW,sBAAsB;IACtC,6EAA6E;IAC7E,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,sCAAsC;IACtC,UAAU,EAAE,MAAM,CAAC;IACnB,0CAA0C;IAC1C,SAAS,EAAE,MAAM,CAAC;IAClB,6DAA6D;IAC7D,MAAM,CAAC,EAAE,QAAQ,GAAG,KAAK,GAAG,WAAW,CAAC;CACxC;AAoBD,qBAAa,eAAe;IAC3B,OAAO,CAAC,IAAI,CAAiC;IAC7C,OAAO,CAAC,KAAK,CAAiB;IAC9B,OAAO,CAAC,MAAM,CAAM;IACpB,OAAO,CAAC,MAAM,CAAS;IACvB,OAAO,CAAC,YAAY,CAAgB;IAEpC,YAAY,IAAI,EAAE,sBAAsB,EA+BvC;IAED,6DAA6D;IAC7D,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,CAErB;IAED,IAAI,QAAQ,IAAI,OAAO,CAEtB;IAED,OAAO,CAAC,QAAQ;IAqBhB,OAAO,CAAC,IAAI;IAQZ;;;;;;;;;;;;OAYG;IACG,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,CAAC,SAAK,EAAE,SAAS,GAAE,eAAyB,GAAG,OAAO,CAAC,eAAe,EAAE,CAAC,CAKlG;IAED;;;;OAIG;IACG,IAAI,CAAC,KAAK,EAAE,KAAK,CAAC;QAAE,EAAE,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC,GAAG,OAAO,CAAC,mBAAmB,CAAC,CAGnF;IAED,qEAAqE;IAC/D,IAAI,IAAI,OAAO,CAAC,mBAAmB,CAAC,CAGzC;IAED,yDAAyD;IACnD,MAAM,CAAC,EAAE,EAAE,MAAM,GAAG,OAAO,CAAC,OAAO,CAAC,CAGzC;IAED,uDAAuD;IACjD,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAE7B;IAED,gDAAgD;IAC1C,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,CAE1B;IAED,+DAA+D;IACzD,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,CAI3B;CACD","sourcesContent":["/**\n * Client for the `embsearch` stdio daemon (vendored from\n * github.com/kolisachint/embeddingsearchtools ts/client.ts).\n *\n * Spawns `embsearch serve` once and talks newline-delimited JSON over its\n * stdin/stdout. The process stays alive so the model and index load a single\n * time; every query after startup is hot.\n */\n\nimport { type ChildProcessWithoutNullStreams, spawn } from \"child_process\";\n\nexport interface EmbSearchResult {\n\tid: string;\n\tscore: number;\n}\n\nexport interface EmbSearchDaemonInfo {\n\tmodelId: string;\n\tdim: number;\n\tcount: number;\n}\n\nexport interface EmbSearchBulkResult {\n\tinserted: number;\n\tupdated: number;\n}\n\n/** Which daemon-side retriever answers a query (embsearch >= 0.2.0). */\nexport type DaemonRetriever = \"dense\" | \"lexical\" | \"hybrid\";\n\nexport interface EmbSearchClientOptions {\n\t/** Open/create the store with a BM25 lexical index alongside the vectors. */\n\thybrid?: boolean;\n\t/** Path to the `embsearch` binary. */\n\tbinaryPath: string;\n\t/** Store directory passed as `--path`. */\n\tstorePath: string;\n\t/** Metric for a freshly created store. Default: \"cosine\". */\n\tmetric?: \"cosine\" | \"dot\" | \"euclidean\";\n}\n\ninterface Pending {\n\tresolve: (value: EmbSearchRawResponse) => void;\n\treject: (err: Error) => void;\n}\n\ninterface EmbSearchRawResponse {\n\tok: boolean;\n\terror?: string;\n\tresults?: EmbSearchResult[];\n\tinserted?: boolean;\n\tremoved?: boolean;\n\tcount?: number;\n\tinserted_count?: number;\n\tupdated_count?: number;\n\tmodel_id?: string;\n\tdim?: number;\n}\n\nexport class EmbSearchClient {\n\tprivate proc: ChildProcessWithoutNullStreams;\n\tprivate queue: Pending[] = [];\n\tprivate buffer = \"\";\n\tprivate closed = false;\n\tprivate readyPromise: Promise<void>;\n\n\tconstructor(opts: EmbSearchClientOptions) {\n\t\tconst args = [\"serve\", \"--path\", opts.storePath];\n\t\tif (opts.metric) args.push(\"--metric\", opts.metric);\n\t\t// Hybrid-ness is fixed when a store is created: passing --hybrid against\n\t\t// an existing non-hybrid store warns and is ignored daemon-side, so a\n\t\t// hybrid store needs its own directory.\n\t\tif (opts.hybrid) args.push(\"--hybrid\");\n\n\t\tthis.proc = spawn(opts.binaryPath, args, { stdio: [\"pipe\", \"pipe\", \"pipe\"] });\n\n\t\tthis.proc.stdout.setEncoding(\"utf8\");\n\t\tthis.proc.stdout.on(\"data\", (chunk: string) => this.onStdout(chunk));\n\t\t// Swallow the informational stderr banner; errors surface via responses/exit.\n\t\tthis.proc.stderr.resume();\n\n\t\t// Readiness = the daemon answering a ping (probes the actual request loop).\n\t\tthis.readyPromise = this.send({ op: \"ping\" }).then(() => undefined);\n\t\t// Spawn failures reject the pending ping via the exit handler; mark the\n\t\t// promise handled so a failed spawn doesn't raise an unhandled rejection\n\t\t// before ready() is awaited.\n\t\tthis.readyPromise.catch(() => {});\n\n\t\tthis.proc.on(\"error\", (err) => {\n\t\t\tthis.closed = true;\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t\tthis.proc.on(\"exit\", (code) => {\n\t\t\tthis.closed = true;\n\t\t\tconst err = new Error(`embsearch daemon exited (code ${code})`);\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t}\n\n\t/** Resolves once the daemon has loaded the model + index. */\n\tready(): Promise<void> {\n\t\treturn this.readyPromise;\n\t}\n\n\tget isClosed(): boolean {\n\t\treturn this.closed;\n\t}\n\n\tprivate onStdout(chunk: string): void {\n\t\tthis.buffer += chunk;\n\t\t// Each response is one line; dispatch FIFO against the pending queue.\n\t\tfor (let nl = this.buffer.indexOf(\"\\n\"); nl !