@kolisachint/hoocode-agent 0.5.0 → 0.5.2

package/CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## [0.5.2] - 2026-08-08
+
+## [0.5.1] - 2026-08-08
+
 ## [0.5.0] - 2026-08-08
 
 ## [0.4.168] - 2026-08-08

package/dist/core/search/hybrid-search.d.ts

@@ -104,6 +104,27 @@ export interface RunSearchResult {
  * grep. Past the cap the remainder keeps its fused position.
  */
 export declare function hoistStaleCandidates(candidates: readonly FusedCandidate[], staleFiles: ReadonlySet<string>): FusedCandidate[];
+/**
+ * Fold overlapping or adjacent spans of the same file into their best-ranked
+ * occurrence.
+ *
+ * Chunks overlap by design (see the chunker's `CHUNK_OVERLAP_LINES`) and each
+ * one is a separate id, so neighbouring chunks of one region survive fusion as
+ * separate candidates and take separate result slots — showing the model code
+ * it already has. Measured on the 62-query set at the tool's default
+ * `limit=5`, 43 queries had such a pair in their top 5 and 58 of 310 slots
+ * went to repeated code.
+ *
+ * Merging is free recall: the union of two overlapping spans matches exactly
+ * what either matched, so nothing is gained by widening — the gain is entirely
+ * the slot handed back to the ranked tail. Recall@5 0.597 -> 0.677, Recall@1
+ * unchanged (merging cannot alter the top result).
+ *
+ * Adjacency counts as overlap (`endLine + 1`): two chunks that abut describe
+ * one continuous region, and rendering them as separate results implies a gap
+ * that is not there.
+ */
+export declare function mergeOverlappingSpans(candidates: readonly FusedCandidate[]): FusedCandidate[];
 export declare function retrieveCandidates(options: RetrieveOptions): Promise<RetrieveResult>;
 export declare function runSearch(options: RunSearchOptions): Promise<RunSearchResult>;
 //# sourceMappingURL=hybrid-search.d.ts.map