== -1; nl = this.buffer.indexOf(\"\\n\")) {\n\t\t\tconst line = this.buffer.slice(0, nl).trim();\n\t\t\tthis.buffer = this.buffer.slice(nl + 1);\n\t\t\tif (!line) continue;\n\t\t\tconst pending = this.queue.shift();\n\t\t\tif (!pending) continue;\n\t\t\tlet msg: EmbSearchRawResponse;\n\t\t\ttry {\n\t\t\t\tmsg = JSON.parse(line) as EmbSearchRawResponse;\n\t\t\t} catch {\n\t\t\t\tpending.reject(new Error(`bad response: ${line}`));\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tif (msg.ok) pending.resolve(msg);\n\t\t\telse pending.reject(new Error(msg.error ?? \"unknown error\"));\n\t\t}\n\t}\n\n\tprivate send(req: Record<string, unknown>): Promise<EmbSearchRawResponse> {\n\t\tif (this.closed) return Promise.reject(new Error(\"embsearch client is closed\"));\n\t\treturn new Promise<EmbSearchRawResponse>((resolve, reject) => {\n\t\t\tthis.queue.push({ resolve, reject });\n\t\t\tthis.proc.stdin.write(`${JSON.stringify(req)}\\n`);\n\t\t});\n\t}\n\n\t/**\n\t * Search for the top-`k` matches for `text`.\n\t *\n\t * `retriever` selects which leg answers:\n\t *  - `dense` (default) — vector search, the historical behaviour;\n\t *  - `lexical` — BM25 only, raw scores, no embedding computed;\n\t *  - `hybrid` — both, pre-fused by the daemon's own RRF constant.\n\t *\n\t * `lexical` and `hybrid` need a store created with `--hybrid`. Prefer\n\t * `lexical` over `hybrid` when fusing here: `hybrid` collapses both legs\n\t * into one RRF score, discarding the per-retriever ranks the trace records\n\t * and preventing n-way fusion with the grep leg.\n\t */\n\tasync query(text: string, k = 10, retriever: DaemonRetriever = \"dense\"): Promise<EmbSearchResult[]> {\n\t\tconst res = await this.send(\n\t\t\tretriever === \"dense\" ? { op: \"query\", text, k } : { op: \"query\", text, k, retriever },\n\t\t);\n\t\treturn res.results ?? [];\n\t}\n\n\t/**\n\t * Batched insert-or-replace. One embedding inference for the whole batch —\n\t * the fast path for bulk indexing. Keep batches modest (e.g. 32–64) so a\n\t * concurrent query is not stuck behind a huge inference.\n\t */\n\tasync bulk(items: Array<{ id: string; text: string }>): Promise<EmbSearchBulkResult> {\n\t\tconst res = await this.send({ op: \"bulk\", items });\n\t\treturn { inserted: res.inserted_count ?? 0, updated: res.updated_count ?? 0 };\n\t}\n\n\t/** Model id, dimensionality, and live vector count of the daemon. */\n\tasync info(): Promise<EmbSearchDaemonInfo> {\n\t\tconst res = await this.send({ op: \"info\" });\n\t\treturn { modelId: res.model_id ?? \"\", dim: res.dim ?? 0, count: res.count ?? 0 };\n\t}\n\n\t/** Remove a record. Resolves to `true` if it existed. */\n\tasync remove(id: string): Promise<boolean> {\n\t\tconst res = await this.send({ op: \"remove\", id });\n\t\treturn res.removed === true;\n\t}\n\n\t/** Reclaim tombstoned rows left behind by `remove`. */\n\tasync compact(): Promise<void> {\n\t\tawait this.send({ op: \"compact\" });\n\t}\n\n\t/** Persist the index to the store directory. */\n\tasync save(): Promise<void> {\n\t\tawait this.send({ op: \"save\" });\n\t}\n\n\t/** Shut the daemon down, closing stdin so it exits cleanly. */\n\tasync close(): Promise<void> {\n\t\tif (this.closed) return;\n\t\tthis.proc.stdin.end();\n\t\tawait new Promise<void>((resolve) => this.proc.on(\"exit\", () => resolve()));\n\t}\n}\n"]}

package/dist/core/embsearch/client.js

@@ -17,6 +17,11 @@ export class EmbSearchClient {
         const args = ["serve", "--path", opts.storePath];
         if (opts.metric)
             args.push("--metric", opts.metric);
+        // Hybrid-ness is fixed when a store is created: passing --hybrid against
+        // an existing non-hybrid store warns and is ignored daemon-side, so a
+        // hybrid store needs its own directory.
+        if (opts.hybrid)
+            args.push("--hybrid");
         this.proc = spawn(opts.binaryPath, args, { stdio: ["pipe", "pipe", "pipe"] });
         this.proc.stdout.setEncoding("utf8");
         this.proc.stdout.on("data", (chunk) => this.onStdout(chunk));
@@ -80,9 +85,21 @@ export class EmbSearchClient {
             this.proc.stdin.write(`${JSON.stringify(req)}\n`);
         });
     }
-    /** Search for the top-`k` matches for `text`. */
-    async query(text, k = 10) {
-        const res = await this.send({ op: "query", text, k });
+    /**
+     * Search for the top-`k` matches for `text`.
+     *
+     * `retriever` selects which leg answers:
+     *  - `dense` (default) — vector search, the historical behaviour;
+     *  - `lexical` — BM25 only, raw scores, no embedding computed;
+     *  - `hybrid` — both, pre-fused by the daemon's own RRF constant.
+     *
+     * `lexical` and `hybrid` need a store created with `--hybrid`. Prefer
+     * `lexical` over `hybrid` when fusing here: `hybrid` collapses both legs
+     * into one RRF score, discarding the per-retriever ranks the trace records
+     * and preventing n-way fusion with the grep leg.
+     */
+    async query(text, k = 10, retriever = "dense") {
+        const res = await this.send(retriever === "dense" ? { op: "query", text, k } : { op: "query", text, k, retriever });
         return res.results ?? [];
     }
     /**

package/dist/core/embsearch/client.js.map

@@ -1 +1 @@
-{"version":3,"file":"client.js","sourceRoot":"","sources":["../../../src/core/embsearch/client.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAuC,KAAK,EAAE,MAAM,eAAe,CAAC;AA6C3E,MAAM,OAAO,eAAe;IACnB,IAAI,CAAiC;IACrC,KAAK,GAAc,EAAE,CAAC;IACtB,MAAM,GAAG,EAAE,CAAC;IACZ,MAAM,GAAG,KAAK,CAAC;IACf,YAAY,CAAgB;IAEpC,YAAY,IAA4B,EAAE;QACzC,MAAM,IAAI,GAAG,CAAC,OAAO,EAAE,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC;QACjD,IAAI,IAAI,CAAC,MAAM;YAAE,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;QAEpD,IAAI,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;QAE9E,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC;QACrC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,KAAa,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC;QACrE,8EAA8E;QAC9E,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC;QAE1B,4EAA4E;QAC5E,IAAI,CAAC,YAAY,GAAG,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;QACpE,wEAAwE;QACxE,yEAAyE;QACzE,6BAA6B;QAC7B,IAAI,CAAC,YAAY,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAC,CAAC,CAAC,CAAC;QAElC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC;YAC9B,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;YACnB,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;gBAAE,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAAA,CACpD,CAAC,CAAC;QACH,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC;YAC9B,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;YACnB,MAAM,GAAG,GAAG,IAAI,KAAK,CAAC,iCAAiC,IAAI,GAAG,CAAC,CAAC;YAChE,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;gBAAE,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAAA,CACpD,CAAC,CAAC;IAAA,CACH;IAED,6DAA6D;IAC7D,KAAK,GAAkB;QACtB,OAAO,IAAI,CAAC,YAAY,CAAC;IAAA,CACzB;IAED,IAAI,QAAQ,GAAY;QACvB,OAAO,IAAI,CAAC,MAAM,CAAC;IAAA,CACnB;IAEO,QAAQ,CAAC,KAAa,EAAQ;QACrC,IAAI,CAAC,MAAM,IAAI,KAAK,CAAC;QACrB,sEAAsE;QACtE,KAAK,IAAI,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC;YACpF,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC7C,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;YACxC,IAAI,CAAC,IAAI;gBAAE,SAAS;YACpB,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC;YACnC,IAAI,CAAC,OAAO;gBAAE,SAAS;YACvB,IAAI,GAAyB,CAAC;YAC9B,IAAI,CAAC;gBACJ,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAyB,CAAC;YAChD,CAAC;YAAC,MAAM,CAAC;gBACR,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,iBAAiB,IAAI,EAAE,CAAC,CAAC,CAAC;gBACnD,SAAS;YACV,CAAC;YACD,IAAI,GAAG,CAAC,EAAE;gBAAE,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;;gBAC5B,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,GAAG,CAAC,KAAK,IAAI,eAAe,CAAC,CAAC,CAAC;QAC9D,CAAC;IAAA,CACD;IAEO,IAAI,CAAC,GAA4B,EAAiC;QACzE,IAAI,IAAI,CAAC,MAAM;YAAE,OAAO,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,4BAA4B,CAAC,CAAC,CAAC;QAChF,OAAO,IAAI,OAAO,CAAuB,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE,CAAC;YAC7D,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,CAAC;YACrC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAAA,CAClD,CAAC,CAAC;IAAA,CACH;IAED,iDAAiD;IACjD,KAAK,CAAC,KAAK,CAAC,IAAY,EAAE,CAAC,GAAG,EAAE,EAA8B;QAC7D,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC;QACtD,OAAO,GAAG,CAAC,OAAO,IAAI,EAAE,CAAC;IAAA,CACzB;IAED;;;;OAIG;IACH,KAAK,CAAC,IAAI,CAAC,KAA0C,EAAgC;QACpF,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;QACnD,OAAO,EAAE,QAAQ,EAAE,GAAG,CAAC,cAAc,IAAI,CAAC,EAAE,OAAO,EAAE,GAAG,CAAC,aAAa,IAAI,CAAC,EAAE,CAAC;IAAA,CAC9E;IAED,qEAAqE;IACrE,KAAK,CAAC,IAAI,GAAiC;QAC1C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;QAC5C,OAAO,EAAE,OAAO,EAAE,GAAG,CAAC,QAAQ,IAAI,EAAE,EAAE,GAAG,EAAE,GAAG,CAAC,GAAG,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,IAAI,CAAC,EAAE,CAAC;IAAA,CACjF;IAED,yDAAyD;IACzD,KAAK,CAAC,MAAM,CAAC,EAAU,EAAoB;QAC1C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,CAAC;QAClD,OAAO,GAAG,CAAC,OAAO,KAAK,IAAI,CAAC;IAAA,CAC5B;IAED,uDAAuD;IACvD,KAAK,CAAC,OAAO,GAAkB;QAC9B,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,SAAS,EAAE,CAAC,CAAC;IAAA,CACnC;IAED,gDAAgD;IAChD,KAAK,CAAC,IAAI,GAAkB;QAC3B,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;IAAA,CAChC;IAED,+DAA+D;IAC/D,KAAK,CAAC,KAAK,GAAkB;QAC5B,IAAI,IAAI,CAAC,MAAM;YAAE,OAAO;QACxB,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC;QACtB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,MAAM,EAAE,GAAG,EAAE,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IAAA,CAC5E;CACD","sourcesContent":["/**\n * Client for the `embsearch` stdio daemon (vendored from\n * github.com/kolisachint/embeddingsearchtools ts/client.ts).\n *\n * Spawns `embsearch serve` once and talks newline-delimited JSON over its\n * stdin/stdout. The process stays alive so the model and index load a single\n * time; every query after startup is hot.\n */\n\nimport { type ChildProcessWithoutNullStreams, spawn } from \"child_process\";\n\nexport interface EmbSearchResult {\n\tid: string;\n\tscore: number;\n}\n\nexport interface EmbSearchDaemonInfo {\n\tmodelId: string;\n\tdim: number;\n\tcount: number;\n}\n\nexport interface EmbSearchBulkResult {\n\tinserted: number;\n\tupdated: number;\n}\n\nexport interface EmbSearchClientOptions {\n\t/** Path to the `embsearch` binary. */\n\tbinaryPath: string;\n\t/** Store directory passed as `--path`. */\n\tstorePath: string;\n\t/** Metric for a freshly created store. Default: \"cosine\". */\n\tmetric?: \"cosine\" | \"dot\" | \"euclidean\";\n}\n\ninterface Pending {\n\tresolve: (value: EmbSearchRawResponse) => void;\n\treject: (err: Error) => void;\n}\n\ninterface EmbSearchRawResponse {\n\tok: boolean;\n\terror?: string;\n\tresults?: EmbSearchResult[];\n\tinserted?: boolean;\n\tremoved?: boolean;\n\tcount?: number;\n\tinserted_count?: number;\n\tupdated_count?: number;\n\tmodel_id?: string;\n\tdim?: number;\n}\n\nexport class EmbSearchClient {\n\tprivate proc: ChildProcessWithoutNullStreams;\n\tprivate queue: Pending[] = [];\n\tprivate buffer = \"\";\n\tprivate closed = false;\n\tprivate readyPromise: Promise<void>;\n\n\tconstructor(opts: EmbSearchClientOptions) {\n\t\tconst args = [\"serve\", \"--path\", opts.storePath];\n\t\tif (opts.metric) args.push(\"--metric\", opts.metric);\n\n\t\tthis.proc = spawn(opts.binaryPath, args, { stdio: [\"pipe\", \"pipe\", \"pipe\"] });\n\n\t\tthis.proc.stdout.setEncoding(\"utf8\");\n\t\tthis.proc.stdout.on(\"data\", (chunk: string) => this.onStdout(chunk));\n\t\t// Swallow the informational stderr banner; errors surface via responses/exit.