package/dist/core/search/hybrid-search.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"hybrid-search.d.ts","sourceRoot":"","sources":["../../../src/core/search/hybrid-search.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,mCAAmC,CAAC;AAS1E,OAAO,KAAK,EAAiB,cAAc,EAAa,kBAAkB,EAAE,UAAU,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAuCxH,MAAM,WAAW,eAAe;IAC/B,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,CAAC,EAAE,UAAU,CAAC;IAClB,kDAAkD;IAClD,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,yCAAyC;IACzC,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,wFAAwF;IACxF,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,wEAAwE;IACxE,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB;;;;;;;;OAQG;IACH,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB;;;;;;;;;;;OAWG;IACH,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB;;;;OAIG;IACH,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,OAAO,CAAC,EAAE,gBAAgB,CAAC;IAC3B,MAAM,CAAC,EAAE,WAAW,CAAC;CACrB;AAED,MAAM,WAAW,cAAc;IAC9B,UAAU,EAAE,cAAc,EAAE,CAAC;IAC7B,YAAY,EAAE,kBAAkB,CAAC;IACjC,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,WAAW,CAAC,YAAY,CAAC,CAAC;IACtC,UAAU,EAAE,WAAW,CAAC,YAAY,CAAC,CAAC;IACtC,uDAAuD;IACvD,QAAQ,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;IAC3C,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,CAAC,EAAE,WAAW,CAAC,QAAQ,CAAC,CAAC;CAC/B;AAED,MAAM,WAAW,gBAAiB,SAAQ,eAAe;IACxD,oDAAoD;IACpD,WAAW,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,eAAe;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,YAAY,EAAE,kBAAkB,CAAC;IACjC,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,WAAW,EAAE,MAAM,CAAC;IACpB,uDAAuD;IACvD,QAAQ,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;CAC3C;AAYD;;;;;;;;;;;;;;;GAeG;AACH,wBAAgB,oBAAoB,CACnC,UAAU,EAAE,SAAS,cAAc,EAAE,EACrC,UAAU,EAAE,WAAW,CAAC,MAAM,CAAC,GAC7B,cAAc,EAAE,CAQlB;AAED,wBAAsB,kBAAkB,CAAC,OAAO,EAAE,eAAe,GAAG,OAAO,CAAC,cAAc,CAAC,CAuK1F;AAED,wBAAsB,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,eAAe,CAAC,CAyBnF","sourcesContent":["/**\n * Hybrid search orchestrator: resolve mode, run retrievers in parallel, fuse\n * by rank, expand within budget, trace everything\n * (docs/hybrid-retrieval-design.md).\n *\n * Single-retriever modes flow through the same pipeline — rrfFuse over one\n * list preserves its order — so lexical, semantic, and hybrid all produce the\n * same result shape and the same trace record.\n *\n * `retrieveCandidates` is the candidate-level core (also used by the eval\n * harness, which needs forced modes, a configurable `k`, and no trace\n * pollution); `runSearch` wraps it with span expansion and tracing for the\n * tool.\n */\n\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport { adaptGrepHits, type ChunkLookup } from \"./adapter.js\";\nimport { assembleContext } from \"./context-assembler.js\";\nimport { crossEncoderRerank } from \"./cross-rerank.js\";\nimport { runLexicalRetriever } from \"./lexical-retriever.js\";\nimport { resolveSearchMode } from \"./mode.js\";\nimport { rerankCandidates } from \"./rerank.js\";\nimport { DEFAULT_RRF_K, rrfFuse } from \"./rrf.js\";\nimport { writeSearchTrace } from \"./trace.js\";\nimport type { CandidateSpan, FusedCandidate, RankedHit, ResolvedSearchMode, SearchMode, SearchTrace } from \"./types.js\";\n\n/** Raw grep line-hits fetched per query (pre-collapse). */\nconst LEXICAL_MATCH_LIMIT = 200;\n/** Adapted lexical candidates entering fusion. The eval gate showed the\n *  uncapped lexical tail diluting hybrid below plain semantic: lexical\n *  precision is front-loaded by the adapter's term-evidence ranking, while\n *  RRF weighs a rank-30 lexical candidate like a rank-30 embedding hit. */\nconst LEXICAL_FUSION_CAP = 20;\n/**\n * Embedding hits fetched per query when fusing with the grep leg.\n *\n * Deliberately shallow. The grep list is capped at {@link LEXICAL_FUSION_CAP}\n * because its tail is unranked noise, so a deep dense pool here simply\n * outnumbers it: raising this to 200 alongside a 20-candidate grep list cost\n * `auto +rr` 0.472 -> 0.450 MRR.\n */\nconst EMBED_TOP_K = 50;\n/**\n * Per-leg depth when both retrievers are ranked ones (dense + BM25).\n *\n * Four times {@link FUSED_WINDOW}, matching what the daemon's `query_hybrid`\n * does internally (`pool = 4·k`). Fusing at the same depth as the window loses\n * any candidate ranked well by one retriever but just outside the other's\n * top-50 — worth 12pp R@10 and 8pp R@50 for `bm25+dense`, which is exactly\n * what closed the gap to daemon-side fusion.\n */\nconst FUSION_POOL_TOP_K = 200;\n/** BM25 depth, matching {@link FUSION_POOL_TOP_K} so neither ranked leg is\n *  structurally advantaged by pool size. The daemon returns only documents\n *  sharing a query term, so this is an upper bound, not a fill. */\nconst BM25_TOP_K = FUSION_POOL_TOP_K;\n/** Candidates from unindexed/stale files moved ahead of the fused window.\n *  Small on purpose: this answers \"what did I just write\", not \"search the\n *  working tree\". */\nconst STALE_HOIST_CAP = 5;\n/** Fused candidates kept for reranking / final slicing. */\nconst FUSED_WINDOW = 50;\n\nexport interface RetrieveOptions {\n\tcwd: string;\n\tquery: string;\n\tmode?: SearchMode;\n\t/** Optional glob filter applied to file paths. */\n\tglob?: string;\n\t/** Maximum fused candidates returned. */\n\tlimit?: number;\n\t/** RRF constant override (eval harness sweeps this). Default: {@link DEFAULT_RRF_K}. */\n\trrfK?: number;\n\t/** Rerank the fused top-50 before slicing to `limit`. Default: true. */\n\trerank?: boolean;\n\t/**\n\t * Ask the daemon to fuse its own BM25 index with the vectors and return one\n\t * already-fused ranking, instead of taking a dense-only list. Needs a store\n\t * built with `--hybrid`. The fused list arrives as a single \"embed\" leg,\n\t * because a pre-fused ranking has no per-retriever structure left to record.\n\t *\n\t * Prefer {@link bm25Leg}: fusing here keeps the legs separable in the trace\n\t * and lets the grep leg participate.\n\t */\n\tdaemonHybrid?: boolean;\n\t/**\n\t * Fetch the daemon's BM25 index as its own ranked list and fuse it here,\n\t * alongside dense and grep.\n\t *\n\t * Defaults to on wherever the daemon can serve it, because BM25 is the\n\t * better lexical leg on the indexed corpus: Recall@50 0.790 -> 0.879, 6 of\n\t * 62 queries better and 0 worse (p <= 0.05). When it is on, the grep leg\n\t * narrows to files the index has not read — see `staleFiles`.\n\t *\n\t * Set `false` to force ripgrep as the only lexical leg; the eval harness\n\t * does this to keep measuring what the old rows measured.\n\t */\n\tbm25Leg?: boolean;\n\t/**\n\t * Reorder the fused shortlist with the daemon's cross-encoder instead of\n\t * the deterministic reranker. Needs embsearch >= 0.3.0; costs one model\n\t * pass per scored candidate.\n\t */\n\tcrossEncoder?: boolean;\n\tservice?: EmbsearchService;\n\tsignal?: AbortSignal;\n}\n\nexport interface RetrieveResult {\n\tcandidates: FusedCandidate[];\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tindexPhase: SearchTrace[\"indexPhase\"];\n\tretrievers: SearchTrace[\"retrievers\"];\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n\trrfK: number;\n\trerank?: SearchTrace[\"rerank\"];\n}\n\nexport interface RunSearchOptions extends RetrieveOptions {\n\t/** Approximate token budget for the result text. */\n\ttokenBudget?: number;\n}\n\nexport interface RunSearchResult {\n\ttext: string;\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tresultCount: number;\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n}\n\nfunction normalizeSearchGlob(glob: string | undefined): string | undefined {\n\tif (!glob) return undefined;\n\t// Match fd/rg semantics: a slash-containing glob is anchored anywhere in\n\t// the tree, so prepend \"**/\" unless it already starts with a slash or \"**/\".\n\tif (glob.includes(\"/\") && !glob.startsWith(\"/\") && !glob.startsWith(\"**/\")) {\n\t\treturn `**/${glob}`;\n\t}\n\treturn glob;\n}\n\n/**\n * Move candidates from files the index has not read to the front.\n *\n * They are there because grep found them and nothing else could: the index is\n * ranking a stale copy of the file, or has never seen it. Left to fuse, they\n * lose — RRF rewards agreement, and one leg reporting a single document is\n * outvoted by two legs agreeing on hundreds. Measured: scoping grep to stale\n * files without this hoist scored 25% on the live-edit set where unscoped grep\n * scored 100%, because the fused window filled with consensus hits about the\n * indexed copy.\n *\n * Capped, because \"stale\" scales with how far behind the index is. A few\n * edited files is the case this exists for; a fresh checkout makes everything\n * stale, and hoisting all of it would quietly turn hybrid search back into\n * grep. Past the cap the remainder keeps its fused position.\n */\nexport function hoistStaleCandidates(\n\tcandidates: readonly FusedCandidate[],\n\tstaleFiles: ReadonlySet<string>,\n): FusedCandidate[] {\n\tif (staleFiles.size === 0) return [...candidates];\n\tconst hoisted: FusedCandidate[] = [];\n\tconst rest: FusedCandidate[] = [];\n\tfor (const candidate of candidates) {\n\t\t(staleFiles.has(candidate.path) && hoisted.length < STALE_HOIST_CAP ? hoisted : rest).push(candidate);\n\t}\n\treturn [...hoisted, ...rest];\n}\n\nexport async function retrieveCandidates(options: RetrieveOptions): Promise<RetrieveResult> {\n\tconst { cwd, query, service, signal } = options;\n\tconst glob = normalizeSearchGlob(options.glob);\n\tconst requestedMode = options.mode ?? \"auto\";\n\tconst limit = Math.max(1, options.limit ?? 10);\n\tconst rrfK = options.rrfK ?? DEFAULT_RRF_K;\n\n\tconst state = service?.getState();\n\tconst embedAvailable = service?.isAvailable() ?? false;\n\tconst embedUnavailableReason =\n\t\tstate === undefined\n\t\t\t? \"semantic index is not enabled\"\n\t\t\t: state.phase === \"unavailable\" || state.phase === \"skipped\"\n\t\t\t\t? state.reason\n\t\t\t\t: state.phase === \"idle\"\n\t\t\t\t\t? \"semantic index has not started\"\n\t\t\t\t\t: undefined;\n\n\tconst resolution = resolveSearchMode(query, requestedMode, embedAvailable, embedUnavailableReason);\n\tconst mode = resolution.mode;\n\n\t// Map lexical hits onto indexed chunk ids whenever the sidecar is usable,\n\t// even in lexical-only mode, so identities line up across modes.\n\tconst lookupChunk: ChunkLookup | undefined = embedAvailable\n\t\t? (rel, line) => service!.findEnclosingChunk(rel, line)\n\t\t: undefined;\n\n\tconst spans = new Map<string, CandidateSpan>();\n\t/** Files the index does not have current content for, when grep is scoped\n\t *  to them. Read again after fusion — see the hoist below. */\n\tconst staleSet = new Set<string>();\n\tconst lists: RankedHit[][] = [];\n\tconst retrieverStats: SearchTrace[\"retrievers\"] = {};\n\tconst errors: Error[] = [];\n\n\tconst runLexical = async (scopeToStale = false): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Scoped run: grep covers only what the index has not read yet, so it\n\t\t\t// adds the one thing BM25 cannot see without re-voting on documents\n\t\t\t// BM25 already ranked. Fusing two lexical views of the same corpus is\n\t\t\t// what made the three-leg configuration lose (18 of 62 queries worse).\n\t\t\tconst paths = scopeToStale ? service?.staleFiles(signal) : undefined;\n\t\t\tif (paths) for (const rel of paths) staleSet.add(rel);\n\t\t\tconst lineHits = await runLexicalRetriever({ cwd, query, limit: LEXICAL_MATCH_LIMIT, glob, signal, paths });\n\t\t\tconst adapted = adaptGrepHits(lineHits, lookupChunk);\n\t\t\t// In single-retriever lexical mode the full list is the result; in\n\t\t\t// hybrid, only the front-loaded head is trustworthy enough to vote.\n\t\t\tconst hits = mode === \"hybrid\" ? adapted.hits.slice(0, LEXICAL_FUSION_CAP) : adapted.hits;\n\t\t\tfor (const [id, span] of adapted.spans) if (!spans.has(id)) spans.set(id, span);\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runEmbed = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// The flat index pads top-k with whatever exists; with the cosine\n\t\t\t// metric the store uses, score <= 0 means \"no relation at all\", so\n\t\t\t// those padding hits would cast RRF votes on pure noise.\n\t\t\t// A BM25 leg is itself ranked, so the pair can afford — and needs — the\n\t\t\t// deeper pool; the grep leg cannot (see EMBED_TOP_K).\n\t\t\tconst topK = options.bm25Leg ? FUSION_POOL_TOP_K : EMBED_TOP_K;\n\t\t\tconst chunkHits = (\n\t\t\t\tawait service!.searchChunks(query, topK, glob, options.daemonHybrid ? \"hybrid\" : \"dense\")\n\t\t\t).filter((hit) => hit.score > 0);\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"embed\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runBm25 = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Raw BM25 sums; only the ordering enters fusion, the score is a\n\t\t\t// diagnostic. Unlike the dense leg there is no zero-score padding to\n\t\t\t// filter — the daemon omits documents sharing no query term.\n\t\t\tconst chunkHits = await service!.searchChunks(query, BM25_TOP_K, glob, \"lexical\");\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"bm25\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tif (!spans.has(hit.id)) {\n\t\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t\t}\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\t// BM25 is the better lexical leg where the index is current: it beats grep\n\t// on Recall@50 (0.790 -> 0.879, 6 queries better and 0 worse, p <= 0.05)\n\t// because it ranks the whole corpus rather than truncating a match stream.\n\t// It is also blind to anything indexed later than it was written, which is\n\t// precisely where grep still wins — so grep runs scoped to that set instead\n\t// of being dropped or left to duplicate BM25 over the whole tree.\n\tconst bm25AsLexicalLeg =\n\t\t(options.bm25Leg ?? service?.supportsLexicalRetriever() ?? false) && embedAvailable && mode !== \"lexical\";\n\n\tconst runs: Promise<void>[] = [];\n\tif (mode === \"lexical\" || mode === \"hybrid\") runs.push(runLexical(bm25AsLexicalLeg));\n\tif (mode === \"semantic\" || mode === \"hybrid\") runs.push(runEmbed());\n\t// `semantic` gets the BM25 leg too: the caller asked for the index, and\n\t// BM25 is part of it. Only an explicit `lexical` request excludes it.\n\tif (bm25AsLexicalLeg) runs.push(runBm25());\n\tawait Promise.all(runs);\n\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\t// A partial failure in hybrid degrades to whichever retriever survived;\n\t// only a total loss is an error.\n\tif (lists.length === 0) throw errors[0] ?? new Error(\"search produced no retriever results\");\n\n\tconst fused = rrfFuse(lists, rrfK).slice(0, FUSED_WINDOW);\n\tlet candidates: FusedCandidate[] = [];\n\tfor (const hit of fused) {\n\t\tconst span = spans.get(hit.id);\n\t\tif (span) candidates.push({ ...hit, ...span });\n\t}\n\n\tlet rerankInfo: SearchTrace[\"rerank\"];\n\tif (options.crossEncoder && service?.supportsCrossEncoder()) {\n\t\tconst reranked = await crossEncoderRerank(query, candidates, cwd, service);\n\t\trerankInfo = { applied: true, candidateCount: reranked.scored, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t} else if (options.rerank !== false) {\n\t\tconst reranked = rerankCandidates(query, candidates, cwd);\n\t\trerankInfo = { applied: true, candidateCount: candidates.length, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t}\n\t// After reranking, not before: both rerankers weight the fused prior, and a\n\t// lone grep hit from an unindexed file has the lowest prior there is, so\n\t// hoisting first would simply be undone. The reranker still orders the\n\t// hoisted set against itself.\n\tcandidates = hoistStaleCandidates(candidates, staleSet).slice(0, limit);\n\n\treturn {\n\t\tcandidates,\n\t\tresolvedMode: mode,\n\t\tdegradedReason: resolution.degradedReason,\n\t\tindexPhase: state?.phase === \"ready\" ? \"ready\" : state?.phase === \"indexing\" ? \"indexing\" : \"unavailable\",\n\t\tretrievers: retrieverStats,\n\t\tindexing: state?.phase === \"indexing\" ? { done: state.done, total: state.total } : undefined,\n\t\trrfK,\n\t\trerank: rerankInfo,\n\t};\n}\n\nexport async function runSearch(options: RunSearchOptions): Promise<RunSearchResult> {\n\tconst retrieved = await retrieveCandidates(options);\n\n\tconst assembled = assembleContext(retrieved.candidates, { cwd: options.cwd, tokenBudget: options.tokenBudget });\n\n\twriteSearchTrace(options.cwd, {\n\t\ttimestampMs: Date.now(),\n\t\tquery: options.query,\n\t\trequestedMode: options.mode ?? \"auto\",\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tindexPhase: retrieved.indexPhase,\n\t\trrfK: retrieved.resolvedMode === \"hybrid\" ? retrieved.rrfK : undefined,\n\t\tretrievers: retrieved.retrievers,\n\t\tfused: retrieved.candidates.map(({ id, rrfScore, ranks, rawScores }) => ({ id, rrfScore, ranks, rawScores })),\n\t\trerank: retrieved.rerank,\n\t});\n\n\treturn {\n\t\ttext: assembled.text,\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tresultCount: retrieved.candidates.length,\n\t\tindexing: retrieved.indexing,\n\t};\n}\n"]}