\n\t\tthis.proc.stderr.resume();\n\n\t\t// Readiness = the daemon answering a ping (probes the actual request loop).\n\t\tthis.readyPromise = this.send({ op: \"ping\" }).then(() => undefined);\n\t\t// Spawn failures reject the pending ping via the exit handler; mark the\n\t\t// promise handled so a failed spawn doesn't raise an unhandled rejection\n\t\t// before ready() is awaited.\n\t\tthis.readyPromise.catch(() => {});\n\n\t\tthis.proc.on(\"error\", (err) => {\n\t\t\tthis.closed = true;\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t\tthis.proc.on(\"exit\", (code) => {\n\t\t\tthis.closed = true;\n\t\t\tconst err = new Error(`embsearch daemon exited (code ${code})`);\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t}\n\n\t/** Resolves once the daemon has loaded the model + index. */\n\tready(): Promise<void> {\n\t\treturn this.readyPromise;\n\t}\n\n\tget isClosed(): boolean {\n\t\treturn this.closed;\n\t}\n\n\tprivate onStdout(chunk: string): void {\n\t\tthis.buffer += chunk;\n\t\t// Each response is one line; dispatch FIFO against the pending queue.\n\t\tfor (let nl = this.buffer.indexOf(\"\\n\"); nl !== -1; nl = this.buffer.indexOf(\"\\n\")) {\n\t\t\tconst line = this.buffer.slice(0, nl).trim();\n\t\t\tthis.buffer = this.buffer.slice(nl + 1);\n\t\t\tif (!line) continue;\n\t\t\tconst pending = this.queue.shift();\n\t\t\tif (!pending) continue;\n\t\t\tlet msg: EmbSearchRawResponse;\n\t\t\ttry {\n\t\t\t\tmsg = JSON.parse(line) as EmbSearchRawResponse;\n\t\t\t} catch {\n\t\t\t\tpending.reject(new Error(`bad response: ${line}`));\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tif (msg.ok) pending.resolve(msg);\n\t\t\telse pending.reject(new Error(msg.error ?? \"unknown error\"));\n\t\t}\n\t}\n\n\tprivate send(req: Record<string, unknown>): Promise<EmbSearchRawResponse> {\n\t\tif (this.closed) return Promise.reject(new Error(\"embsearch client is closed\"));\n\t\treturn new Promise<EmbSearchRawResponse>((resolve, reject) => {\n\t\t\tthis.queue.push({ resolve, reject });\n\t\t\tthis.proc.stdin.write(`${JSON.stringify(req)}\\n`);\n\t\t});\n\t}\n\n\t/** Search for the top-`k` matches for `text`. */\n\tasync query(text: string, k = 10): Promise<EmbSearchResult[]> {\n\t\tconst res = await this.send({ op: \"query\", text, k });\n\t\treturn res.results ?? [];\n\t}\n\n\t/**\n\t * Batched insert-or-replace. One embedding inference for the whole batch —\n\t * the fast path for bulk indexing. Keep batches modest (e.g. 32–64) so a\n\t * concurrent query is not stuck behind a huge inference.\n\t */\n\tasync bulk(items: Array<{ id: string; text: string }>): Promise<EmbSearchBulkResult> {\n\t\tconst res = await this.send({ op: \"bulk\", items });\n\t\treturn { inserted: res.inserted_count ?? 0, updated: res.updated_count ?? 0 };\n\t}\n\n\t/** Model id, dimensionality, and live vector count of the daemon. */\n\tasync info(): Promise<EmbSearchDaemonInfo> {\n\t\tconst res = await this.send({ op: \"info\" });\n\t\treturn { modelId: res.model_id ?? \"\", dim: res.dim ?? 0, count: res.count ?? 0 };\n\t}\n\n\t/** Remove a record. Resolves to `true` if it existed. */\n\tasync remove(id: string): Promise<boolean> {\n\t\tconst res = await this.send({ op: \"remove\", id });\n\t\treturn res.removed === true;\n\t}\n\n\t/** Reclaim tombstoned rows left behind by `remove`. */\n\tasync compact(): Promise<void> {\n\t\tawait this.send({ op: \"compact\" });\n\t}\n\n\t/** Persist the index to the store directory. */\n\tasync save(): Promise<void> {\n\t\tawait this.send({ op: \"save\" });\n\t}\n\n\t/** Shut the daemon down, closing stdin so it exits cleanly. */\n\tasync close(): Promise<void> {\n\t\tif (this.closed) return;\n\t\tthis.proc.stdin.end();\n\t\tawait new Promise<void>((resolve) => this.proc.on(\"exit\", () => resolve()));\n\t}\n}\n"]}
+{"version":3,"file":"client.js","sourceRoot":"","sources":["../../../src/core/embsearch/client.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAuC,KAAK,EAAE,MAAM,eAAe,CAAC;AAkD3E,MAAM,OAAO,eAAe;IACnB,IAAI,CAAiC;IACrC,KAAK,GAAc,EAAE,CAAC;IACtB,MAAM,GAAG,EAAE,CAAC;IACZ,MAAM,GAAG,KAAK,CAAC;IACf,YAAY,CAAgB;IAEpC,YAAY,IAA4B,EAAE;QACzC,MAAM,IAAI,GAAG,CAAC,OAAO,EAAE,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC;QACjD,IAAI,IAAI,CAAC,MAAM;YAAE,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC;QACpD,yEAAyE;QACzE,sEAAsE;QACtE,wCAAwC;QACxC,IAAI,IAAI,CAAC,MAAM;YAAE,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;QAEvC,IAAI,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC;QAE9E,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC;QACrC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,KAAa,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC;QACrE,8EAA8E;QAC9E,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC;QAE1B,4EAA4E;QAC5E,IAAI,CAAC,YAAY,GAAG,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;QACpE,wEAAwE;QACxE,yEAAyE;QACzE,6BAA6B;QAC7B,IAAI,CAAC,YAAY,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAC,CAAC,CAAC,CAAC;QAElC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,OAAO,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC;YAC9B,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;YACnB,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;gBAAE,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAAA,CACpD,CAAC,CAAC;QACH,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,MAAM,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC;YAC9B,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC;YACnB,MAAM,GAAG,GAAG,IAAI,KAAK,CAAC,iCAAiC,IAAI,GAAG,CAAC,CAAC;YAChE,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;gBAAE,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAAA,CACpD,CAAC,CAAC;IAAA,CACH;IAED,6DAA6D;IAC7D,KAAK,GAAkB;QACtB,OAAO,IAAI,CAAC,YAAY,CAAC;IAAA,CACzB;IAED,IAAI,QAAQ,GAAY;QACvB,OAAO,IAAI,CAAC,MAAM,CAAC;IAAA,CACnB;IAEO,QAAQ,CAAC,KAAa,EAAQ;QACrC,IAAI,CAAC,MAAM,IAAI,KAAK,CAAC;QACrB,sEAAsE;QACtE,KAAK,IAAI,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC;YACpF,MAAM,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;YAC7C,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC;YACxC,IAAI,CAAC,IAAI;gBAAE,SAAS;YACpB,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC;YACnC,IAAI,CAAC,OAAO;gBAAE,SAAS;YACvB,IAAI,GAAyB,CAAC;YAC9B,IAAI,CAAC;gBACJ,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAyB,CAAC;YAChD,CAAC;YAAC,MAAM,CAAC;gBACR,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,iBAAiB,IAAI,EAAE,CAAC,CAAC,CAAC;gBACnD,SAAS;YACV,CAAC;YACD,IAAI,GAAG,CAAC,EAAE;gBAAE,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;;gBAC5B,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,GAAG,CAAC,KAAK,IAAI,eAAe,CAAC,CAAC,CAAC;QAC9D,CAAC;IAAA,CACD;IAEO,IAAI,CAAC,GAA4B,EAAiC;QACzE,IAAI,IAAI,CAAC,MAAM;YAAE,OAAO,OAAO,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,4BAA4B,CAAC,CAAC,CAAC;QAChF,OAAO,IAAI,OAAO,CAAuB,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE,CAAC;YAC7D,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,CAAC;YACrC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAAA,CAClD,CAAC,CAAC;IAAA,CACH;IAED;;;;;;;;;;;;OAYG;IACH,KAAK,CAAC,KAAK,CAAC,IAAY,EAAE,CAAC,GAAG,EAAE,EAAE,SAAS,GAAoB,OAAO,EAA8B;QACnG,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAC1B,SAAS,KAAK,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,EAAE,SAAS,EAAE,CACtF,CAAC;QACF,OAAO,GAAG,CAAC,OAAO,IAAI,EAAE,CAAC;IAAA,CACzB;IAED;;;;OAIG;IACH,KAAK,CAAC,IAAI,CAAC,KAA0C,EAAgC;QACpF,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;QACnD,OAAO,EAAE,QAAQ,EAAE,GAAG,CAAC,cAAc,IAAI,CAAC,EAAE,OAAO,EAAE,GAAG,CAAC,aAAa,IAAI,CAAC,EAAE,CAAC;IAAA,CAC9E;IAED,qEAAqE;IACrE,KAAK,CAAC,IAAI,GAAiC;QAC1C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;QAC5C,OAAO,EAAE,OAAO,EAAE,GAAG,CAAC,QAAQ,IAAI,EAAE,EAAE,GAAG,EAAE,GAAG,CAAC,GAAG,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,IAAI,CAAC,EAAE,CAAC;IAAA,CACjF;IAED,yDAAyD;IACzD,KAAK,CAAC,MAAM,CAAC,EAAU,EAAoB;QAC1C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,EAAE,EAAE,CAAC,CAAC;QAClD,OAAO,GAAG,CAAC,OAAO,KAAK,IAAI,CAAC;IAAA,CAC5B;IAED,uDAAuD;IACvD,KAAK,CAAC,OAAO,GAAkB;QAC9B,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,SAAS,EAAE,CAAC,CAAC;IAAA,CACnC;IAED,gDAAgD;IAChD,KAAK,CAAC,IAAI,GAAkB;QAC3B,MAAM,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,CAAC;IAAA,CAChC;IAED,+DAA+D;IAC/D,KAAK,CAAC,KAAK,GAAkB;QAC5B,IAAI,IAAI,CAAC,MAAM;YAAE,OAAO;QACxB,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC;QACtB,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,MAAM,EAAE,GAAG,EAAE,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IAAA,CAC5E;CACD","sourcesContent":["/**\n * Client for the `embsearch` stdio daemon (vendored from\n * github.com/kolisachint/embeddingsearchtools ts/client.ts).\n *\n * Spawns `embsearch serve` once and talks newline-delimited JSON over its\n * stdin/stdout. The process stays alive so the model and index load a single\n * time; every query after startup is hot.\n */\n\nimport { type ChildProcessWithoutNullStreams, spawn } from \"child_process\";\n\nexport interface EmbSearchResult {\n\tid: string;\n\tscore: number;\n}\n\nexport interface EmbSearchDaemonInfo {\n\tmodelId: string;\n\tdim: number;\n\tcount: number;\n}\n\nexport interface EmbSearchBulkResult {\n\tinserted: number;\n\tupdated: number;\n}\n\n/** Which daemon-side retriever answers a query (embsearch >= 0.2.0). */\nexport type DaemonRetriever = \"dense\" | \"lexical\" | \"hybrid\";\n\nexport interface EmbSearchClientOptions {\n\t/** Open/create the store with a BM25 lexical index alongside the vectors. */\n\thybrid?: boolean;\n\t/** Path to the `embsearch` binary. */\n\tbinaryPath: string;\n\t/** Store directory passed as `--path`. */\n\tstorePath: string;\n\t/** Metric for a freshly created store. Default: \"cosine\". */\n\tmetric?: \"cosine\" | \"dot\" | \"euclidean\";\n}\n\ninterface Pending {\n\tresolve: (value: EmbSearchRawResponse) => void;\n\treject: (err: Error) => void;\n}\n\ninterface EmbSearchRawResponse {\n\tok: boolean;\n\terror?: string;\n\tresults?: EmbSearchResult[];\n\tinserted?: boolean;\n\tremoved?: boolean;\n\tcount?: number;\n\tinserted_count?: number;\n\tupdated_count?: number;\n\tmodel_id?: string;\n\tdim?: number;\n}\n\nexport class EmbSearchClient {\n\tprivate proc: ChildProcessWithoutNullStreams;\n\tprivate queue: Pending[] = [];\n\tprivate buffer = \"\";\n\tprivate closed = false;\n\tprivate readyPromise: Promise<void>;\n\n\tconstructor(opts: EmbSearchClientOptions) {\n\t\tconst args = [\"serve\", \"--path\", opts.storePath];\n\t\tif (opts.metric) args.push(\"--metric\", opts.metric);\n\t\t// Hybrid-ness is fixed when a store is created: passing --hybrid against\n\t\t// an existing non-hybrid store warns and is ignored daemon-side, so a\n\t\t// hybrid store needs its own directory.\n\t\tif (opts.hybrid) args.push(\"--hybrid\");\n\n\t\tthis.proc = spawn(opts.binaryPath, args, { stdio: [\"pipe\", \"pipe\", \"pipe\"] });\n\n\t\tthis.proc.stdout.setEncoding(\"utf8\");\n\t\tthis.proc.stdout.on(\"data\", (chunk: string) => this.onStdout(chunk));\n\t\t// Swallow the informational stderr banner; errors surface via responses/exit.\n\t\tthis.proc.stderr.resume();\n\n\t\t// Readiness = the daemon answering a ping (probes the actual request loop).