+{"version":3,"file":"hybrid-search.d.ts","sourceRoot":"","sources":["../../../src/core/search/hybrid-search.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,mCAAmC,CAAC;AAS1E,OAAO,KAAK,EAAiB,cAAc,EAAa,kBAAkB,EAAE,UAAU,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAuCxH,MAAM,WAAW,eAAe;IAC/B,GAAG,EAAE,MAAM,CAAC;IACZ,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,CAAC,EAAE,UAAU,CAAC;IAClB,kDAAkD;IAClD,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,yCAAyC;IACzC,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,wFAAwF;IACxF,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,wEAAwE;IACxE,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB;;;;;;;;OAQG;IACH,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB;;;;;;;;;;;OAWG;IACH,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB;;;;OAIG;IACH,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,OAAO,CAAC,EAAE,gBAAgB,CAAC;IAC3B,MAAM,CAAC,EAAE,WAAW,CAAC;CACrB;AAED,MAAM,WAAW,cAAc;IAC9B,UAAU,EAAE,cAAc,EAAE,CAAC;IAC7B,YAAY,EAAE,kBAAkB,CAAC;IACjC,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,WAAW,CAAC,YAAY,CAAC,CAAC;IACtC,UAAU,EAAE,WAAW,CAAC,YAAY,CAAC,CAAC;IACtC,uDAAuD;IACvD,QAAQ,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;IAC3C,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,CAAC,EAAE,WAAW,CAAC,QAAQ,CAAC,CAAC;CAC/B;AAED,MAAM,WAAW,gBAAiB,SAAQ,eAAe;IACxD,oDAAoD;IACpD,WAAW,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,eAAe;IAC/B,IAAI,EAAE,MAAM,CAAC;IACb,YAAY,EAAE,kBAAkB,CAAC;IACjC,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,WAAW,EAAE,MAAM,CAAC;IACpB,uDAAuD;IACvD,QAAQ,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;CAC3C;AAYD;;;;;;;;;;;;;;;GAeG;AACH,wBAAgB,oBAAoB,CACnC,UAAU,EAAE,SAAS,cAAc,EAAE,EACrC,UAAU,EAAE,WAAW,CAAC,MAAM,CAAC,GAC7B,cAAc,EAAE,CAQlB;AAED;;;;;;;;;;;;;;;;;;;GAmBG;AACH,wBAAgB,qBAAqB,CAAC,UAAU,EAAE,SAAS,cAAc,EAAE,GAAG,cAAc,EAAE,CA+B7F;AAED,wBAAsB,kBAAkB,CAAC,OAAO,EAAE,eAAe,GAAG,OAAO,CAAC,cAAc,CAAC,CAyK1F;AAED,wBAAsB,SAAS,CAAC,OAAO,EAAE,gBAAgB,GAAG,OAAO,CAAC,eAAe,CAAC,CA+BnF","sourcesContent":["/**\n * Hybrid search orchestrator: resolve mode, run retrievers in parallel, fuse\n * by rank, expand within budget, trace everything\n * (docs/hybrid-retrieval-design.md).\n *\n * Single-retriever modes flow through the same pipeline — rrfFuse over one\n * list preserves its order — so lexical, semantic, and hybrid all produce the\n * same result shape and the same trace record.\n *\n * `retrieveCandidates` is the candidate-level core (also used by the eval\n * harness, which needs forced modes, a configurable `k`, and no trace\n * pollution); `runSearch` wraps it with span expansion and tracing for the\n * tool.\n */\n\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport { adaptGrepHits, type ChunkLookup } from \"./adapter.js\";\nimport { assembleContext } from \"./context-assembler.js\";\nimport { crossEncoderRerank } from \"./cross-rerank.js\";\nimport { runLexicalRetriever } from \"./lexical-retriever.js\";\nimport { resolveSearchMode } from \"./mode.js\";\nimport { rerankCandidates } from \"./rerank.js\";\nimport { DEFAULT_RRF_K, rrfFuse } from \"./rrf.js\";\nimport { writeSearchTrace } from \"./trace.js\";\nimport type { CandidateSpan, FusedCandidate, RankedHit, ResolvedSearchMode, SearchMode, SearchTrace } from \"./types.js\";\n\n/** Raw grep line-hits fetched per query (pre-collapse). */\nconst LEXICAL_MATCH_LIMIT = 200;\n/** Adapted lexical candidates entering fusion. The eval gate showed the\n *  uncapped lexical tail diluting hybrid below plain semantic: lexical\n *  precision is front-loaded by the adapter's term-evidence ranking, while\n *  RRF weighs a rank-30 lexical candidate like a rank-30 embedding hit. */\nconst LEXICAL_FUSION_CAP = 20;\n/**\n * Embedding hits fetched per query when fusing with the grep leg.\n *\n * Deliberately shallow. The grep list is capped at {@link LEXICAL_FUSION_CAP}\n * because its tail is unranked noise, so a deep dense pool here simply\n * outnumbers it: raising this to 200 alongside a 20-candidate grep list cost\n * `auto +rr` 0.472 -> 0.450 MRR.\n */\nconst EMBED_TOP_K = 50;\n/**\n * Per-leg depth when both retrievers are ranked ones (dense + BM25).\n *\n * Four times {@link FUSED_WINDOW}, matching what the daemon's `query_hybrid`\n * does internally (`pool = 4·k`). Fusing at the same depth as the window loses\n * any candidate ranked well by one retriever but just outside the other's\n * top-50 — worth 12pp R@10 and 8pp R@50 for `bm25+dense`, which is exactly\n * what closed the gap to daemon-side fusion.\n */\nconst FUSION_POOL_TOP_K = 200;\n/** BM25 depth, matching {@link FUSION_POOL_TOP_K} so neither ranked leg is\n *  structurally advantaged by pool size. The daemon returns only documents\n *  sharing a query term, so this is an upper bound, not a fill. */\nconst BM25_TOP_K = FUSION_POOL_TOP_K;\n/** Candidates from unindexed/stale files moved ahead of the fused window.\n *  Small on purpose: this answers \"what did I just write\", not \"search the\n *  working tree\". */\nconst STALE_HOIST_CAP = 5;\n/** Fused candidates kept for reranking / final slicing. */\nconst FUSED_WINDOW = 50;\n\nexport interface RetrieveOptions {\n\tcwd: string;\n\tquery: string;\n\tmode?: SearchMode;\n\t/** Optional glob filter applied to file paths. */\n\tglob?: string;\n\t/** Maximum fused candidates returned. */\n\tlimit?: number;\n\t/** RRF constant override (eval harness sweeps this). Default: {@link DEFAULT_RRF_K}. */\n\trrfK?: number;\n\t/** Rerank the fused top-50 before slicing to `limit`. Default: true. */\n\trerank?: boolean;\n\t/**\n\t * Ask the daemon to fuse its own BM25 index with the vectors and return one\n\t * already-fused ranking, instead of taking a dense-only list. Needs a store\n\t * built with `--hybrid`. The fused list arrives as a single \"embed\" leg,\n\t * because a pre-fused ranking has no per-retriever structure left to record.\n\t *\n\t * Prefer {@link bm25Leg}: fusing here keeps the legs separable in the trace\n\t * and lets the grep leg participate.\n\t */\n\tdaemonHybrid?: boolean;\n\t/**\n\t * Fetch the daemon's BM25 index as its own ranked list and fuse it here,\n\t * alongside dense and grep.\n\t *\n\t * Defaults to on wherever the daemon can serve it, because BM25 is the\n\t * better lexical leg on the indexed corpus: Recall@50 0.790 -> 0.879, 6 of\n\t * 62 queries better and 0 worse (p <= 0.05). When it is on, the grep leg\n\t * narrows to files the index has not read — see `staleFiles`.\n\t *\n\t * Set `false` to force ripgrep as the only lexical leg; the eval harness\n\t * does this to keep measuring what the old rows measured.\n\t */\n\tbm25Leg?: boolean;\n\t/**\n\t * Reorder the fused shortlist with the daemon's cross-encoder instead of\n\t * the deterministic reranker. Needs embsearch >= 0.3.0; costs one model\n\t * pass per scored candidate.\n\t */\n\tcrossEncoder?: boolean;\n\tservice?: EmbsearchService;\n\tsignal?: AbortSignal;\n}\n\nexport interface RetrieveResult {\n\tcandidates: FusedCandidate[];\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tindexPhase: SearchTrace[\"indexPhase\"];\n\tretrievers: SearchTrace[\"retrievers\"];\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n\trrfK: number;\n\trerank?: SearchTrace[\"rerank\"];\n}\n\nexport interface RunSearchOptions extends RetrieveOptions {\n\t/** Approximate token budget for the result text. */\n\ttokenBudget?: number;\n}\n\nexport interface RunSearchResult {\n\ttext: string;\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tresultCount: number;\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n}\n\nfunction normalizeSearchGlob(glob: string | undefined): string | undefined {\n\tif (!glob) return undefined;\n\t// Match fd/rg semantics: a slash-containing glob is anchored anywhere in\n\t// the tree, so prepend \"**/\" unless it already starts with a slash or \"**/\".\n\tif (glob.includes(\"/\") && !glob.startsWith(\"/\") && !glob.startsWith(\"**/\")) {\n\t\treturn `**/${glob}`;\n\t}\n\treturn glob;\n}\n\n/**\n * Move candidates from files the index has not read to the front.\n *\n * They are there because grep found them and nothing else could: the index is\n * ranking a stale copy of the file, or has never seen it. Left to fuse, they\n * lose — RRF rewards agreement, and one leg reporting a single document is\n * outvoted by two legs agreeing on hundreds. Measured: scoping grep to stale\n * files without this hoist scored 25% on the live-edit set where unscoped grep\n * scored 100%, because the fused window filled with consensus hits about the\n * indexed copy.\n *\n * Capped, because \"stale\" scales with how far behind the index is. A few\n * edited files is the case this exists for; a fresh checkout makes everything\n * stale, and hoisting all of it would quietly turn hybrid search back into\n * grep. Past the cap the remainder keeps its fused position.\n */\nexport function hoistStaleCandidates(\n\tcandidates: readonly FusedCandidate[],\n\tstaleFiles: ReadonlySet<string>,\n): FusedCandidate[] {\n\tif (staleFiles.size === 0) return [...candidates];\n\tconst hoisted: FusedCandidate[] = [];\n\tconst rest: FusedCandidate[] = [];\n\tfor (const candidate of candidates) {\n\t\t(staleFiles.has(candidate.path) && hoisted.length < STALE_HOIST_CAP ? hoisted : rest).push(candidate);\n\t}\n\treturn [...hoisted, ...rest];\n}\n\n/**\n * Fold overlapping or adjacent spans of the same file into their best-ranked\n * occurrence.\n *\n * Chunks overlap by design (see the chunker's `CHUNK_OVERLAP_LINES`) and each\n * one is a separate id, so neighbouring chunks of one region survive fusion as\n * separate candidates and take separate result slots — showing the model code\n * it already has. Measured on the 62-query set at the tool's default\n * `limit=5`, 43 queries had such a pair in their top 5 and 58 of 310 slots\n * went to repeated code.\n *\n * Merging is free recall: the union of two overlapping spans matches exactly\n * what either matched, so nothing is gained by widening — the gain is entirely\n * the slot handed back to the ranked tail. Recall@5 0.597 -> 0.677, Recall@1\n * unchanged (merging cannot alter the top result).\n *\n * Adjacency counts as overlap (`endLine + 1`): two chunks that abut describe\n * one continuous region, and rendering them as separate results implies a gap\n * that is not there.\n */\nexport function mergeOverlappingSpans(candidates: readonly FusedCandidate[]): FusedCandidate[] {\n\tconst merged: FusedCandidate[] = [];\n\tfor (const candidate of candidates) {\n\t\tconst into = merged.find(\n\t\t\t(m) =>\n\t\t\t\tm.path === candidate.path && candidate.startLine <= m.endLine + 1 && m.startLine <= candidate.endLine + 1,\n\t\t);\n\t\tif (!into) {\n\t\t\tmerged.push({ ...candidate, ranks: { ...candidate.ranks }, rawScores: { ...candidate.rawScores } });\n\t\t\tcontinue;\n\t\t}\n\t\tinto.startLine = Math.min(into.startLine, candidate.startLine);\n\t\tinto.endLine = Math.max(into.endLine, candidate.endLine);\n\t\t// Keep the folded id: the trace is the only record that retrieval found\n\t\t// it, and dropping it would make a merged result look like a miss.\n\t\tif (!into.mergedFrom) into.mergedFrom = [];\n\t\tinto.mergedFrom.push(candidate.id);\n\t\t// The merged region really was reached by every retriever that found any\n\t\t// part of it, so the source label and the trace should say so. Best rank\n\t\t// per source wins; the fused score stays the survivor's, which is what\n\t\t// ordered it here.\n\t\tfor (const [source, rank] of Object.entries(candidate.ranks)) {\n\t\t\tconst held = into.ranks[source as keyof typeof into.ranks];\n\t\t\tif (held === undefined || rank < held) into.ranks[source as keyof typeof into.ranks] = rank;\n\t\t}\n\t\tfor (const [source, score] of Object.entries(candidate.rawScores)) {\n\t\t\tconst held = into.rawScores[source as keyof typeof into.rawScores];\n\t\t\tif (held === undefined || score > held) into.rawScores[source as keyof typeof into.rawScores] = score;\n\t\t}\n\t}\n\treturn merged;\n}\n\nexport async function retrieveCandidates(options: RetrieveOptions): Promise<RetrieveResult> {\n\tconst { cwd, query, service, signal } = options;\n\tconst glob = normalizeSearchGlob(options.glob);\n\tconst requestedMode = options.mode ?? \"auto\";\n\tconst limit = Math.max(1, options.limit ?? 10);\n\tconst rrfK = options.rrfK ?? DEFAULT_RRF_K;\n\n\tconst state = service?.getState();\n\tconst embedAvailable = service?.isAvailable() ?? false;\n\tconst embedUnavailableReason =\n\t\tstate === undefined\n\t\t\t? \"semantic index is not enabled\"\n\t\t\t: state.phase === \"unavailable\" || state.phase === \"skipped\"\n\t\t\t\t? state.reason\n\t\t\t\t: state.phase === \"idle\"\n\t\t\t\t\t? \"semantic index has not started\"\n\t\t\t\t\t: undefined;\n\n\tconst resolution = resolveSearchMode(query, requestedMode, embedAvailable, embedUnavailableReason);\n\tconst mode = resolution.mode;\n\n\t// Map lexical hits onto indexed chunk ids whenever the sidecar is usable,\n\t// even in lexical-only mode, so identities line up across modes.\n\tconst lookupChunk: ChunkLookup | undefined = embedAvailable\n\t\t? (rel, line) => service!.findEnclosingChunk(rel, line)\n\t\t: undefined;\n\n\tconst spans = new Map<string, CandidateSpan>();\n\t/** Files the index does not have current content for, when grep is scoped\n\t *  to them. Read again after fusion — see the hoist below. */\n\tconst staleSet = new Set<string>();\n\tconst lists: RankedHit[][] = [];\n\tconst retrieverStats: SearchTrace[\"retrievers\"] = {};\n\tconst errors: Error[] = [];\n\n\tconst runLexical = async (scopeToStale = false): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Scoped run: grep covers only what the index has not read yet, so it\n\t\t\t// adds the one thing BM25 cannot see without re-voting on documents\n\t\t\t// BM25 already ranked. Fusing two lexical views of the same corpus is\n\t\t\t// what made the three-leg configuration lose (18 of 62 queries worse).\n\t\t\tconst paths = scopeToStale ? service?.staleFiles(signal) : undefined;\n\t\t\tif (paths) for (const rel of paths) staleSet.add(rel);\n\t\t\tconst lineHits = await runLexicalRetriever({ cwd, query, limit: LEXICAL_MATCH_LIMIT, glob, signal, paths });\n\t\t\tconst adapted = adaptGrepHits(lineHits, lookupChunk);\n\t\t\t// In single-retriever lexical mode the full list is the result; in\n\t\t\t// hybrid, only the front-loaded head is trustworthy enough to vote.\n\t\t\tconst hits = mode === \"hybrid\" ? adapted.hits.slice(0, LEXICAL_FUSION_CAP) : adapted.hits;\n\t\t\tfor (const [id, span] of adapted.spans) if (!spans.has(id)) spans.set(id, span);\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runEmbed = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// The flat index pads top-k with whatever exists; with the cosine\n\t\t\t// metric the store uses, score <= 0 means \"no relation at all\", so\n\t\t\t// those padding hits would cast RRF votes on pure noise.\n\t\t\t// A BM25 leg is itself ranked, so the pair can afford — and needs — the\n\t\t\t// deeper pool; the grep leg cannot (see EMBED_TOP_K).\n\t\t\tconst topK = options.bm25Leg ? FUSION_POOL_TOP_K : EMBED_TOP_K;\n\t\t\tconst chunkHits = (\n\t\t\t\tawait service!.searchChunks(query, topK, glob, options.daemonHybrid ? \"hybrid\" : \"dense\")\n\t\t\t).filter((hit) => hit.score > 0);\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"embed\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runBm25 = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Raw BM25 sums; only the ordering enters fusion, the score is a\n\t\t\t// diagnostic. Unlike the dense leg there is no zero-score padding to\n\t\t\t// filter — the daemon omits documents sharing no query term.\n\t\t\tconst chunkHits = await service!.searchChunks(query, BM25_TOP_K, glob, \"lexical\");\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"bm25\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tif (!spans.has(hit.id)) {\n\t\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t\t}\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\t// BM25 is the better lexical leg where the index is current: it beats grep\n\t// on Recall@50 (0.790 -> 0.879, 6 queries better and 0 worse, p <= 0.05)\n\t// because it ranks the whole corpus rather than truncating a match stream.\n\t// It is also blind to anything indexed later than it was written, which is\n\t// precisely where grep still wins — so grep runs scoped to that set instead\n\t// of being dropped or left to duplicate BM25 over the whole tree.\n\tconst bm25AsLexicalLeg =\n\t\t(options.bm25Leg ?? service?.supportsLexicalRetriever() ?? false) && embedAvailable && mode !== \"lexical\";\n\n\tconst runs: Promise<void>[] = [];\n\tif (mode === \"lexical\" || mode === \"hybrid\") runs.push(runLexical(bm25AsLexicalLeg));\n\tif (mode === \"semantic\" || mode === \"hybrid\") runs.push(runEmbed());\n\t// `semantic` gets the BM25 leg too: the caller asked for the index, and\n\t// BM25 is part of it. Only an explicit `lexical` request excludes it.\n\tif (bm25AsLexicalLeg) runs.push(runBm25());\n\tawait Promise.all(runs);\n\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\t// A partial failure in hybrid degrades to whichever retriever survived;\n\t// only a total loss is an error.\n\tif (lists.length === 0) throw errors[0] ?? new Error(\"search produced no retriever results\");\n\n\tconst fused = rrfFuse(lists, rrfK).slice(0, FUSED_WINDOW);\n\tlet candidates: FusedCandidate[] = [];\n\tfor (const hit of fused) {\n\t\tconst span = spans.get(hit.id);\n\t\tif (span) candidates.push({ ...hit, ...span });\n\t}\n\n\tlet rerankInfo: SearchTrace[\"rerank\"];\n\tif (options.crossEncoder && service?.supportsCrossEncoder()) {\n\t\tconst reranked = await crossEncoderRerank(query, candidates, cwd, service);\n\t\trerankInfo = { applied: true, candidateCount: reranked.scored, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t} else if (options.rerank !== false) {\n\t\tconst reranked = rerankCandidates(query, candidates, cwd);\n\t\trerankInfo = { applied: true, candidateCount: candidates.length, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t}\n\t// After reranking, not before: both rerankers weight the fused prior, and a\n\t// lone grep hit from an unindexed file has the lowest prior there is, so\n\t// hoisting first would simply be undone. The reranker still orders the\n\t// hoisted set against itself.\n\t// Merge before slicing, so a slot freed by a duplicate is refilled from the\n\t// ranked tail rather than left empty.\n\tcandidates = mergeOverlappingSpans(hoistStaleCandidates(candidates, staleSet)).slice(0, limit);\n\n\treturn {\n\t\tcandidates,\n\t\tresolvedMode: mode,\n\t\tdegradedReason: resolution.degradedReason,\n\t\tindexPhase: state?.phase === \"ready\" ? \"ready\" : state?.phase === \"indexing\" ? \"indexing\" : \"unavailable\",\n\t\tretrievers: retrieverStats,\n\t\tindexing: state?.phase === \"indexing\" ? { done: state.done, total: state.total } : undefined,\n\t\trrfK,\n\t\trerank: rerankInfo,\n\t};\n}\n\nexport async function runSearch(options: RunSearchOptions): Promise<RunSearchResult> {\n\tconst retrieved = await retrieveCandidates(options);\n\n\tconst assembled = assembleContext(retrieved.candidates, { cwd: options.cwd, tokenBudget: options.tokenBudget });\n\n\twriteSearchTrace(options.cwd, {\n\t\ttimestampMs: Date.now(),\n\t\tquery: options.query,\n\t\trequestedMode: options.mode ?? \"auto\",\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tindexPhase: retrieved.indexPhase,\n\t\trrfK: retrieved.resolvedMode === \"hybrid\" ? retrieved.rrfK : undefined,\n\t\tretrievers: retrieved.retrievers,\n\t\tfused: retrieved.candidates.map(({ id, rrfScore, ranks, rawScores, mergedFrom }) => ({\n\t\t\tid,\n\t\t\trrfScore,\n\t\t\tranks,\n\t\t\trawScores,\n\t\t\t...(mergedFrom ? { mergedFrom } : {}),\n\t\t})),\n\t\trerank: retrieved.rerank,\n\t});\n\n\treturn {\n\t\ttext: assembled.text,\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tresultCount: retrieved.candidates.length,\n\t\tindexing: retrieved.indexing,\n\t};\n}\n"]}

package/dist/core/search/hybrid-search.js

@@ -92,6 +92,58 @@ export function hoistStaleCandidates(candidates, staleFiles) {
     }
     return [...hoisted, ...rest];
 }
+/**
+ * Fold overlapping or adjacent spans of the same file into their best-ranked
+ * occurrence.
+ *
+ * Chunks overlap by design (see the chunker's `CHUNK_OVERLAP_LINES`) and each
+ * one is a separate id, so neighbouring chunks of one region survive fusion as
+ * separate candidates and take separate result slots — showing the model code
+ * it already has. Measured on the 62-query set at the tool's default
+ * `limit=5`, 43 queries had such a pair in their top 5 and 58 of 310 slots
+ * went to repeated code.
+ *
+ * Merging is free recall: the union of two overlapping spans matches exactly
+ * what either matched, so nothing is gained by widening — the gain is entirely
+ * the slot handed back to the ranked tail. Recall@5 0.597 -> 0.677, Recall@1
+ * unchanged (merging cannot alter the top result).
+ *
+ * Adjacency counts as overlap (`endLine + 1`): two chunks that abut describe
+ * one continuous region, and rendering them as separate results implies a gap
+ * that is not there.
+ */
+export function mergeOverlappingSpans(candidates) {
+    const merged = [];
+    for (const candidate of candidates) {
+        const into = merged.find((m) => m.path === candidate.path && candidate.startLine <= m.endLine + 1 && m.startLine <= candidate.endLine + 1);
+        if (!into) {
+            merged.push({ ...candidate, ranks: { ...candidate.ranks }, rawScores: { ...candidate.rawScores } });
+            continue;
+        }
+        into.startLine = Math.min(into.startLine, candidate.startLine);
+        into.endLine = Math.max(into.endLine, candidate.endLine);
+        // Keep the folded id: the trace is the only record that retrieval found
+        // it, and dropping it would make a merged result look like a miss.
+        if (!into.mergedFrom)
+            into.mergedFrom = [];
+        into.mergedFrom.push(candidate.id);
+        // The merged region really was reached by every retriever that found any
+        // part of it, so the source label and the trace should say so. Best rank
+        // per source wins; the fused score stays the survivor's, which is what
+        // ordered it here.
+        for (const [source, rank] of Object.entries(candidate.ranks)) {
+            const held = into.ranks[source];
+            if (held === undefined || rank < held)
+                into.ranks[source] = rank;
+        }
+        for (const [source, score] of Object.entries(candidate.rawScores)) {
+            const held = into.rawScores[source];
+            if (held === undefined || score > held)
+                into.rawScores[source] = score;
+        }
+    }
+    return merged;
+}
 export async function retrieveCandidates(options) {
     const { cwd, query, service, signal } = options;
     const glob = normalizeSearchGlob(options.glob);
@@ -246,7 +298,9 @@ export async function retrieveCandidates(options) {
     // lone grep hit from an unindexed file has the lowest prior there is, so
     // hoisting first would simply be undone. The reranker still orders the
     // hoisted set against itself.
-    candidates = hoistStaleCandidates(candidates, staleSet).slice(0, limit);
+    // Merge before slicing, so a slot freed by a duplicate is refilled from the
+    // ranked tail rather than left empty.
+    candidates = mergeOverlappingSpans(hoistStaleCandidates(candidates, staleSet)).slice(0, limit);
     return {
         candidates,
         resolvedMode: mode,
@@ -270,7 +324,13 @@ export async function runSearch(options) {
         indexPhase: retrieved.indexPhase,
         rrfK: retrieved.resolvedMode === "hybrid" ? retrieved.rrfK : undefined,
         retrievers: retrieved.retrievers,
-        fused: retrieved.candidates.map(({ id, rrfScore, ranks, rawScores }) => ({ id, rrfScore, ranks, rawScores })),
+        fused: retrieved.candidates.map(({ id, rrfScore, ranks, rawScores, mergedFrom }) => ({
+            id,
+            rrfScore,
+            ranks,
+            rawScores,
+            ...(mergedFrom ? { mergedFrom } : {}),
+        })),
         rerank: retrieved.rerank,
     });
     return {