\n\t\tthis.readyPromise = this.send({ op: \"ping\" }).then(() => undefined);\n\t\t// Spawn failures reject the pending ping via the exit handler; mark the\n\t\t// promise handled so a failed spawn doesn't raise an unhandled rejection\n\t\t// before ready() is awaited.\n\t\tthis.readyPromise.catch(() => {});\n\n\t\tthis.proc.on(\"error\", (err) => {\n\t\t\tthis.closed = true;\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t\tthis.proc.on(\"exit\", (code) => {\n\t\t\tthis.closed = true;\n\t\t\tconst err = new Error(`embsearch daemon exited (code ${code})`);\n\t\t\tfor (const p of this.queue.splice(0)) p.reject(err);\n\t\t});\n\t}\n\n\t/** Resolves once the daemon has loaded the model + index. */\n\tready(): Promise<void> {\n\t\treturn this.readyPromise;\n\t}\n\n\tget isClosed(): boolean {\n\t\treturn this.closed;\n\t}\n\n\tprivate onStdout(chunk: string): void {\n\t\tthis.buffer += chunk;\n\t\t// Each response is one line; dispatch FIFO against the pending queue.\n\t\tfor (let nl = this.buffer.indexOf(\"\\n\"); nl !== -1; nl = this.buffer.indexOf(\"\\n\")) {\n\t\t\tconst line = this.buffer.slice(0, nl).trim();\n\t\t\tthis.buffer = this.buffer.slice(nl + 1);\n\t\t\tif (!line) continue;\n\t\t\tconst pending = this.queue.shift();\n\t\t\tif (!pending) continue;\n\t\t\tlet msg: EmbSearchRawResponse;\n\t\t\ttry {\n\t\t\t\tmsg = JSON.parse(line) as EmbSearchRawResponse;\n\t\t\t} catch {\n\t\t\t\tpending.reject(new Error(`bad response: ${line}`));\n\t\t\t\tcontinue;\n\t\t\t}\n\t\t\tif (msg.ok) pending.resolve(msg);\n\t\t\telse pending.reject(new Error(msg.error ?? \"unknown error\"));\n\t\t}\n\t}\n\n\tprivate send(req: Record<string, unknown>): Promise<EmbSearchRawResponse> {\n\t\tif (this.closed) return Promise.reject(new Error(\"embsearch client is closed\"));\n\t\treturn new Promise<EmbSearchRawResponse>((resolve, reject) => {\n\t\t\tthis.queue.push({ resolve, reject });\n\t\t\tthis.proc.stdin.write(`${JSON.stringify(req)}\\n`);\n\t\t});\n\t}\n\n\t/**\n\t * Search for the top-`k` matches for `text`.\n\t *\n\t * `retriever` selects which leg answers:\n\t *  - `dense` (default) — vector search, the historical behaviour;\n\t *  - `lexical` — BM25 only, raw scores, no embedding computed;\n\t *  - `hybrid` — both, pre-fused by the daemon's own RRF constant.\n\t *\n\t * `lexical` and `hybrid` need a store created with `--hybrid`. Prefer\n\t * `lexical` over `hybrid` when fusing here: `hybrid` collapses both legs\n\t * into one RRF score, discarding the per-retriever ranks the trace records\n\t * and preventing n-way fusion with the grep leg.\n\t */\n\tasync query(text: string, k = 10, retriever: DaemonRetriever = \"dense\"): Promise<EmbSearchResult[]> {\n\t\tconst res = await this.send(\n\t\t\tretriever === \"dense\" ? { op: \"query\", text, k } : { op: \"query\", text, k, retriever },\n\t\t);\n\t\treturn res.results ?? [];\n\t}\n\n\t/**\n\t * Batched insert-or-replace. One embedding inference for the whole batch —\n\t * the fast path for bulk indexing. Keep batches modest (e.g. 32–64) so a\n\t * concurrent query is not stuck behind a huge inference.\n\t */\n\tasync bulk(items: Array<{ id: string; text: string }>): Promise<EmbSearchBulkResult> {\n\t\tconst res = await this.send({ op: \"bulk\", items });\n\t\treturn { inserted: res.inserted_count ?? 0, updated: res.updated_count ?? 0 };\n\t}\n\n\t/** Model id, dimensionality, and live vector count of the daemon. */\n\tasync info(): Promise<EmbSearchDaemonInfo> {\n\t\tconst res = await this.send({ op: \"info\" });\n\t\treturn { modelId: res.model_id ?? \"\", dim: res.dim ?? 0, count: res.count ?? 0 };\n\t}\n\n\t/** Remove a record. Resolves to `true` if it existed. */\n\tasync remove(id: string): Promise<boolean> {\n\t\tconst res = await this.send({ op: \"remove\", id });\n\t\treturn res.removed === true;\n\t}\n\n\t/** Reclaim tombstoned rows left behind by `remove`. */\n\tasync compact(): Promise<void> {\n\t\tawait this.send({ op: \"compact\" });\n\t}\n\n\t/** Persist the index to the store directory. */\n\tasync save(): Promise<void> {\n\t\tawait this.send({ op: \"save\" });\n\t}\n\n\t/** Shut the daemon down, closing stdin so it exits cleanly. */\n\tasync close(): Promise<void> {\n\t\tif (this.closed) return;\n\t\tthis.proc.stdin.end();\n\t\tawait new Promise<void>((resolve) => this.proc.on(\"exit\", () => resolve()));\n\t}\n}\n"]}

package/dist/core/embsearch/embsearch-service.d.ts

@@ -13,6 +13,7 @@
  * Every failure degrades to `unavailable` with a reason; nothing here ever
  * blocks session startup or affects grep/find.
  */
+import { type DaemonRetriever } from "./client.js";
 export type EmbsearchState = {
     phase: "idle";
 } | {
@@ -39,6 +40,17 @@ export interface EmbsearchServiceOptions {
     binaryPath?: string;
     /** Minimum indexable bytes before indexing kicks in. */
     thresholdBytes: number;
+    /**
+     * Override the store location. Only the eval harness sets this, so a
+     * second index (e.g. a BM25-hybrid store) can exist for the same repo
+     * without colliding with the primary one.
+     */
+    storeDir?: string;
+    /**
+     * Create the store with the daemon's BM25 lexical index. Fixed at store
+     * creation, so this must pair with a distinct `storeDir`.
+     */
+    hybridStore?: boolean;
     /** Progress callback for UI (footer / stderr lines). */
     onProgress?: (state: EmbsearchState) => void;
 }
@@ -58,11 +70,16 @@ export declare class EmbsearchService {
     private meta;
     private state;
     private disposed;
+    /** Whether the resolved binary serves `retriever: "lexical"`. */
+    private lexicalRetriever;
     constructor(options: EmbsearchServiceOptions);
     getState(): EmbsearchState;
     /** Semantic search is usable (index ready, or still building with partial data). */
     isAvailable(): boolean;
     private setState;
+    /** Whether the running daemon can serve a BM25-only query. */
+    supportsLexicalRetriever(): boolean;
+    private probeLexicalRetrieverSupport;
     private resolveBinary;
     /**
      * Scan, threshold-check, and (when needed) index in the background.