package/dist/core/search/hybrid-search.js.map

@@ -1 +1 @@
-{"version":3,"file":"hybrid-search.js","sourceRoot":"","sources":["../../../src/core/search/hybrid-search.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAGH,OAAO,EAAE,aAAa,EAAoB,MAAM,cAAc,CAAC;AAC/D,OAAO,EAAE,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACzD,OAAO,EAAE,kBAAkB,EAAE,MAAM,mBAAmB,CAAC;AACvD,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAC7D,OAAO,EAAE,iBAAiB,EAAE,MAAM,WAAW,CAAC;AAC9C,OAAO,EAAE,gBAAgB,EAAE,MAAM,aAAa,CAAC;AAC/C,OAAO,EAAE,aAAa,EAAE,OAAO,EAAE,MAAM,UAAU,CAAC;AAClD,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAG9C,2DAA2D;AAC3D,MAAM,mBAAmB,GAAG,GAAG,CAAC;AAChC;;;2EAG2E;AAC3E,MAAM,kBAAkB,GAAG,EAAE,CAAC;AAC9B;;;;;;;GAOG;AACH,MAAM,WAAW,GAAG,EAAE,CAAC;AACvB;;;;;;;;GAQG;AACH,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAC9B;;mEAEmE;AACnE,MAAM,UAAU,GAAG,iBAAiB,CAAC;AACrC;;qBAEqB;AACrB,MAAM,eAAe,GAAG,CAAC,CAAC;AAC1B,2DAA2D;AAC3D,MAAM,YAAY,GAAG,EAAE,CAAC;AAyExB,SAAS,mBAAmB,CAAC,IAAwB,EAAsB;IAC1E,IAAI,CAAC,IAAI;QAAE,OAAO,SAAS,CAAC;IAC5B,yEAAyE;IACzE,6EAA6E;IAC7E,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;QAC5E,OAAO,MAAM,IAAI,EAAE,CAAC;IACrB,CAAC;IACD,OAAO,IAAI,CAAC;AAAA,CACZ;AAED;;;;;;;;;;;;;;;GAeG;AACH,MAAM,UAAU,oBAAoB,CACnC,UAAqC,EACrC,UAA+B,EACZ;IACnB,IAAI,UAAU,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,GAAG,UAAU,CAAC,CAAC;IAClD,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,MAAM,IAAI,GAAqB,EAAE,CAAC;IAClC,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACpC,CAAC,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,MAAM,GAAG,eAAe,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IACvG,CAAC;IACD,OAAO,CAAC,GAAG,OAAO,EAAE,GAAG,IAAI,CAAC,CAAC;AAAA,CAC7B;AAED,MAAM,CAAC,KAAK,UAAU,kBAAkB,CAAC,OAAwB,EAA2B;IAC3F,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC;IAChD,MAAM,IAAI,GAAG,mBAAmB,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAC/C,MAAM,aAAa,GAAG,OAAO,CAAC,IAAI,IAAI,MAAM,CAAC;IAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC;IAC/C,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,aAAa,CAAC;IAE3C,MAAM,KAAK,GAAG,OAAO,EAAE,QAAQ,EAAE,CAAC;IAClC,MAAM,cAAc,GAAG,OAAO,EAAE,WAAW,EAAE,IAAI,KAAK,CAAC;IACvD,MAAM,sBAAsB,GAC3B,KAAK,KAAK,SAAS;QAClB,CAAC,CAAC,+BAA+B;QACjC,CAAC,CAAC,KAAK,CAAC,KAAK,KAAK,aAAa,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS;YAC3D,CAAC,CAAC,KAAK,CAAC,MAAM;YACd,CAAC,CAAC,KAAK,CAAC,KAAK,KAAK,MAAM;gBACvB,CAAC,CAAC,gCAAgC;gBAClC,CAAC,CAAC,SAAS,CAAC;IAEhB,MAAM,UAAU,GAAG,iBAAiB,CAAC,KAAK,EAAE,aAAa,EAAE,cAAc,EAAE,sBAAsB,CAAC,CAAC;IACnG,MAAM,IAAI,GAAG,UAAU,CAAC,IAAI,CAAC;IAE7B,0EAA0E;IAC1E,iEAAiE;IACjE,MAAM,WAAW,GAA4B,cAAc;QAC1D,CAAC,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,CAAC,OAAQ,CAAC,kBAAkB,CAAC,GAAG,EAAE,IAAI,CAAC;QACvD,CAAC,CAAC,SAAS,CAAC;IAEb,MAAM,KAAK,GAAG,IAAI,GAAG,EAAyB,CAAC;IAC/C;oEAC8D;IAC9D,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;IACnC,MAAM,KAAK,GAAkB,EAAE,CAAC;IAChC,MAAM,cAAc,GAA8B,EAAE,CAAC;IACrD,MAAM,MAAM,GAAY,EAAE,CAAC;IAE3B,MAAM,UAAU,GAAG,KAAK,EAAE,YAAY,GAAG,KAAK,EAAiB,EAAE,CAAC;QACjE,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC7B,IAAI,CAAC;YACJ,sEAAsE;YACtE,oEAAoE;YACpE,sEAAsE;YACtE,uEAAuE;YACvE,MAAM,KAAK,GAAG,YAAY,CAAC,CAAC,CAAC,OAAO,EAAE,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;YACrE,IAAI,KAAK;gBAAE,KAAK,MAAM,GAAG,IAAI,KAAK;oBAAE,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACtD,MAAM,QAAQ,GAAG,MAAM,mBAAmB,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,mBAAmB,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;YAC5G,MAAM,OAAO,GAAG,aAAa,CAAC,QAAQ,EAAE,WAAW,CAAC,CAAC;YACrD,mEAAmE;YACnE,oEAAoE;YACpE,MAAM,IAAI,GAAG,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC;YAC1F,KAAK,MAAM,CAAC,EAAE,EAAE,IAAI,CAAC,IAAI,OAAO,CAAC,KAAK;gBAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;oBAAE,KAAK,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;YAChF,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC;QACpF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3D,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;QAC1E,CAAC;IAAA,CACD,CAAC;IAEF,MAAM,QAAQ,GAAG,KAAK,IAAmB,EAAE,CAAC;QAC3C,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC7B,IAAI,CAAC;YACJ,kEAAkE;YAClE,mEAAmE;YACnE,yDAAyD;YACzD,4EAAwE;YACxE,sDAAsD;YACtD,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,WAAW,CAAC;YAC/D,MAAM,SAAS,GAAG,CACjB,MAAM,OAAQ,CAAC,YAAY,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CACzF,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;YACjC,MAAM,IAAI,GAAgB,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;gBACpD,EAAE,EAAE,GAAG,CAAC,EAAE;gBACV,IAAI,EAAE,CAAC,GAAG,CAAC;gBACX,KAAK,EAAE,GAAG,CAAC,KAAK;gBAChB,MAAM,EAAE,OAAO;aACf,CAAC,CAAC,CAAC;YACJ,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;gBAC7B,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,GAAG,CAAC,SAAS,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;YACvF,CAAC;YACD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,cAAc,CAAC,KAAK,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC;QACrF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3D,cAAc,CAAC,KAAK,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;QAC3E,CAAC;IAAA,CACD,CAAC;IAEF,MAAM,OAAO,GAAG,KAAK,IAAmB,EAAE,CAAC;QAC1C,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC7B,IAAI,CAAC;YACJ,iEAAiE;YACjE,qEAAqE;YACrE,+DAA6D;YAC7D,MAAM,SAAS,GAAG,MAAM,OAAQ,CAAC,YAAY,CAAC,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC;YAClF,MAAM,IAAI,GAAgB,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;gBACpD,EAAE,EAAE,GAAG,CAAC,EAAE;gBACV,IAAI,EAAE,CAAC,GAAG,CAAC;gBACX,KAAK,EAAE,GAAG,CAAC,KAAK;gBAChB,MAAM,EAAE,MAAM;aACd,CAAC,CAAC,CAAC;YACJ,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;gBAC7B,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,CAAC;oBACxB,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,GAAG,CAAC,SAAS,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;gBACvF,CAAC;YACF,CAAC;YACD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC;QACpF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3D,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;QAC1E,CAAC;IAAA,CACD,CAAC;IAEF,2EAA2E;IAC3E,yEAAyE;IACzE,2EAA2E;IAC3E,2EAA2E;IAC3E,8EAA4E;IAC5E,kEAAkE;IAClE,MAAM,gBAAgB,GACrB,CAAC,OAAO,CAAC,OAAO,IAAI,OAAO,EAAE,wBAAwB,EAAE,IAAI,KAAK,CAAC,IAAI,cAAc,IAAI,IAAI,KAAK,SAAS,CAAC;IAE3G,MAAM,IAAI,GAAoB,EAAE,CAAC;IACjC,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,QAAQ;QAAE,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,gBAAgB,CAAC,CAAC,CAAC;IACrF,IAAI,IAAI,KAAK,UAAU,IAAI,IAAI,KAAK,QAAQ;QAAE,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC;IACpE,wEAAwE;IACxE,sEAAsE;IACtE,IAAI,gBAAgB;QAAE,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC,CAAC;IAC3C,MAAM,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACxB,IAAI,MAAM,EAAE,OAAO;QAAE,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC;IAC1D,wEAAwE;IACxE,iCAAiC;IACjC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,MAAM,MAAM,CAAC,CAAC,CAAC,IAAI,IAAI,KAAK,CAAC,sCAAsC,CAAC,CAAC;IAE7F,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;IAC1D,IAAI,UAAU,GAAqB,EAAE,CAAC;IACtC,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC/B,IAAI,IAAI;YAAE,UAAU,CAAC,IAAI,CAAC,EAAE,GAAG,GAAG,EAAE,GAAG,IAAI,EAAE,CAAC,CAAC;IAChD,CAAC;IAED,IAAI,UAAiC,CAAC;IACtC,IAAI,OAAO,CAAC,YAAY,IAAI,OAAO,EAAE,oBAAoB,EAAE,EAAE,CAAC;QAC7D,MAAM,QAAQ,GAAG,MAAM,kBAAkB,CAAC,KAAK,EAAE,UAAU,EAAE,GAAG,EAAE,OAAO,CAAC,CAAC;QAC3E,UAAU,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,cAAc,EAAE,QAAQ,CAAC,MAAM,EAAE,SAAS,EAAE,QAAQ,CAAC,SAAS,EAAE,CAAC;QAC/F,UAAU,GAAG,QAAQ,CAAC,UAAU,CAAC;IAClC,CAAC;SAAM,IAAI,OAAO,CAAC,MAAM,KAAK,KAAK,EAAE,CAAC;QACrC,MAAM,QAAQ,GAAG,gBAAgB,CAAC,KAAK,EAAE,UAAU,EAAE,GAAG,CAAC,CAAC;QAC1D,UAAU,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,cAAc,EAAE,UAAU,CAAC,MAAM,EAAE,SAAS,EAAE,QAAQ,CAAC,SAAS,EAAE,CAAC;QACjG,UAAU,GAAG,QAAQ,CAAC,UAAU,CAAC;IAClC,CAAC;IACD,4EAA4E;IAC5E,yEAAyE;IACzE,uEAAuE;IACvE,8BAA8B;IAC9B,UAAU,GAAG,oBAAoB,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAExE,OAAO;QACN,UAAU;QACV,YAAY,EAAE,IAAI;QAClB,cAAc,EAAE,UAAU,CAAC,cAAc;QACzC,UAAU,EAAE,KAAK,EAAE,KAAK,KAAK,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,KAAK,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,aAAa;QACzG,UAAU,EAAE,cAAc;QAC1B,QAAQ,EAAE,KAAK,EAAE,KAAK,KAAK,UAAU,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS;QAC5F,IAAI;QACJ,MAAM,EAAE,UAAU;KAClB,CAAC;AAAA,CACF;AAED,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,OAAyB,EAA4B;IACpF,MAAM,SAAS,GAAG,MAAM,kBAAkB,CAAC,OAAO,CAAC,CAAC;IAEpD,MAAM,SAAS,GAAG,eAAe,CAAC,SAAS,CAAC,UAAU,EAAE,EAAE,GAAG,EAAE,OAAO,CAAC,GAAG,EAAE,WAAW,EAAE,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC;IAEhH,gBAAgB,CAAC,OAAO,CAAC,GAAG,EAAE;QAC7B,WAAW,EAAE,IAAI,CAAC,GAAG,EAAE;QACvB,KAAK,EAAE,OAAO,CAAC,KAAK;QACpB,aAAa,EAAE,OAAO,CAAC,IAAI,IAAI,MAAM;QACrC,YAAY,EAAE,SAAS,CAAC,YAAY;QACpC,cAAc,EAAE,SAAS,CAAC,cAAc;QACxC,UAAU,EAAE,SAAS,CAAC,UAAU;QAChC,IAAI,EAAE,SAAS,CAAC,YAAY,KAAK,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS;QACtE,UAAU,EAAE,SAAS,CAAC,UAAU;QAChC,KAAK,EAAE,SAAS,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC,CAAC;QAC7G,MAAM,EAAE,SAAS,CAAC,MAAM;KACxB,CAAC,CAAC;IAEH,OAAO;QACN,IAAI,EAAE,SAAS,CAAC,IAAI;QACpB,YAAY,EAAE,SAAS,CAAC,YAAY;QACpC,cAAc,EAAE,SAAS,CAAC,cAAc;QACxC,WAAW,EAAE,SAAS,CAAC,UAAU,CAAC,MAAM;QACxC,QAAQ,EAAE,SAAS,CAAC,QAAQ;KAC5B,CAAC;AAAA,CACF","sourcesContent":["/**\n * Hybrid search orchestrator: resolve mode, run retrievers in parallel, fuse\n * by rank, expand within budget, trace everything\n * (docs/hybrid-retrieval-design.md).\n *\n * Single-retriever modes flow through the same pipeline — rrfFuse over one\n * list preserves its order — so lexical, semantic, and hybrid all produce the\n * same result shape and the same trace record.\n *\n * `retrieveCandidates` is the candidate-level core (also used by the eval\n * harness, which needs forced modes, a configurable `k`, and no trace\n * pollution); `runSearch` wraps it with span expansion and tracing for the\n * tool.\n */\n\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport { adaptGrepHits, type ChunkLookup } from \"./adapter.js\";\nimport { assembleContext } from \"./context-assembler.js\";\nimport { crossEncoderRerank } from \"./cross-rerank.js\";\nimport { runLexicalRetriever } from \"./lexical-retriever.js\";\nimport { resolveSearchMode } from \"./mode.js\";\nimport { rerankCandidates } from \"./rerank.js\";\nimport { DEFAULT_RRF_K, rrfFuse } from \"./rrf.js\";\nimport { writeSearchTrace } from \"./trace.js\";\nimport type { CandidateSpan, FusedCandidate, RankedHit, ResolvedSearchMode, SearchMode, SearchTrace } from \"./types.js\";\n\n/** Raw grep line-hits fetched per query (pre-collapse). */\nconst LEXICAL_MATCH_LIMIT = 200;\n/** Adapted lexical candidates entering fusion. The eval gate showed the\n *  uncapped lexical tail diluting hybrid below plain semantic: lexical\n *  precision is front-loaded by the adapter's term-evidence ranking, while\n *  RRF weighs a rank-30 lexical candidate like a rank-30 embedding hit. */\nconst LEXICAL_FUSION_CAP = 20;\n/**\n * Embedding hits fetched per query when fusing with the grep leg.\n *\n * Deliberately shallow. The grep list is capped at {@link LEXICAL_FUSION_CAP}\n * because its tail is unranked noise, so a deep dense pool here simply\n * outnumbers it: raising this to 200 alongside a 20-candidate grep list cost\n * `auto +rr` 0.472 -> 0.450 MRR.\n */\nconst EMBED_TOP_K = 50;\n/**\n * Per-leg depth when both retrievers are ranked ones (dense + BM25).\n *\n * Four times {@link FUSED_WINDOW}, matching what the daemon's `query_hybrid`\n * does internally (`pool = 4·k`). Fusing at the same depth as the window loses\n * any candidate ranked well by one retriever but just outside the other's\n * top-50 — worth 12pp R@10 and 8pp R@50 for `bm25+dense`, which is exactly\n * what closed the gap to daemon-side fusion.\n */\nconst FUSION_POOL_TOP_K = 200;\n/** BM25 depth, matching {@link FUSION_POOL_TOP_K} so neither ranked leg is\n *  structurally advantaged by pool size. The daemon returns only documents\n *  sharing a query term, so this is an upper bound, not a fill. */\nconst BM25_TOP_K = FUSION_POOL_TOP_K;\n/** Candidates from unindexed/stale files moved ahead of the fused window.\n *  Small on purpose: this answers \"what did I just write\", not \"search the\n *  working tree\". */\nconst STALE_HOIST_CAP = 5;\n/** Fused candidates kept for reranking / final slicing. */\nconst FUSED_WINDOW = 50;\n\nexport interface RetrieveOptions {\n\tcwd: string;\n\tquery: string;\n\tmode?: SearchMode;\n\t/** Optional glob filter applied to file paths. */\n\tglob?: string;\n\t/** Maximum fused candidates returned. */\n\tlimit?: number;\n\t/** RRF constant override (eval harness sweeps this). Default: {@link DEFAULT_RRF_K}. */\n\trrfK?: number;\n\t/** Rerank the fused top-50 before slicing to `limit`. Default: true. */\n\trerank?: boolean;\n\t/**\n\t * Ask the daemon to fuse its own BM25 index with the vectors and return one\n\t * already-fused ranking, instead of taking a dense-only list. Needs a store\n\t * built with `--hybrid`. The fused list arrives as a single \"embed\" leg,\n\t * because a pre-fused ranking has no per-retriever structure left to record.\n\t *\n\t * Prefer {@link bm25Leg}: fusing here keeps the legs separable in the trace\n\t * and lets the grep leg participate.\n\t */\n\tdaemonHybrid?: boolean;\n\t/**\n\t * Fetch the daemon's BM25 index as its own ranked list and fuse it here,\n\t * alongside dense and grep.\n\t *\n\t * Defaults to on wherever the daemon can serve it, because BM25 is the\n\t * better lexical leg on the indexed corpus: Recall@50 0.790 -> 0.879, 6 of\n\t * 62 queries better and 0 worse (p <= 0.05). When it is on, the grep leg\n\t * narrows to files the index has not read — see `staleFiles`.\n\t *\n\t * Set `false` to force ripgrep as the only lexical leg; the eval harness\n\t * does this to keep measuring what the old rows measured.\n\t */\n\tbm25Leg?: boolean;\n\t/**\n\t * Reorder the fused shortlist with the daemon's cross-encoder instead of\n\t * the deterministic reranker. Needs embsearch >= 0.3.0; costs one model\n\t * pass per scored candidate.\n\t */\n\tcrossEncoder?: boolean;\n\tservice?: EmbsearchService;\n\tsignal?: AbortSignal;\n}\n\nexport interface RetrieveResult {\n\tcandidates: FusedCandidate[];\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tindexPhase: SearchTrace[\"indexPhase\"];\n\tretrievers: SearchTrace[\"retrievers\"];\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n\trrfK: number;\n\trerank?: SearchTrace[\"rerank\"];\n}\n\nexport interface RunSearchOptions extends RetrieveOptions {\n\t/** Approximate token budget for the result text. */\n\ttokenBudget?: number;\n}\n\nexport interface RunSearchResult {\n\ttext: string;\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tresultCount: number;\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n}\n\nfunction normalizeSearchGlob(glob: string | undefined): string | undefined {\n\tif (!glob) return undefined;\n\t// Match fd/rg semantics: a slash-containing glob is anchored anywhere in\n\t// the tree, so prepend \"**/\" unless it already starts with a slash or \"**/\".\n\tif (glob.includes(\"/\") && !glob.startsWith(\"/\") && !glob.startsWith(\"**/\")) {\n\t\treturn `**/${glob}`;\n\t}\n\treturn glob;\n}\n\n/**\n * Move candidates from files the index has not read to the front.\n *\n * They are there because grep found them and nothing else could: the index is\n * ranking a stale copy of the file, or has never seen it. Left to fuse, they\n * lose — RRF rewards agreement, and one leg reporting a single document is\n * outvoted by two legs agreeing on hundreds. Measured: scoping grep to stale\n * files without this hoist scored 25% on the live-edit set where unscoped grep\n * scored 100%, because the fused window filled with consensus hits about the\n * indexed copy.\n *\n * Capped, because \"stale\" scales with how far behind the index is. A few\n * edited files is the case this exists for; a fresh checkout makes everything\n * stale, and hoisting all of it would quietly turn hybrid search back into\n * grep. Past the cap the remainder keeps its fused position.\n */\nexport function hoistStaleCandidates(\n\tcandidates: readonly FusedCandidate[],\n\tstaleFiles: ReadonlySet<string>,\n): FusedCandidate[] {\n\tif (staleFiles.size === 0) return [...candidates];\n\tconst hoisted: FusedCandidate[] = [];\n\tconst rest: FusedCandidate[] = [];\n\tfor (const candidate of candidates) {\n\t\t(staleFiles.has(candidate.path) && hoisted.length < STALE_HOIST_CAP ? hoisted : rest).push(candidate);\n\t}\n\treturn [...hoisted, ...rest];\n}\n\nexport async function retrieveCandidates(options: RetrieveOptions): Promise<RetrieveResult> {\n\tconst { cwd, query, service, signal } = options;\n\tconst glob = normalizeSearchGlob(options.glob);\n\tconst requestedMode = options.mode ?? \"auto\";\n\tconst limit = Math.max(1, options.limit ?? 10);\n\tconst rrfK = options.rrfK ?? DEFAULT_RRF_K;\n\n\tconst state = service?.getState();\n\tconst embedAvailable = service?.isAvailable() ?? false;\n\tconst embedUnavailableReason =\n\t\tstate === undefined\n\t\t\t? \"semantic index is not enabled\"\n\t\t\t: state.phase === \"unavailable\" || state.phase === \"skipped\"\n\t\t\t\t? state.reason\n\t\t\t\t: state.phase === \"idle\"\n\t\t\t\t\t? \"semantic index has not started\"\n\t\t\t\t\t: undefined;\n\n\tconst resolution = resolveSearchMode(query, requestedMode, embedAvailable, embedUnavailableReason);\n\tconst mode = resolution.mode;\n\n\t// Map lexical hits onto indexed chunk ids whenever the sidecar is usable,\n\t// even in lexical-only mode, so identities line up across modes.\n\tconst lookupChunk: ChunkLookup | undefined = embedAvailable\n\t\t? (rel, line) => service!.findEnclosingChunk(rel, line)\n\t\t: undefined;\n\n\tconst spans = new Map<string, CandidateSpan>();\n\t/** Files the index does not have current content for, when grep is scoped\n\t *  to them. Read again after fusion — see the hoist below. */\n\tconst staleSet = new Set<string>();\n\tconst lists: RankedHit[][] = [];\n\tconst retrieverStats: SearchTrace[\"retrievers\"] = {};\n\tconst errors: Error[] = [];\n\n\tconst runLexical = async (scopeToStale = false): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Scoped run: grep covers only what the index has not read yet, so it\n\t\t\t// adds the one thing BM25 cannot see without re-voting on documents\n\t\t\t// BM25 already ranked. Fusing two lexical views of the same corpus is\n\t\t\t// what made the three-leg configuration lose (18 of 62 queries worse).\n\t\t\tconst paths = scopeToStale ? service?.staleFiles(signal) : undefined;\n\t\t\tif (paths) for (const rel of paths) staleSet.add(rel);\n\t\t\tconst lineHits = await runLexicalRetriever({ cwd, query, limit: LEXICAL_MATCH_LIMIT, glob, signal, paths });\n\t\t\tconst adapted = adaptGrepHits(lineHits, lookupChunk);\n\t\t\t// In single-retriever lexical mode the full list is the result; in\n\t\t\t// hybrid, only the front-loaded head is trustworthy enough to vote.\n\t\t\tconst hits = mode === \"hybrid\" ? adapted.hits.slice(0, LEXICAL_FUSION_CAP) : adapted.hits;\n\t\t\tfor (const [id, span] of adapted.spans) if (!spans.has(id)) spans.set(id, span);\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runEmbed = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// The flat index pads top-k with whatever exists; with the cosine\n\t\t\t// metric the store uses, score <= 0 means \"no relation at all\", so\n\t\t\t// those padding hits would cast RRF votes on pure noise.\n\t\t\t// A BM25 leg is itself ranked, so the pair can afford — and needs — the\n\t\t\t// deeper pool; the grep leg cannot (see EMBED_TOP_K).\n\t\t\tconst topK = options.bm25Leg ? FUSION_POOL_TOP_K : EMBED_TOP_K;\n\t\t\tconst chunkHits = (\n\t\t\t\tawait service!.searchChunks(query, topK, glob, options.daemonHybrid ? \"hybrid\" : \"dense\")\n\t\t\t).filter((hit) => hit.score > 0);\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"embed\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runBm25 = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Raw BM25 sums; only the ordering enters fusion, the score is a\n\t\t\t// diagnostic. Unlike the dense leg there is no zero-score padding to\n\t\t\t// filter — the daemon omits documents sharing no query term.\n\t\t\tconst chunkHits = await service!.searchChunks(query, BM25_TOP_K, glob, \"lexical\");\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"bm25\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tif (!spans.has(hit.id)) {\n\t\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t\t}\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\t// BM25 is the better lexical leg where the index is current: it beats grep\n\t// on Recall@50 (0.790 -> 0.879, 6 queries better and 0 worse, p <= 0.05)\n\t// because it ranks the whole corpus rather than truncating a match stream.\n\t// It is also blind to anything indexed later than it was written, which is\n\t// precisely where grep still wins — so grep runs scoped to that set instead\n\t// of being dropped or left to duplicate BM25 over the whole tree.\n\tconst bm25AsLexicalLeg =\n\t\t(options.bm25Leg ?? service?.supportsLexicalRetriever() ?? false) && embedAvailable && mode !== \"lexical\";\n\n\tconst runs: Promise<void>[] = [];\n\tif (mode === \"lexical\" || mode === \"hybrid\") runs.push(runLexical(bm25AsLexicalLeg));\n\tif (mode === \"semantic\" || mode === \"hybrid\") runs.push(runEmbed());\n\t// `semantic` gets the BM25 leg too: the caller asked for the index, and\n\t// BM25 is part of it. Only an explicit `lexical` request excludes it.\n\tif (bm25AsLexicalLeg) runs.push(runBm25());\n\tawait Promise.all(runs);\n\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\t// A partial failure in hybrid degrades to whichever retriever survived;\n\t// only a total loss is an error.\n\tif (lists.length === 0) throw errors[0] ?? new Error(\"search produced no retriever results\");\n\n\tconst fused = rrfFuse(lists, rrfK).slice(0, FUSED_WINDOW);\n\tlet candidates: FusedCandidate[] = [];\n\tfor (const hit of fused) {\n\t\tconst span = spans.get(hit.id);\n\t\tif (span) candidates.push({ ...hit, ...span });\n\t}\n\n\tlet rerankInfo: SearchTrace[\"rerank\"];\n\tif (options.crossEncoder && service?.supportsCrossEncoder()) {\n\t\tconst reranked = await crossEncoderRerank(query, candidates, cwd, service);\n\t\trerankInfo = { applied: true, candidateCount: reranked.scored, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t} else if (options.rerank !== false) {\n\t\tconst reranked = rerankCandidates(query, candidates, cwd);\n\t\trerankInfo = { applied: true, candidateCount: candidates.length, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t}\n\t// After reranking, not before: both rerankers weight the fused prior, and a\n\t// lone grep hit from an unindexed file has the lowest prior there is, so\n\t// hoisting first would simply be undone. The reranker still orders the\n\t// hoisted set against itself.\n\tcandidates = hoistStaleCandidates(candidates, staleSet).slice(0, limit);\n\n\treturn {\n\t\tcandidates,\n\t\tresolvedMode: mode,\n\t\tdegradedReason: resolution.degradedReason,\n\t\tindexPhase: state?.phase === \"ready\" ? \"ready\" : state?.phase === \"indexing\" ? \"indexing\" : \"unavailable\",\n\t\tretrievers: retrieverStats,\n\t\tindexing: state?.phase === \"indexing\" ? { done: state.done, total: state.total } : undefined,\n\t\trrfK,\n\t\trerank: rerankInfo,\n\t};\n}\n\nexport async function runSearch(options: RunSearchOptions): Promise<RunSearchResult> {\n\tconst retrieved = await retrieveCandidates(options);\n\n\tconst assembled = assembleContext(retrieved.candidates, { cwd: options.cwd, tokenBudget: options.tokenBudget });\n\n\twriteSearchTrace(options.cwd, {\n\t\ttimestampMs: Date.now(),\n\t\tquery: options.query,\n\t\trequestedMode: options.mode ?? \"auto\",\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tindexPhase: retrieved.indexPhase,\n\t\trrfK: retrieved.resolvedMode === \"hybrid\" ? retrieved.rrfK : undefined,\n\t\tretrievers: retrieved.retrievers,\n\t\tfused: retrieved.candidates.map(({ id, rrfScore, ranks, rawScores }) => ({ id, rrfScore, ranks, rawScores })),\n\t\trerank: retrieved.rerank,\n\t});\n\n\treturn {\n\t\ttext: assembled.text,\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tresultCount: retrieved.candidates.length,\n\t\tindexing: retrieved.indexing,\n\t};\n}\n"]}
+{"version":3,"file":"hybrid-search.js","sourceRoot":"","sources":["../../../src/core/search/hybrid-search.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAGH,OAAO,EAAE,aAAa,EAAoB,MAAM,cAAc,CAAC;AAC/D,OAAO,EAAE,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACzD,OAAO,EAAE,kBAAkB,EAAE,MAAM,mBAAmB,CAAC;AACvD,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAC7D,OAAO,EAAE,iBAAiB,EAAE,MAAM,WAAW,CAAC;AAC9C,OAAO,EAAE,gBAAgB,EAAE,MAAM,aAAa,CAAC;AAC/C,OAAO,EAAE,aAAa,EAAE,OAAO,EAAE,MAAM,UAAU,CAAC;AAClD,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC;AAG9C,2DAA2D;AAC3D,MAAM,mBAAmB,GAAG,GAAG,CAAC;AAChC;;;2EAG2E;AAC3E,MAAM,kBAAkB,GAAG,EAAE,CAAC;AAC9B;;;;;;;GAOG;AACH,MAAM,WAAW,GAAG,EAAE,CAAC;AACvB;;;;;;;;GAQG;AACH,MAAM,iBAAiB,GAAG,GAAG,CAAC;AAC9B;;mEAEmE;AACnE,MAAM,UAAU,GAAG,iBAAiB,CAAC;AACrC;;qBAEqB;AACrB,MAAM,eAAe,GAAG,CAAC,CAAC;AAC1B,2DAA2D;AAC3D,MAAM,YAAY,GAAG,EAAE,CAAC;AAyExB,SAAS,mBAAmB,CAAC,IAAwB,EAAsB;IAC1E,IAAI,CAAC,IAAI;QAAE,OAAO,SAAS,CAAC;IAC5B,yEAAyE;IACzE,6EAA6E;IAC7E,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;QAC5E,OAAO,MAAM,IAAI,EAAE,CAAC;IACrB,CAAC;IACD,OAAO,IAAI,CAAC;AAAA,CACZ;AAED;;;;;;;;;;;;;;;GAeG;AACH,MAAM,UAAU,oBAAoB,CACnC,UAAqC,EACrC,UAA+B,EACZ;IACnB,IAAI,UAAU,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,GAAG,UAAU,CAAC,CAAC;IAClD,MAAM,OAAO,GAAqB,EAAE,CAAC;IACrC,MAAM,IAAI,GAAqB,EAAE,CAAC;IAClC,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACpC,CAAC,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,MAAM,GAAG,eAAe,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;IACvG,CAAC;IACD,OAAO,CAAC,GAAG,OAAO,EAAE,GAAG,IAAI,CAAC,CAAC;AAAA,CAC7B;AAED;;;;;;;;;;;;;;;;;;;GAmBG;AACH,MAAM,UAAU,qBAAqB,CAAC,UAAqC,EAAoB;IAC9F,MAAM,MAAM,GAAqB,EAAE,CAAC;IACpC,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACpC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CACvB,CAAC,CAAC,EAAE,EAAE,CACL,CAAC,CAAC,IAAI,KAAK,SAAS,CAAC,IAAI,IAAI,SAAS,CAAC,SAAS,IAAI,CAAC,CAAC,OAAO,GAAG,CAAC,IAAI,CAAC,CAAC,SAAS,IAAI,SAAS,CAAC,OAAO,GAAG,CAAC,CAC1G,CAAC;QACF,IAAI,CAAC,IAAI,EAAE,CAAC;YACX,MAAM,CAAC,IAAI,CAAC,EAAE,GAAG,SAAS,EAAE,KAAK,EAAE,EAAE,GAAG,SAAS,CAAC,KAAK,EAAE,EAAE,SAAS,EAAE,EAAE,GAAG,SAAS,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC;YACpG,SAAS;QACV,CAAC;QACD,IAAI,CAAC,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,SAAS,CAAC,SAAS,CAAC,CAAC;QAC/D,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,OAAO,EAAE,SAAS,CAAC,OAAO,CAAC,CAAC;QACzD,wEAAwE;QACxE,mEAAmE;QACnE,IAAI,CAAC,IAAI,CAAC,UAAU;YAAE,IAAI,CAAC,UAAU,GAAG,EAAE,CAAC;QAC3C,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,CAAC,CAAC;QACnC,yEAAyE;QACzE,yEAAyE;QACzE,uEAAuE;QACvE,mBAAmB;QACnB,KAAK,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC;YAC9D,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,MAAiC,CAAC,CAAC;YAC3D,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,GAAG,IAAI;gBAAE,IAAI,CAAC,KAAK,CAAC,MAAiC,CAAC,GAAG,IAAI,CAAC;QAC7F,CAAC;QACD,KAAK,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,SAAS,CAAC,EAAE,CAAC;YACnE,MAAM,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,MAAqC,CAAC,CAAC;YACnE,IAAI,IAAI,KAAK,SAAS,IAAI,KAAK,GAAG,IAAI;gBAAE,IAAI,CAAC,SAAS,CAAC,MAAqC,CAAC,GAAG,KAAK,CAAC;QACvG,CAAC;IACF,CAAC;IACD,OAAO,MAAM,CAAC;AAAA,CACd;AAED,MAAM,CAAC,KAAK,UAAU,kBAAkB,CAAC,OAAwB,EAA2B;IAC3F,MAAM,EAAE,GAAG,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC;IAChD,MAAM,IAAI,GAAG,mBAAmB,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAC/C,MAAM,aAAa,GAAG,OAAO,CAAC,IAAI,IAAI,MAAM,CAAC;IAC7C,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC;IAC/C,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,aAAa,CAAC;IAE3C,MAAM,KAAK,GAAG,OAAO,EAAE,QAAQ,EAAE,CAAC;IAClC,MAAM,cAAc,GAAG,OAAO,EAAE,WAAW,EAAE,IAAI,KAAK,CAAC;IACvD,MAAM,sBAAsB,GAC3B,KAAK,KAAK,SAAS;QAClB,CAAC,CAAC,+BAA+B;QACjC,CAAC,CAAC,KAAK,CAAC,KAAK,KAAK,aAAa,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS;YAC3D,CAAC,CAAC,KAAK,CAAC,MAAM;YACd,CAAC,CAAC,KAAK,CAAC,KAAK,KAAK,MAAM;gBACvB,CAAC,CAAC,gCAAgC;gBAClC,CAAC,CAAC,SAAS,CAAC;IAEhB,MAAM,UAAU,GAAG,iBAAiB,CAAC,KAAK,EAAE,aAAa,EAAE,cAAc,EAAE,sBAAsB,CAAC,CAAC;IACnG,MAAM,IAAI,GAAG,UAAU,CAAC,IAAI,CAAC;IAE7B,0EAA0E;IAC1E,iEAAiE;IACjE,MAAM,WAAW,GAA4B,cAAc;QAC1D,CAAC,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,CAAC,OAAQ,CAAC,kBAAkB,CAAC,GAAG,EAAE,IAAI,CAAC;QACvD,CAAC,CAAC,SAAS,CAAC;IAEb,MAAM,KAAK,GAAG,IAAI,GAAG,EAAyB,CAAC;IAC/C;oEAC8D;IAC9D,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;IACnC,MAAM,KAAK,GAAkB,EAAE,CAAC;IAChC,MAAM,cAAc,GAA8B,EAAE,CAAC;IACrD,MAAM,MAAM,GAAY,EAAE,CAAC;IAE3B,MAAM,UAAU,GAAG,KAAK,EAAE,YAAY,GAAG,KAAK,EAAiB,EAAE,CAAC;QACjE,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC7B,IAAI,CAAC;YACJ,sEAAsE;YACtE,oEAAoE;YACpE,sEAAsE;YACtE,uEAAuE;YACvE,MAAM,KAAK,GAAG,YAAY,CAAC,CAAC,CAAC,OAAO,EAAE,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;YACrE,IAAI,KAAK;gBAAE,KAAK,MAAM,GAAG,IAAI,KAAK;oBAAE,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACtD,MAAM,QAAQ,GAAG,MAAM,mBAAmB,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,KAAK,EAAE,mBAAmB,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;YAC5G,MAAM,OAAO,GAAG,aAAa,CAAC,QAAQ,EAAE,WAAW,CAAC,CAAC;YACrD,mEAAmE;YACnE,oEAAoE;YACpE,MAAM,IAAI,GAAG,IAAI,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC;YAC1F,KAAK,MAAM,CAAC,EAAE,EAAE,IAAI,CAAC,IAAI,OAAO,CAAC,KAAK;gBAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;oBAAE,KAAK,CAAC,GAAG,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC;YAChF,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC;QACpF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3D,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;QAC1E,CAAC;IAAA,CACD,CAAC;IAEF,MAAM,QAAQ,GAAG,KAAK,IAAmB,EAAE,CAAC;QAC3C,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC7B,IAAI,CAAC;YACJ,kEAAkE;YAClE,mEAAmE;YACnE,yDAAyD;YACzD,4EAAwE;YACxE,sDAAsD;YACtD,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,WAAW,CAAC;YAC/D,MAAM,SAAS,GAAG,CACjB,MAAM,OAAQ,CAAC,YAAY,CAAC,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CACzF,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;YACjC,MAAM,IAAI,GAAgB,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;gBACpD,EAAE,EAAE,GAAG,CAAC,EAAE;gBACV,IAAI,EAAE,CAAC,GAAG,CAAC;gBACX,KAAK,EAAE,GAAG,CAAC,KAAK;gBAChB,MAAM,EAAE,OAAO;aACf,CAAC,CAAC,CAAC;YACJ,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;gBAC7B,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,GAAG,CAAC,SAAS,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;YACvF,CAAC;YACD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,cAAc,CAAC,KAAK,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC;QACrF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3D,cAAc,CAAC,KAAK,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;QAC3E,CAAC;IAAA,CACD,CAAC;IAEF,MAAM,OAAO,GAAG,KAAK,IAAmB,EAAE,CAAC;QAC1C,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAC7B,IAAI,CAAC;YACJ,iEAAiE;YACjE,qEAAqE;YACrE,+DAA6D;YAC7D,MAAM,SAAS,GAAG,MAAM,OAAQ,CAAC,YAAY,CAAC,KAAK,EAAE,UAAU,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC;YAClF,MAAM,IAAI,GAAgB,SAAS,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;gBACpD,EAAE,EAAE,GAAG,CAAC,EAAE;gBACV,IAAI,EAAE,CAAC,GAAG,CAAC;gBACX,KAAK,EAAE,GAAG,CAAC,KAAK;gBAChB,MAAM,EAAE,MAAM;aACd,CAAC,CAAC,CAAC;YACJ,KAAK,MAAM,GAAG,IAAI,SAAS,EAAE,CAAC;gBAC7B,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,CAAC;oBACxB,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,SAAS,EAAE,GAAG,CAAC,SAAS,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,CAAC,CAAC;gBACvF,CAAC;YACF,CAAC;YACD,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,IAAI,CAAC,MAAM,EAAE,CAAC;QACpF,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACZ,MAAM,CAAC,IAAI,CAAC,CAAC,YAAY,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAC3D,cAAc,CAAC,IAAI,GAAG,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;QAC1E,CAAC;IAAA,CACD,CAAC;IAEF,2EAA2E;IAC3E,yEAAyE;IACzE,2EAA2E;IAC3E,2EAA2E;IAC3E,8EAA4E;IAC5E,kEAAkE;IAClE,MAAM,gBAAgB,GACrB,CAAC,OAAO,CAAC,OAAO,IAAI,OAAO,EAAE,wBAAwB,EAAE,IAAI,KAAK,CAAC,IAAI,cAAc,IAAI,IAAI,KAAK,SAAS,CAAC;IAE3G,MAAM,IAAI,GAAoB,EAAE,CAAC;IACjC,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,KAAK,QAAQ;QAAE,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,gBAAgB,CAAC,CAAC,CAAC;IACrF,IAAI,IAAI,KAAK,UAAU,IAAI,IAAI,KAAK,QAAQ;QAAE,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC;IACpE,wEAAwE;IACxE,sEAAsE;IACtE,IAAI,gBAAgB;QAAE,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC,CAAC;IAC3C,MAAM,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACxB,IAAI,MAAM,EAAE,OAAO;QAAE,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC;IAC1D,wEAAwE;IACxE,iCAAiC;IACjC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,MAAM,MAAM,CAAC,CAAC,CAAC,IAAI,IAAI,KAAK,CAAC,sCAAsC,CAAC,CAAC;IAE7F,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,YAAY,CAAC,CAAC;IAC1D,IAAI,UAAU,GAAqB,EAAE,CAAC;IACtC,KAAK,MAAM,GAAG,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC/B,IAAI,IAAI;YAAE,UAAU,CAAC,IAAI,CAAC,EAAE,GAAG,GAAG,EAAE,GAAG,IAAI,EAAE,CAAC,CAAC;IAChD,CAAC;IAED,IAAI,UAAiC,CAAC;IACtC,IAAI,OAAO,CAAC,YAAY,IAAI,OAAO,EAAE,oBAAoB,EAAE,EAAE,CAAC;QAC7D,MAAM,QAAQ,GAAG,MAAM,kBAAkB,CAAC,KAAK,EAAE,UAAU,EAAE,GAAG,EAAE,OAAO,CAAC,CAAC;QAC3E,UAAU,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,cAAc,EAAE,QAAQ,CAAC,MAAM,EAAE,SAAS,EAAE,QAAQ,CAAC,SAAS,EAAE,CAAC;QAC/F,UAAU,GAAG,QAAQ,CAAC,UAAU,CAAC;IAClC,CAAC;SAAM,IAAI,OAAO,CAAC,MAAM,KAAK,KAAK,EAAE,CAAC;QACrC,MAAM,QAAQ,GAAG,gBAAgB,CAAC,KAAK,EAAE,UAAU,EAAE,GAAG,CAAC,CAAC;QAC1D,UAAU,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,cAAc,EAAE,UAAU,CAAC,MAAM,EAAE,SAAS,EAAE,QAAQ,CAAC,SAAS,EAAE,CAAC;QACjG,UAAU,GAAG,QAAQ,CAAC,UAAU,CAAC;IAClC,CAAC;IACD,4EAA4E;IAC5E,yEAAyE;IACzE,uEAAuE;IACvE,8BAA8B;IAC9B,4EAA4E;IAC5E,sCAAsC;IACtC,UAAU,GAAG,qBAAqB,CAAC,oBAAoB,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAE/F,OAAO;QACN,UAAU;QACV,YAAY,EAAE,IAAI;QAClB,cAAc,EAAE,UAAU,CAAC,cAAc;QACzC,UAAU,EAAE,KAAK,EAAE,KAAK,KAAK,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,EAAE,KAAK,KAAK,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,aAAa;QACzG,UAAU,EAAE,cAAc;QAC1B,QAAQ,EAAE,KAAK,EAAE,KAAK,KAAK,UAAU,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS;QAC5F,IAAI;QACJ,MAAM,EAAE,UAAU;KAClB,CAAC;AAAA,CACF;AAED,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,OAAyB,EAA4B;IACpF,MAAM,SAAS,GAAG,MAAM,kBAAkB,CAAC,OAAO,CAAC,CAAC;IAEpD,MAAM,SAAS,GAAG,eAAe,CAAC,SAAS,CAAC,UAAU,EAAE,EAAE,GAAG,EAAE,OAAO,CAAC,GAAG,EAAE,WAAW,EAAE,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC;IAEhH,gBAAgB,CAAC,OAAO,CAAC,GAAG,EAAE;QAC7B,WAAW,EAAE,IAAI,CAAC,GAAG,EAAE;QACvB,KAAK,EAAE,OAAO,CAAC,KAAK;QACpB,aAAa,EAAE,OAAO,CAAC,IAAI,IAAI,MAAM;QACrC,YAAY,EAAE,SAAS,CAAC,YAAY;QACpC,cAAc,EAAE,SAAS,CAAC,cAAc;QACxC,UAAU,EAAE,SAAS,CAAC,UAAU;QAChC,IAAI,EAAE,SAAS,CAAC,YAAY,KAAK,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS;QACtE,UAAU,EAAE,SAAS,CAAC,UAAU;QAChC,KAAK,EAAE,SAAS,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAE,UAAU,EAAE,EAAE,EAAE,CAAC,CAAC;YACpF,EAAE;YACF,QAAQ;YACR,KAAK;YACL,SAAS;YACT,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SACrC,CAAC,CAAC;QACH,MAAM,EAAE,SAAS,CAAC,MAAM;KACxB,CAAC,CAAC;IAEH,OAAO;QACN,IAAI,EAAE,SAAS,CAAC,IAAI;QACpB,YAAY,EAAE,SAAS,CAAC,YAAY;QACpC,cAAc,EAAE,SAAS,CAAC,cAAc;QACxC,WAAW,EAAE,SAAS,CAAC,UAAU,CAAC,MAAM;QACxC,QAAQ,EAAE,SAAS,CAAC,QAAQ;KAC5B,CAAC;AAAA,CACF","sourcesContent":["/**\n * Hybrid search orchestrator: resolve mode, run retrievers in parallel, fuse\n * by rank, expand within budget, trace everything\n * (docs/hybrid-retrieval-design.md).\n *\n * Single-retriever modes flow through the same pipeline — rrfFuse over one\n * list preserves its order — so lexical, semantic, and hybrid all produce the\n * same result shape and the same trace record.\n *\n * `retrieveCandidates` is the candidate-level core (also used by the eval\n * harness, which needs forced modes, a configurable `k`, and no trace\n * pollution); `runSearch` wraps it with span expansion and tracing for the\n * tool.