@@ -75,7 +92,7 @@ export declare class EmbsearchService {
     /** Top-`k` semantic hits as `path` + line range + score. */
     search(query: string, k?: number): Promise<SemanticHit[]>;
     /** Top-`k` semantic hits including their chunk ids, for rank fusion. */
-    searchChunks(query: string, k?: number, glob?: string): Promise<SemanticChunkHit[]>;
+    searchChunks(query: string, k?: number, glob?: string, retriever?: DaemonRetriever): Promise<SemanticChunkHit[]>;
     /**
      * Resolve a repo-relative path + line to its enclosing indexed chunk, or
      * undefined when the file/line is not covered by the index. Chunks overlap

package/dist/core/embsearch/embsearch-service.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"embsearch-service.d.ts","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AA2BH,MAAM,MAAM,cAAc,GACvB;IAAE,KAAK,EAAE,MAAM,CAAA;CAAE,GACjB;IAAE,KAAK,EAAE,SAAS,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,GACpC;IAAE,KAAK,EAAE,aAAa,CAAC;IAAC,aAAa,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,GAAG,IAAI,CAAA;CAAE,GAC1E;IAAE,KAAK,EAAE,UAAU,CAAC;IAAC,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,GAClD;IAAE,KAAK,EAAE,OAAO,CAAC;IAAC,UAAU,EAAE,MAAM,CAAA;CAAE,GACtC;IAAE,KAAK,EAAE,aAAa,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAAC;AAE5C,MAAM,WAAW,uBAAuB;IACvC,GAAG,EAAE,MAAM,CAAC;IACZ,gFAAgF;IAChF,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,wDAAwD;IACxD,cAAc,EAAE,MAAM,CAAC;IACvB,wDAAwD;IACxD,UAAU,CAAC,EAAE,CAAC,KAAK,EAAE,cAAc,KAAK,IAAI,CAAC;CAC7C;AAED,MAAM,WAAW,WAAW;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,gBAAiB,SAAQ,WAAW;IACpD,sFAAoF;IACpF,EAAE,EAAE,MAAM,CAAC;CACX;AAED,qBAAa,gBAAgB;IAC5B,OAAO,CAAC,QAAQ,CAAC,OAAO,CAA0B;IAClD,OAAO,CAAC,MAAM,CAA8B;IAC5C,OAAO,CAAC,IAAI,CAAwB;IACpC,OAAO,CAAC,KAAK,CAAqC;IAClD,OAAO,CAAC,QAAQ,CAAS;IAEzB,YAAY,OAAO,EAAE,uBAAuB,EAE3C;IAED,QAAQ,IAAI,cAAc,CAEzB;IAED,oFAAoF;IACpF,WAAW,IAAI,OAAO,CAErB;IAED,OAAO,CAAC,QAAQ;YAKF,aAAa;IAY3B;;;OAGG;IACG,KAAK,CAAC,MAAM,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,IAAI,CAAC,CAQ/C;YAEa,GAAG;YAmCH,iBAAiB;IAqF/B,OAAO,CAAC,WAAW;IAMnB,4DAA4D;IACtD,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,SAAK,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC,CAE1D;IAED,wEAAwE;IAClE,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,SAAK,EAAE,IAAI,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,gBAAgB,EAAE,CAAC,CAqBpF;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAG,EAAE,MAAM,EACX,IAAI,EAAE,MAAM,GACV;QAAE,EAAE,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAU9E;YAEa,WAAW;IAYzB,kEAAkE;IAC5D,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAW7B;CACD;AAWD,wBAAgB,wBAAwB,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,EAAE,gBAAgB,GAAG,IAAI,CAMrF;AAED,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,MAAM,GAAG,gBAAgB,GAAG,SAAS,CAE7E;AAED,wBAAgB,0BAA0B,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,CAE5D","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.d.ts","sourceRoot":"","sources":["../../../src/core/embsearch/embsearch-service.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAOH,OAAO,EAAE,KAAK,eAAe,EAAmB,MAAM,aAAa,CAAC;AA4CpE,MAAM,MAAM,cAAc,GACvB;IAAE,KAAK,EAAE,MAAM,CAAA;CAAE,GACjB;IAAE,KAAK,EAAE,SAAS,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,GACpC;IAAE,KAAK,EAAE,aAAa,CAAC;IAAC,aAAa,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,GAAG,IAAI,CAAA;CAAE,GAC1E;IAAE,KAAK,EAAE,UAAU,CAAC;IAAC,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,GAClD;IAAE,KAAK,EAAE,OAAO,CAAC;IAAC,UAAU,EAAE,MAAM,CAAA;CAAE,GACtC;IAAE,KAAK,EAAE,aAAa,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,CAAC;AAE5C,MAAM,WAAW,uBAAuB;IACvC,GAAG,EAAE,MAAM,CAAC;IACZ,gFAAgF;IAChF,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,wDAAwD;IACxD,cAAc,EAAE,MAAM,CAAC;IACvB;;;;OAIG;IACH,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB;;;OAGG;IACH,WAAW,CAAC,EAAE,OAAO,CAAC;IACtB,wDAAwD;IACxD,UAAU,CAAC,EAAE,CAAC,KAAK,EAAE,cAAc,KAAK,IAAI,CAAC;CAC7C;AAED,MAAM,WAAW,WAAW;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,KAAK,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,gBAAiB,SAAQ,WAAW;IACpD,sFAAoF;IACpF,EAAE,EAAE,MAAM,CAAC;CACX;AAED,qBAAa,gBAAgB;IAC5B,OAAO,CAAC,QAAQ,CAAC,OAAO,CAA0B;IAClD,OAAO,CAAC,MAAM,CAA8B;IAC5C,OAAO,CAAC,IAAI,CAAwB;IACpC,OAAO,CAAC,KAAK,CAAqC;IAClD,OAAO,CAAC,QAAQ,CAAS;IACzB,iEAAiE;IACjE,OAAO,CAAC,gBAAgB,CAAS;IAEjC,YAAY,OAAO,EAAE,uBAAuB,EAE3C;IAED,QAAQ,IAAI,cAAc,CAEzB;IAED,oFAAoF;IACpF,WAAW,IAAI,OAAO,CAErB;IAED,OAAO,CAAC,QAAQ;IAKhB,8DAA8D;IAC9D,wBAAwB,IAAI,OAAO,CAElC;IAED,OAAO,CAAC,4BAA4B;YAUtB,aAAa;IAY3B;;;OAGG;IACG,KAAK,CAAC,MAAM,CAAC,EAAE,WAAW,GAAG,OAAO,CAAC,IAAI,CAAC,CAQ/C;YAEa,GAAG;YAyCH,iBAAiB;IAqF/B,OAAO,CAAC,WAAW;IAMnB,4DAA4D;IACtD,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,CAAC,SAAK,GAAG,OAAO,CAAC,WAAW,EAAE,CAAC,CAE1D;IAED,wEAAwE;IAClE,YAAY,CACjB,KAAK,EAAE,MAAM,EACb,CAAC,SAAK,EACN,IAAI,CAAC,EAAE,MAAM,EACb,SAAS,GAAE,eAAyB,GAClC,OAAO,CAAC,gBAAgB,EAAE,CAAC,CA2B7B;IAED;;;;;OAKG;IACH,kBAAkB,CACjB,GAAG,EAAE,MAAM,EACX,IAAI,EAAE,MAAM,GACV;QAAE,EAAE,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,GAAG,SAAS,CAU9E;YAEa,WAAW;IAYzB,kEAAkE;IAC5D,OAAO,IAAI,OAAO,CAAC,IAAI,CAAC,CAW7B;CACD;AAWD,wBAAgB,wBAAwB,CAAC,GAAG,EAAE,MAAM,EAAE,OAAO,EAAE,gBAAgB,GAAG,IAAI,CAMrF;AAED,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,MAAM,GAAG,gBAAgB,GAAG,SAAS,CAE7E;AAED,wBAAgB,0BAA0B,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,CAE5D","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 { type DaemonRetriever, 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/**\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\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\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\tprivate probeLexicalRetrieverSupport(binary: string): boolean {\n\t\ttry {\n\t\t\tconst out = execFileSync(binary, [\"--version\"], { encoding: \"utf-8\", timeout: 10_000 });\n\t\t\tconst version = parseBinaryVersion(out);\n\t\t\treturn version !== undefined && atLeast(version, MIN_LEXICAL_RETRIEVER_VERSION);\n\t\t} catch {\n\t\t\treturn false;\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\tthis.lexicalRetriever = this.probeLexicalRetrieverSupport(binary);\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/embsearch/embsearch-service.js

@@ -13,6 +13,7 @@
  * Every failure degrades to `unavailable` with a reason; nothing here ever
  * blocks session startup or affects grep/find.