\n */\n\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport { adaptGrepHits, type ChunkLookup } from \"./adapter.js\";\nimport { assembleContext } from \"./context-assembler.js\";\nimport { crossEncoderRerank } from \"./cross-rerank.js\";\nimport { runLexicalRetriever } from \"./lexical-retriever.js\";\nimport { resolveSearchMode } from \"./mode.js\";\nimport { rerankCandidates } from \"./rerank.js\";\nimport { DEFAULT_RRF_K, rrfFuse } from \"./rrf.js\";\nimport { writeSearchTrace } from \"./trace.js\";\nimport type { CandidateSpan, FusedCandidate, RankedHit, ResolvedSearchMode, SearchMode, SearchTrace } from \"./types.js\";\n\n/** Raw grep line-hits fetched per query (pre-collapse). */\nconst LEXICAL_MATCH_LIMIT = 200;\n/** Adapted lexical candidates entering fusion. The eval gate showed the\n *  uncapped lexical tail diluting hybrid below plain semantic: lexical\n *  precision is front-loaded by the adapter's term-evidence ranking, while\n *  RRF weighs a rank-30 lexical candidate like a rank-30 embedding hit. */\nconst LEXICAL_FUSION_CAP = 20;\n/**\n * Embedding hits fetched per query when fusing with the grep leg.\n *\n * Deliberately shallow. The grep list is capped at {@link LEXICAL_FUSION_CAP}\n * because its tail is unranked noise, so a deep dense pool here simply\n * outnumbers it: raising this to 200 alongside a 20-candidate grep list cost\n * `auto +rr` 0.472 -> 0.450 MRR.\n */\nconst EMBED_TOP_K = 50;\n/**\n * Per-leg depth when both retrievers are ranked ones (dense + BM25).\n *\n * Four times {@link FUSED_WINDOW}, matching what the daemon's `query_hybrid`\n * does internally (`pool = 4·k`). Fusing at the same depth as the window loses\n * any candidate ranked well by one retriever but just outside the other's\n * top-50 — worth 12pp R@10 and 8pp R@50 for `bm25+dense`, which is exactly\n * what closed the gap to daemon-side fusion.\n */\nconst FUSION_POOL_TOP_K = 200;\n/** BM25 depth, matching {@link FUSION_POOL_TOP_K} so neither ranked leg is\n *  structurally advantaged by pool size. The daemon returns only documents\n *  sharing a query term, so this is an upper bound, not a fill. */\nconst BM25_TOP_K = FUSION_POOL_TOP_K;\n/** Candidates from unindexed/stale files moved ahead of the fused window.\n *  Small on purpose: this answers \"what did I just write\", not \"search the\n *  working tree\". */\nconst STALE_HOIST_CAP = 5;\n/** Fused candidates kept for reranking / final slicing. */\nconst FUSED_WINDOW = 50;\n\nexport interface RetrieveOptions {\n\tcwd: string;\n\tquery: string;\n\tmode?: SearchMode;\n\t/** Optional glob filter applied to file paths. */\n\tglob?: string;\n\t/** Maximum fused candidates returned. */\n\tlimit?: number;\n\t/** RRF constant override (eval harness sweeps this). Default: {@link DEFAULT_RRF_K}. */\n\trrfK?: number;\n\t/** Rerank the fused top-50 before slicing to `limit`. Default: true. */\n\trerank?: boolean;\n\t/**\n\t * Ask the daemon to fuse its own BM25 index with the vectors and return one\n\t * already-fused ranking, instead of taking a dense-only list. Needs a store\n\t * built with `--hybrid`. The fused list arrives as a single \"embed\" leg,\n\t * because a pre-fused ranking has no per-retriever structure left to record.\n\t *\n\t * Prefer {@link bm25Leg}: fusing here keeps the legs separable in the trace\n\t * and lets the grep leg participate.\n\t */\n\tdaemonHybrid?: boolean;\n\t/**\n\t * Fetch the daemon's BM25 index as its own ranked list and fuse it here,\n\t * alongside dense and grep.\n\t *\n\t * Defaults to on wherever the daemon can serve it, because BM25 is the\n\t * better lexical leg on the indexed corpus: Recall@50 0.790 -> 0.879, 6 of\n\t * 62 queries better and 0 worse (p <= 0.05). When it is on, the grep leg\n\t * narrows to files the index has not read — see `staleFiles`.\n\t *\n\t * Set `false` to force ripgrep as the only lexical leg; the eval harness\n\t * does this to keep measuring what the old rows measured.\n\t */\n\tbm25Leg?: boolean;\n\t/**\n\t * Reorder the fused shortlist with the daemon's cross-encoder instead of\n\t * the deterministic reranker. Needs embsearch >= 0.3.0; costs one model\n\t * pass per scored candidate.\n\t */\n\tcrossEncoder?: boolean;\n\tservice?: EmbsearchService;\n\tsignal?: AbortSignal;\n}\n\nexport interface RetrieveResult {\n\tcandidates: FusedCandidate[];\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tindexPhase: SearchTrace[\"indexPhase\"];\n\tretrievers: SearchTrace[\"retrievers\"];\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n\trrfK: number;\n\trerank?: SearchTrace[\"rerank\"];\n}\n\nexport interface RunSearchOptions extends RetrieveOptions {\n\t/** Approximate token budget for the result text. */\n\ttokenBudget?: number;\n}\n\nexport interface RunSearchResult {\n\ttext: string;\n\tresolvedMode: ResolvedSearchMode;\n\tdegradedReason?: string;\n\tresultCount: number;\n\t/** Set while the embedding index is still building. */\n\tindexing?: { done: number; total: number };\n}\n\nfunction normalizeSearchGlob(glob: string | undefined): string | undefined {\n\tif (!glob) return undefined;\n\t// Match fd/rg semantics: a slash-containing glob is anchored anywhere in\n\t// the tree, so prepend \"**/\" unless it already starts with a slash or \"**/\".\n\tif (glob.includes(\"/\") && !glob.startsWith(\"/\") && !glob.startsWith(\"**/\")) {\n\t\treturn `**/${glob}`;\n\t}\n\treturn glob;\n}\n\n/**\n * Move candidates from files the index has not read to the front.\n *\n * They are there because grep found them and nothing else could: the index is\n * ranking a stale copy of the file, or has never seen it. Left to fuse, they\n * lose — RRF rewards agreement, and one leg reporting a single document is\n * outvoted by two legs agreeing on hundreds. Measured: scoping grep to stale\n * files without this hoist scored 25% on the live-edit set where unscoped grep\n * scored 100%, because the fused window filled with consensus hits about the\n * indexed copy.\n *\n * Capped, because \"stale\" scales with how far behind the index is. A few\n * edited files is the case this exists for; a fresh checkout makes everything\n * stale, and hoisting all of it would quietly turn hybrid search back into\n * grep. Past the cap the remainder keeps its fused position.\n */\nexport function hoistStaleCandidates(\n\tcandidates: readonly FusedCandidate[],\n\tstaleFiles: ReadonlySet<string>,\n): FusedCandidate[] {\n\tif (staleFiles.size === 0) return [...candidates];\n\tconst hoisted: FusedCandidate[] = [];\n\tconst rest: FusedCandidate[] = [];\n\tfor (const candidate of candidates) {\n\t\t(staleFiles.has(candidate.path) && hoisted.length < STALE_HOIST_CAP ? hoisted : rest).push(candidate);\n\t}\n\treturn [...hoisted, ...rest];\n}\n\n/**\n * Fold overlapping or adjacent spans of the same file into their best-ranked\n * occurrence.\n *\n * Chunks overlap by design (see the chunker's `CHUNK_OVERLAP_LINES`) and each\n * one is a separate id, so neighbouring chunks of one region survive fusion as\n * separate candidates and take separate result slots — showing the model code\n * it already has. Measured on the 62-query set at the tool's default\n * `limit=5`, 43 queries had such a pair in their top 5 and 58 of 310 slots\n * went to repeated code.\n *\n * Merging is free recall: the union of two overlapping spans matches exactly\n * what either matched, so nothing is gained by widening — the gain is entirely\n * the slot handed back to the ranked tail. Recall@5 0.597 -> 0.677, Recall@1\n * unchanged (merging cannot alter the top result).\n *\n * Adjacency counts as overlap (`endLine + 1`): two chunks that abut describe\n * one continuous region, and rendering them as separate results implies a gap\n * that is not there.\n */\nexport function mergeOverlappingSpans(candidates: readonly FusedCandidate[]): FusedCandidate[] {\n\tconst merged: FusedCandidate[] = [];\n\tfor (const candidate of candidates) {\n\t\tconst into = merged.find(\n\t\t\t(m) =>\n\t\t\t\tm.path === candidate.path && candidate.startLine <= m.endLine + 1 && m.startLine <= candidate.endLine + 1,\n\t\t);\n\t\tif (!into) {\n\t\t\tmerged.push({ ...candidate, ranks: { ...candidate.ranks }, rawScores: { ...candidate.rawScores } });\n\t\t\tcontinue;\n\t\t}\n\t\tinto.startLine = Math.min(into.startLine, candidate.startLine);\n\t\tinto.endLine = Math.max(into.endLine, candidate.endLine);\n\t\t// Keep the folded id: the trace is the only record that retrieval found\n\t\t// it, and dropping it would make a merged result look like a miss.\n\t\tif (!into.mergedFrom) into.mergedFrom = [];\n\t\tinto.mergedFrom.push(candidate.id);\n\t\t// The merged region really was reached by every retriever that found any\n\t\t// part of it, so the source label and the trace should say so. Best rank\n\t\t// per source wins; the fused score stays the survivor's, which is what\n\t\t// ordered it here.\n\t\tfor (const [source, rank] of Object.entries(candidate.ranks)) {\n\t\t\tconst held = into.ranks[source as keyof typeof into.ranks];\n\t\t\tif (held === undefined || rank < held) into.ranks[source as keyof typeof into.ranks] = rank;\n\t\t}\n\t\tfor (const [source, score] of Object.entries(candidate.rawScores)) {\n\t\t\tconst held = into.rawScores[source as keyof typeof into.rawScores];\n\t\t\tif (held === undefined || score > held) into.rawScores[source as keyof typeof into.rawScores] = score;\n\t\t}\n\t}\n\treturn merged;\n}\n\nexport async function retrieveCandidates(options: RetrieveOptions): Promise<RetrieveResult> {\n\tconst { cwd, query, service, signal } = options;\n\tconst glob = normalizeSearchGlob(options.glob);\n\tconst requestedMode = options.mode ?? \"auto\";\n\tconst limit = Math.max(1, options.limit ?? 10);\n\tconst rrfK = options.rrfK ?? DEFAULT_RRF_K;\n\n\tconst state = service?.getState();\n\tconst embedAvailable = service?.isAvailable() ?? false;\n\tconst embedUnavailableReason =\n\t\tstate === undefined\n\t\t\t? \"semantic index is not enabled\"\n\t\t\t: state.phase === \"unavailable\" || state.phase === \"skipped\"\n\t\t\t\t? state.reason\n\t\t\t\t: state.phase === \"idle\"\n\t\t\t\t\t? \"semantic index has not started\"\n\t\t\t\t\t: undefined;\n\n\tconst resolution = resolveSearchMode(query, requestedMode, embedAvailable, embedUnavailableReason);\n\tconst mode = resolution.mode;\n\n\t// Map lexical hits onto indexed chunk ids whenever the sidecar is usable,\n\t// even in lexical-only mode, so identities line up across modes.\n\tconst lookupChunk: ChunkLookup | undefined = embedAvailable\n\t\t? (rel, line) => service!.findEnclosingChunk(rel, line)\n\t\t: undefined;\n\n\tconst spans = new Map<string, CandidateSpan>();\n\t/** Files the index does not have current content for, when grep is scoped\n\t *  to them. Read again after fusion — see the hoist below. */\n\tconst staleSet = new Set<string>();\n\tconst lists: RankedHit[][] = [];\n\tconst retrieverStats: SearchTrace[\"retrievers\"] = {};\n\tconst errors: Error[] = [];\n\n\tconst runLexical = async (scopeToStale = false): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Scoped run: grep covers only what the index has not read yet, so it\n\t\t\t// adds the one thing BM25 cannot see without re-voting on documents\n\t\t\t// BM25 already ranked. Fusing two lexical views of the same corpus is\n\t\t\t// what made the three-leg configuration lose (18 of 62 queries worse).\n\t\t\tconst paths = scopeToStale ? service?.staleFiles(signal) : undefined;\n\t\t\tif (paths) for (const rel of paths) staleSet.add(rel);\n\t\t\tconst lineHits = await runLexicalRetriever({ cwd, query, limit: LEXICAL_MATCH_LIMIT, glob, signal, paths });\n\t\t\tconst adapted = adaptGrepHits(lineHits, lookupChunk);\n\t\t\t// In single-retriever lexical mode the full list is the result; in\n\t\t\t// hybrid, only the front-loaded head is trustworthy enough to vote.\n\t\t\tconst hits = mode === \"hybrid\" ? adapted.hits.slice(0, LEXICAL_FUSION_CAP) : adapted.hits;\n\t\t\tfor (const [id, span] of adapted.spans) if (!spans.has(id)) spans.set(id, span);\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.grep = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runEmbed = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// The flat index pads top-k with whatever exists; with the cosine\n\t\t\t// metric the store uses, score <= 0 means \"no relation at all\", so\n\t\t\t// those padding hits would cast RRF votes on pure noise.\n\t\t\t// A BM25 leg is itself ranked, so the pair can afford — and needs — the\n\t\t\t// deeper pool; the grep leg cannot (see EMBED_TOP_K).\n\t\t\tconst topK = options.bm25Leg ? FUSION_POOL_TOP_K : EMBED_TOP_K;\n\t\t\tconst chunkHits = (\n\t\t\t\tawait service!.searchChunks(query, topK, glob, options.daemonHybrid ? \"hybrid\" : \"dense\")\n\t\t\t).filter((hit) => hit.score > 0);\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"embed\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.embed = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\tconst runBm25 = async (): Promise<void> => {\n\t\tconst startedMs = Date.now();\n\t\ttry {\n\t\t\t// Raw BM25 sums; only the ordering enters fusion, the score is a\n\t\t\t// diagnostic. Unlike the dense leg there is no zero-score padding to\n\t\t\t// filter — the daemon omits documents sharing no query term.\n\t\t\tconst chunkHits = await service!.searchChunks(query, BM25_TOP_K, glob, \"lexical\");\n\t\t\tconst hits: RankedHit[] = chunkHits.map((hit, i) => ({\n\t\t\t\tid: hit.id,\n\t\t\t\trank: i + 1,\n\t\t\t\tscore: hit.score,\n\t\t\t\tsource: \"bm25\",\n\t\t\t}));\n\t\t\tfor (const hit of chunkHits) {\n\t\t\t\tif (!spans.has(hit.id)) {\n\t\t\t\t\tspans.set(hit.id, { path: hit.path, startLine: hit.startLine, endLine: hit.endLine });\n\t\t\t\t}\n\t\t\t}\n\t\t\tlists.push(hits);\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: hits.length };\n\t\t} catch (e) {\n\t\t\terrors.push(e instanceof Error ? e : new Error(String(e)));\n\t\t\tretrieverStats.bm25 = { latencyMs: Date.now() - startedMs, hitCount: 0 };\n\t\t}\n\t};\n\n\t// BM25 is the better lexical leg where the index is current: it beats grep\n\t// on Recall@50 (0.790 -> 0.879, 6 queries better and 0 worse, p <= 0.05)\n\t// because it ranks the whole corpus rather than truncating a match stream.\n\t// It is also blind to anything indexed later than it was written, which is\n\t// precisely where grep still wins — so grep runs scoped to that set instead\n\t// of being dropped or left to duplicate BM25 over the whole tree.\n\tconst bm25AsLexicalLeg =\n\t\t(options.bm25Leg ?? service?.supportsLexicalRetriever() ?? false) && embedAvailable && mode !== \"lexical\";\n\n\tconst runs: Promise<void>[] = [];\n\tif (mode === \"lexical\" || mode === \"hybrid\") runs.push(runLexical(bm25AsLexicalLeg));\n\tif (mode === \"semantic\" || mode === \"hybrid\") runs.push(runEmbed());\n\t// `semantic` gets the BM25 leg too: the caller asked for the index, and\n\t// BM25 is part of it. Only an explicit `lexical` request excludes it.\n\tif (bm25AsLexicalLeg) runs.push(runBm25());\n\tawait Promise.all(runs);\n\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\t// A partial failure in hybrid degrades to whichever retriever survived;\n\t// only a total loss is an error.\n\tif (lists.length === 0) throw errors[0] ?? new Error(\"search produced no retriever results\");\n\n\tconst fused = rrfFuse(lists, rrfK).slice(0, FUSED_WINDOW);\n\tlet candidates: FusedCandidate[] = [];\n\tfor (const hit of fused) {\n\t\tconst span = spans.get(hit.id);\n\t\tif (span) candidates.push({ ...hit, ...span });\n\t}\n\n\tlet rerankInfo: SearchTrace[\"rerank\"];\n\tif (options.crossEncoder && service?.supportsCrossEncoder()) {\n\t\tconst reranked = await crossEncoderRerank(query, candidates, cwd, service);\n\t\trerankInfo = { applied: true, candidateCount: reranked.scored, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t} else if (options.rerank !== false) {\n\t\tconst reranked = rerankCandidates(query, candidates, cwd);\n\t\trerankInfo = { applied: true, candidateCount: candidates.length, latencyMs: reranked.latencyMs };\n\t\tcandidates = reranked.candidates;\n\t}\n\t// After reranking, not before: both rerankers weight the fused prior, and a\n\t// lone grep hit from an unindexed file has the lowest prior there is, so\n\t// hoisting first would simply be undone. The reranker still orders the\n\t// hoisted set against itself.\n\t// Merge before slicing, so a slot freed by a duplicate is refilled from the\n\t// ranked tail rather than left empty.\n\tcandidates = mergeOverlappingSpans(hoistStaleCandidates(candidates, staleSet)).slice(0, limit);\n\n\treturn {\n\t\tcandidates,\n\t\tresolvedMode: mode,\n\t\tdegradedReason: resolution.degradedReason,\n\t\tindexPhase: state?.phase === \"ready\" ? \"ready\" : state?.phase === \"indexing\" ? \"indexing\" : \"unavailable\",\n\t\tretrievers: retrieverStats,\n\t\tindexing: state?.phase === \"indexing\" ? { done: state.done, total: state.total } : undefined,\n\t\trrfK,\n\t\trerank: rerankInfo,\n\t};\n}\n\nexport async function runSearch(options: RunSearchOptions): Promise<RunSearchResult> {\n\tconst retrieved = await retrieveCandidates(options);\n\n\tconst assembled = assembleContext(retrieved.candidates, { cwd: options.cwd, tokenBudget: options.tokenBudget });\n\n\twriteSearchTrace(options.cwd, {\n\t\ttimestampMs: Date.now(),\n\t\tquery: options.query,\n\t\trequestedMode: options.mode ?? \"auto\",\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tindexPhase: retrieved.indexPhase,\n\t\trrfK: retrieved.resolvedMode === \"hybrid\" ? retrieved.rrfK : undefined,\n\t\tretrievers: retrieved.retrievers,\n\t\tfused: retrieved.candidates.map(({ id, rrfScore, ranks, rawScores, mergedFrom }) => ({\n\t\t\tid,\n\t\t\trrfScore,\n\t\t\tranks,\n\t\t\trawScores,\n\t\t\t...(mergedFrom ? { mergedFrom } : {}),\n\t\t})),\n\t\trerank: retrieved.rerank,\n\t});\n\n\treturn {\n\t\ttext: assembled.text,\n\t\tresolvedMode: retrieved.resolvedMode,\n\t\tdegradedReason: retrieved.degradedReason,\n\t\tresultCount: retrieved.candidates.length,\n\t\tindexing: retrieved.indexing,\n\t};\n}\n"]}