  */
+import { execFileSync } from "child_process";
 import { readFileSync } from "fs";
 import { minimatch } from "minimatch";
 import { ensureTool } from "../../utils/tools-manager.js";
@@ -27,12 +28,36 @@ const BULK_BATCH_SIZE = 48;
 const BATCH_YIELD_MS = 15;
 /** The Rust mock backend's model id — semantically meaningless, never index with it. */
 const MOCK_MODEL_ID = "mock-hash-v1";
+/**
+ * First embsearch release serving `retriever: "lexical"`.
+ *
+ * The guard matters because the daemon does not reject unknown request fields:
+ * an older binary silently ignores `retriever` and answers with dense results.
+ * Fusing that list a second time as a "bm25" leg would double-count it and
+ * corrupt the ranking with no error anywhere — so refuse instead of degrading.
+ */
+const MIN_LEXICAL_RETRIEVER_VERSION = [0, 2, 0];
+/** `embsearch 0.2.0` -> [0, 2, 0]; undefined when it cannot be parsed. */
+function parseBinaryVersion(output) {
+    const match = output.match(/(\d+)\.(\d+)\.(\d+)/);
+    return match ? [Number(match[1]), Number(match[2]), Number(match[3])] : undefined;
+}
+function atLeast(version, minimum) {
+    for (let i = 0; i < minimum.length; i++) {
+        const part = version[i] ?? 0;
+        if (part !== minimum[i])
+            return part > minimum[i];
+    }
+    return true;
+}
 export class EmbsearchService {
     options;
     client;
     meta;
     state = { phase: "idle" };
     disposed = false;
+    /** Whether the resolved binary serves `retriever: "lexical"`. */
+    lexicalRetriever = false;
     constructor(options) {
         this.options = options;
     }
@@ -47,6 +72,20 @@ export class EmbsearchService {
         this.state = state;
         this.options.onProgress?.(state);
     }
+    /** Whether the running daemon can serve a BM25-only query. */
+    supportsLexicalRetriever() {
+        return this.lexicalRetriever;
+    }
+    probeLexicalRetrieverSupport(binary) {
+        try {
+            const out = execFileSync(binary, ["--version"], { encoding: "utf-8", timeout: 10_000 });
+            const version = parseBinaryVersion(out);
+            return version !== undefined && atLeast(version, MIN_LEXICAL_RETRIEVER_VERSION);
+        }
+        catch {
+            return false;
+        }
+    }
     async resolveBinary() {
         if (this.options.binaryPath) {
             return this.options.binaryPath;
@@ -89,8 +128,13 @@ export class EmbsearchService {
             });
             return;
         }
-        const storeDir = getEmbsearchStoreDir(this.options.cwd);
-        this.client = new EmbSearchClient({ binaryPath: binary, storePath: getVectorStoreDir(storeDir) });
+        this.lexicalRetriever = this.probeLexicalRetrieverSupport(binary);
+        const storeDir = this.options.storeDir ?? getEmbsearchStoreDir(this.options.cwd);
+        this.client = new EmbSearchClient({
+            binaryPath: binary,
+            storePath: getVectorStoreDir(storeDir),
+            hybrid: this.options.hybridStore,
+        });
         await this.client.ready();
         const info = await this.client.info();
         if (info.modelId === MOCK_MODEL_ID) {
@@ -194,11 +238,15 @@ export class EmbsearchService {
         return await this.searchChunks(query, k);
     }
     /** Top-`k` semantic hits including their chunk ids, for rank fusion. */
-    async searchChunks(query, k = 10, glob) {
+    async searchChunks(query, k = 10, glob, retriever = "dense") {
         if (!this.client || this.client.isClosed || !this.meta) {
             throw new Error("semantic index is not available");
         }
-        const results = await this.client.query(query, k);
+        if (retriever === "lexical" && !this.lexicalRetriever) {
+            throw new Error(`embsearch is too old for retriever "lexical" (needs >= ${MIN_LEXICAL_RETRIEVER_VERSION.join(".")}); ` +
+                "an older daemon ignores the field and answers with dense results");
+        }
+        const results = await this.client.query(query, k, retriever);
         const hits = [];
         const matchGlob = (rel) => {
             if (!glob)

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,EAAwB,eAAe,EAAE,MAAM,aAAa,CAAC;AACpE,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;AAEzD,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;IAEjC,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;IAEO,4BAA4B,CAAC,MAAc,EAAW;QAC7D,IAAI,CAAC;YACJ,MAAM,GAAG,GAAG,YAAY,CAAC,MAAM,EAAE,CAAC,WAAW,CAAC,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,CAAC;YACxF,MAAM,OAAO,GAAG,kBAAkB,CAAC,GAAG,CAAC,CAAC;YACxC,OAAO,OAAO,KAAK,SAAS,IAAI,OAAO,CAAC,OAAO,EAAE,6BAA6B,CAAC,CAAC;QACjF,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,KAAK,CAAC;QACd,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,IAAI,CAAC,gBAAgB,GAAG,IAAI,CAAC,4BAA4B,CAAC,MAAM,CAAC,CAAC;QAElE,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 { type DaemonRetriever, 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/**\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\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\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\tprivate probeLexicalRetrieverSupport(binary: string): boolean {\n\t\ttry {\n\t\t\tconst out = execFileSync(binary, [\"--version\"], { encoding: \"utf-8\", timeout: 10_000 });\n\t\t\tconst version = parseBinaryVersion(out);\n\t\t\treturn version !== undefined && atLeast(version, MIN_LEXICAL_RETRIEVER_VERSION);\n\t\t} catch {\n\t\t\treturn false;\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\tthis.lexicalRetriever = this.probeLexicalRetrieverSupport(binary);\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/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"]}