package/dist/core/search/types.d.ts

@@ -35,6 +35,15 @@ export interface FusedHit {
     ranks: Partial<Record<RetrieverSource, number>>;
     /** Raw score at the best rank per retriever. Diagnostics only. */
     rawScores: Partial<Record<RetrieverSource, number>>;
+    /**
+     * Ids folded into this one by span merging, if any.
+     *
+     * The model sees one result per continuous region, but a trace that only
+     * listed survivors would hide that a chunk was retrieved at all — which is
+     * exactly what you need when working out why something ranked where it did.
+     * Diagnostics only.
+     */
+    mergedFrom?: string[];
 }
 /** Line-span identity a candidate id resolves to, for post-fusion expansion. */
 export interface CandidateSpan {

package/dist/core/search/types.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../../src/core/search/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH;;;;;;;;;GASG;AACH,KAAK,eAAe,GAAG,MAAM,GAAG,OAAO,GAAG,MAAM,CAAC;AAEjD,MAAM,MAAM,UAAU,GAAG,MAAM,GAAG,SAAS,GAAG,UAAU,GAAG,QAAQ,CAAC;AACpE,MAAM,MAAM,kBAAkB,GAAG,OAAO,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;AAE7D,MAAM,WAAW,SAAS;IACzB,EAAE,EAAE,MAAM,CAAC;IACX,4DAA4D;IAC5D,IAAI,EAAE,MAAM,CAAC;IACb,uFAAmF;IACnF,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,eAAe,CAAC;CACxB;AAED,MAAM,WAAW,QAAQ;IACxB,EAAE,EAAE,MAAM,CAAC;IACX,QAAQ,EAAE,MAAM,CAAC;IACjB,4CAA4C;IAC5C,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,eAAe,EAAE,MAAM,CAAC,CAAC,CAAC;IAChD,kEAAkE;IAClE,SAAS,EAAE,OAAO,CAAC,MAAM,CAAC,eAAe,EAAE,MAAM,CAAC,CAAC,CAAC;CACpD;AAED,gFAAgF;AAChF,MAAM,WAAW,aAAa;IAC7B,IAAI,EAAE,MAAM,CAAC;IACb,yBAAyB;IACzB,SAAS,EAAE,MAAM,CAAC;IAClB,sEAAsE;IACtE,OAAO,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,cAAe,SAAQ,QAAQ,EAAE,aAAa;CAAG;AAElE;0BAC0B;AAC1B,MAAM,WAAW,WAAW;IAC3B,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,aAAa,EAAE,UAAU,CAAC;IAC1B,YAAY,EAAE,kBAAkB,CAAC;IACjC,wEAAwE;IACxE,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,OAAO,GAAG,UAAU,GAAG,aAAa,CAAC;IACjD,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,eAAe,EAAE;QAAE,SAAS,EAAE,MAAM,CAAC;QAAC,QAAQ,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC,CAAC;IACtF,KAAK,EAAE,QAAQ,EAAE,CAAC;IAClB,MAAM,CAAC,EAAE;QAAE,OAAO,EAAE,OAAO,CAAC;QAAC,cAAc,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAA;KAAE,CAAC;CACzE","sourcesContent":["/**\n * Shared types for hybrid code retrieval (see docs/hybrid-retrieval-design.md).\n *\n * Identity note: candidate ids are per-index-build chunk ids (`relpath#index`)\n * from the embedding sidecar, or synthetic `relpath#L<line>` fallback ids for\n * grep hits in files the index does not cover. They are stable within one\n * query (fusion and span expansion consult the same sidecar snapshot) but not\n * across edits or rebuilds — never persist them as durable references.\n */\n\n/**\n * A retriever whose ranked list can enter fusion.\n *\n * `bm25` is the daemon's Okapi lexical index, fetched as its own leg via the\n * `retriever: \"lexical\"` op rather than pre-fused by `query_hybrid`. Keeping\n * it separate from `grep` matters: they are both \"lexical\" but they fail\n * differently — BM25 has IDF and misses identifiers written in another naming\n * convention, ripgrep has neither IDF nor an index but sees the working tree,\n * including edits made this session.\n */\ntype RetrieverSource = \"grep\" | \"embed\" | \"bm25\";\n\nexport type SearchMode = \"auto\" | \"lexical\" | \"semantic\" | \"hybrid\";\nexport type ResolvedSearchMode = Exclude<SearchMode, \"auto\">;\n\nexport interface RankedHit {\n\tid: string;\n\t/** 1-indexed, gap-free rank within its retriever's list. */\n\trank: number;\n\t/** Retriever-local score (BM25-ish, cosine, …). Diagnostics only — never fused. */\n\tscore?: number;\n\tsource: RetrieverSource;\n}\n\nexport interface FusedHit {\n\tid: string;\n\trrfScore: number;\n\t/** Best rank per contributing retriever. */\n\tranks: Partial<Record<RetrieverSource, number>>;\n\t/** Raw score at the best rank per retriever. Diagnostics only. */\n\trawScores: Partial<Record<RetrieverSource, number>>;\n}\n\n/** Line-span identity a candidate id resolves to, for post-fusion expansion. */\nexport interface CandidateSpan {\n\tpath: string;\n\t/** 1-based inclusive. */\n\tstartLine: number;\n\t/** 1-based inclusive. May exceed the file's length; readers clamp. */\n\tendLine: number;\n}\n\nexport interface FusedCandidate extends FusedHit, CandidateSpan {}\n\n/** Per-call diagnostic record, written to the store-dir trace jsonl — never\n *  into model context. */\nexport interface SearchTrace {\n\ttimestampMs: number;\n\tquery: string;\n\trequestedMode: SearchMode;\n\tresolvedMode: ResolvedSearchMode;\n\t/** Set when the resolved mode is a degradation of the requested one. */\n\tdegradedReason?: string;\n\tindexPhase: \"ready\" | \"indexing\" | \"unavailable\";\n\trrfK?: number;\n\tretrievers: Partial<Record<RetrieverSource, { latencyMs: number; hitCount: number }>>;\n\tfused: FusedHit[];\n\trerank?: { applied: boolean; candidateCount: number; latencyMs: number };\n}\n"]}
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../../src/core/search/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH;;;;;;;;;GASG;AACH,KAAK,eAAe,GAAG,MAAM,GAAG,OAAO,GAAG,MAAM,CAAC;AAEjD,MAAM,MAAM,UAAU,GAAG,MAAM,GAAG,SAAS,GAAG,UAAU,GAAG,QAAQ,CAAC;AACpE,MAAM,MAAM,kBAAkB,GAAG,OAAO,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;AAE7D,MAAM,WAAW,SAAS;IACzB,EAAE,EAAE,MAAM,CAAC;IACX,4DAA4D;IAC5D,IAAI,EAAE,MAAM,CAAC;IACb,uFAAmF;IACnF,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,MAAM,EAAE,eAAe,CAAC;CACxB;AAED,MAAM,WAAW,QAAQ;IACxB,EAAE,EAAE,MAAM,CAAC;IACX,QAAQ,EAAE,MAAM,CAAC;IACjB,4CAA4C;IAC5C,KAAK,EAAE,OAAO,CAAC,MAAM,CAAC,eAAe,EAAE,MAAM,CAAC,CAAC,CAAC;IAChD,kEAAkE;IAClE,SAAS,EAAE,OAAO,CAAC,MAAM,CAAC,eAAe,EAAE,MAAM,CAAC,CAAC,CAAC;IACpD;;;;;;;OAOG;IACH,UAAU,CAAC,EAAE,MAAM,EAAE,CAAC;CACtB;AAED,gFAAgF;AAChF,MAAM,WAAW,aAAa;IAC7B,IAAI,EAAE,MAAM,CAAC;IACb,yBAAyB;IACzB,SAAS,EAAE,MAAM,CAAC;IAClB,sEAAsE;IACtE,OAAO,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,cAAe,SAAQ,QAAQ,EAAE,aAAa;CAAG;AAElE;0BAC0B;AAC1B,MAAM,WAAW,WAAW;IAC3B,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,aAAa,EAAE,UAAU,CAAC;IAC1B,YAAY,EAAE,kBAAkB,CAAC;IACjC,wEAAwE;IACxE,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,UAAU,EAAE,OAAO,GAAG,UAAU,GAAG,aAAa,CAAC;IACjD,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,OAAO,CAAC,MAAM,CAAC,eAAe,EAAE;QAAE,SAAS,EAAE,MAAM,CAAC;QAAC,QAAQ,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC,CAAC;IACtF,KAAK,EAAE,QAAQ,EAAE,CAAC;IAClB,MAAM,CAAC,EAAE;QAAE,OAAO,EAAE,OAAO,CAAC;QAAC,cAAc,EAAE,MAAM,CAAC;QAAC,SAAS,EAAE,MAAM,CAAA;KAAE,CAAC;CACzE","sourcesContent":["/**\n * Shared types for hybrid code retrieval (see docs/hybrid-retrieval-design.md).\n *\n * Identity note: candidate ids are per-index-build chunk ids (`relpath#index`)\n * from the embedding sidecar, or synthetic `relpath#L<line>` fallback ids for\n * grep hits in files the index does not cover. They are stable within one\n * query (fusion and span expansion consult the same sidecar snapshot) but not\n * across edits or rebuilds — never persist them as durable references.\n */\n\n/**\n * A retriever whose ranked list can enter fusion.\n *\n * `bm25` is the daemon's Okapi lexical index, fetched as its own leg via the\n * `retriever: \"lexical\"` op rather than pre-fused by `query_hybrid`. Keeping\n * it separate from `grep` matters: they are both \"lexical\" but they fail\n * differently — BM25 has IDF and misses identifiers written in another naming\n * convention, ripgrep has neither IDF nor an index but sees the working tree,\n * including edits made this session.\n */\ntype RetrieverSource = \"grep\" | \"embed\" | \"bm25\";\n\nexport type SearchMode = \"auto\" | \"lexical\" | \"semantic\" | \"hybrid\";\nexport type ResolvedSearchMode = Exclude<SearchMode, \"auto\">;\n\nexport interface RankedHit {\n\tid: string;\n\t/** 1-indexed, gap-free rank within its retriever's list. */\n\trank: number;\n\t/** Retriever-local score (BM25-ish, cosine, …). Diagnostics only — never fused. */\n\tscore?: number;\n\tsource: RetrieverSource;\n}\n\nexport interface FusedHit {\n\tid: string;\n\trrfScore: number;\n\t/** Best rank per contributing retriever. */\n\tranks: Partial<Record<RetrieverSource, number>>;\n\t/** Raw score at the best rank per retriever. Diagnostics only. */\n\trawScores: Partial<Record<RetrieverSource, number>>;\n\t/**\n\t * Ids folded into this one by span merging, if any.\n\t *\n\t * The model sees one result per continuous region, but a trace that only\n\t * listed survivors would hide that a chunk was retrieved at all — which is\n\t * exactly what you need when working out why something ranked where it did.\n\t * Diagnostics only.\n\t */\n\tmergedFrom?: string[];\n}\n\n/** Line-span identity a candidate id resolves to, for post-fusion expansion. */\nexport interface CandidateSpan {\n\tpath: string;\n\t/** 1-based inclusive. */\n\tstartLine: number;\n\t/** 1-based inclusive. May exceed the file's length; readers clamp. */\n\tendLine: number;\n}\n\nexport interface FusedCandidate extends FusedHit, CandidateSpan {}\n\n/** Per-call diagnostic record, written to the store-dir trace jsonl — never\n *  into model context. */\nexport interface SearchTrace {\n\ttimestampMs: number;\n\tquery: string;\n\trequestedMode: SearchMode;\n\tresolvedMode: ResolvedSearchMode;\n\t/** Set when the resolved mode is a degradation of the requested one. */\n\tdegradedReason?: string;\n\tindexPhase: \"ready\" | \"indexing\" | \"unavailable\";\n\trrfK?: number;\n\tretrievers: Partial<Record<RetrieverSource, { latencyMs: number; hitCount: number }>>;\n\tfused: FusedHit[];\n\trerank?: { applied: boolean; candidateCount: number; latencyMs: number };\n}\n"]}

package/dist/core/search/types.js.map

@@ -1 +1 @@
-{"version":3,"file":"types.js","sourceRoot":"","sources":["../../../src/core/search/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG","sourcesContent":["/**\n * Shared types for hybrid code retrieval (see docs/hybrid-retrieval-design.md).\n *\n * Identity note: candidate ids are per-index-build chunk ids (`relpath#index`)\n * from the embedding sidecar, or synthetic `relpath#L<line>` fallback ids for\n * grep hits in files the index does not cover. They are stable within one\n * query (fusion and span expansion consult the same sidecar snapshot) but not\n * across edits or rebuilds — never persist them as durable references.\n */\n\n/**\n * A retriever whose ranked list can enter fusion.\n *\n * `bm25` is the daemon's Okapi lexical index, fetched as its own leg via the\n * `retriever: \"lexical\"` op rather than pre-fused by `query_hybrid`. Keeping\n * it separate from `grep` matters: they are both \"lexical\" but they fail\n * differently — BM25 has IDF and misses identifiers written in another naming\n * convention, ripgrep has neither IDF nor an index but sees the working tree,\n * including edits made this session.\n */\ntype RetrieverSource = \"grep\" | \"embed\" | \"bm25\";\n\nexport type SearchMode = \"auto\" | \"lexical\" | \"semantic\" | \"hybrid\";\nexport type ResolvedSearchMode = Exclude<SearchMode, \"auto\">;\n\nexport interface RankedHit {\n\tid: string;\n\t/** 1-indexed, gap-free rank within its retriever's list. */\n\trank: number;\n\t/** Retriever-local score (BM25-ish, cosine, …). Diagnostics only — never fused. */\n\tscore?: number;\n\tsource: RetrieverSource;\n}\n\nexport interface FusedHit {\n\tid: string;\n\trrfScore: number;\n\t/** Best rank per contributing retriever. */\n\tranks: Partial<Record<RetrieverSource, number>>;\n\t/** Raw score at the best rank per retriever. Diagnostics only. */\n\trawScores: Partial<Record<RetrieverSource, number>>;\n}\n\n/** Line-span identity a candidate id resolves to, for post-fusion expansion. */\nexport interface CandidateSpan {\n\tpath: string;\n\t/** 1-based inclusive. */\n\tstartLine: number;\n\t/** 1-based inclusive. May exceed the file's length; readers clamp. */\n\tendLine: number;\n}\n\nexport interface FusedCandidate extends FusedHit, CandidateSpan {}\n\n/** Per-call diagnostic record, written to the store-dir trace jsonl — never\n *  into model context. */\nexport interface SearchTrace {\n\ttimestampMs: number;\n\tquery: string;\n\trequestedMode: SearchMode;\n\tresolvedMode: ResolvedSearchMode;\n\t/** Set when the resolved mode is a degradation of the requested one. */\n\tdegradedReason?: string;\n\tindexPhase: \"ready\" | \"indexing\" | \"unavailable\";\n\trrfK?: number;\n\tretrievers: Partial<Record<RetrieverSource, { latencyMs: number; hitCount: number }>>;\n\tfused: FusedHit[];\n\trerank?: { applied: boolean; candidateCount: number; latencyMs: number };\n}\n"]}
+{"version":3,"file":"types.js","sourceRoot":"","sources":["../../../src/core/search/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG","sourcesContent":["/**\n * Shared types for hybrid code retrieval (see docs/hybrid-retrieval-design.md).\n *\n * Identity note: candidate ids are per-index-build chunk ids (`relpath#index`)\n * from the embedding sidecar, or synthetic `relpath#L<line>` fallback ids for\n * grep hits in files the index does not cover. They are stable within one\n * query (fusion and span expansion consult the same sidecar snapshot) but not\n * across edits or rebuilds — never persist them as durable references.\n */\n\n/**\n * A retriever whose ranked list can enter fusion.\n *\n * `bm25` is the daemon's Okapi lexical index, fetched as its own leg via the\n * `retriever: \"lexical\"` op rather than pre-fused by `query_hybrid`. Keeping\n * it separate from `grep` matters: they are both \"lexical\" but they fail\n * differently — BM25 has IDF and misses identifiers written in another naming\n * convention, ripgrep has neither IDF nor an index but sees the working tree,\n * including edits made this session.\n */\ntype RetrieverSource = \"grep\" | \"embed\" | \"bm25\";\n\nexport type SearchMode = \"auto\" | \"lexical\" | \"semantic\" | \"hybrid\";\nexport type ResolvedSearchMode = Exclude<SearchMode, \"auto\">;\n\nexport interface RankedHit {\n\tid: string;\n\t/** 1-indexed, gap-free rank within its retriever's list. */\n\trank: number;\n\t/** Retriever-local score (BM25-ish, cosine, …). Diagnostics only — never fused. */\n\tscore?: number;\n\tsource: RetrieverSource;\n}\n\nexport interface FusedHit {\n\tid: string;\n\trrfScore: number;\n\t/** Best rank per contributing retriever. */\n\tranks: Partial<Record<RetrieverSource, number>>;\n\t/** Raw score at the best rank per retriever. Diagnostics only. */\n\trawScores: Partial<Record<RetrieverSource, number>>;\n\t/**\n\t * Ids folded into this one by span merging, if any.\n\t *\n\t * The model sees one result per continuous region, but a trace that only\n\t * listed survivors would hide that a chunk was retrieved at all — which is\n\t * exactly what you need when working out why something ranked where it did.\n\t * Diagnostics only.\n\t */\n\tmergedFrom?: string[];\n}\n\n/** Line-span identity a candidate id resolves to, for post-fusion expansion. */\nexport interface CandidateSpan {\n\tpath: string;\n\t/** 1-based inclusive. */\n\tstartLine: number;\n\t/** 1-based inclusive. May exceed the file's length; readers clamp. */\n\tendLine: number;\n}\n\nexport interface FusedCandidate extends FusedHit, CandidateSpan {}\n\n/** Per-call diagnostic record, written to the store-dir trace jsonl — never\n *  into model context. */\nexport interface SearchTrace {\n\ttimestampMs: number;\n\tquery: string;\n\trequestedMode: SearchMode;\n\tresolvedMode: ResolvedSearchMode;\n\t/** Set when the resolved mode is a degradation of the requested one. */\n\tdegradedReason?: string;\n\tindexPhase: \"ready\" | \"indexing\" | \"unavailable\";\n\trrfK?: number;\n\tretrievers: Partial<Record<RetrieverSource, { latencyMs: number; hitCount: number }>>;\n\tfused: FusedHit[];\n\trerank?: { applied: boolean; candidateCount: number; latencyMs: number };\n}\n"]}

package/dist/core/tools/search.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"search.d.ts","sourceRoot":"","sources":["../../../src/core/tools/search.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAGH,OAAO,EAAe,IAAI,EAAE,MAAM,SAAS,CAAC;AAE5C,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,mCAAmC,CAAC;AAC1E,OAAO,KAAK,EAAE,cAAc,EAA2B,MAAM,wBAAwB,CAAC;AAEtF,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,oBAAoB,CAAC;AAQ7D,QAAA,MAAM,YAAY;;;;;EAoBhB,CAAC;AAIH,MAAM,WAAW,iBAAiB;IACjC,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,YAAY,CAAC,EAAE,kBAAkB,CAAC;IAClC,QAAQ,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;CAC3C;AAED,MAAM,WAAW,iBAAiB;IACjC;;;;OAIG;IACH,UAAU,CAAC,EAAE,MAAM,gBAAgB,GAAG,SAAS,CAAC;CAChD;AAqCD,wBAAgB,0BAA0B,CACzC,GAAG,EAAE,MAAM,EACX,OAAO,CAAC,EAAE,iBAAiB,GACzB,cAAc,CAAC,OAAO,YAAY,EAAE,iBAAiB,GAAG,SAAS,CAAC,CAoEpE","sourcesContent":["/**\n * The unified `search` tool: ranked code discovery in lexical, semantic, or\n * hybrid (rank-fused) mode. Replaces the old `semantic_search` tool — see\n * docs/hybrid-retrieval-design.md, Decision 1. `grep` stays separate for\n * exact line-level mechanics; this tool answers \"find where X lives\".\n */\n\nimport { Text } from \"@kolisachint/hoocode-tui\";\nimport { type Static, Type } from \"typebox\";\nimport { keyHint } from \"../../modes/interactive/components/keybinding-hints.js\";\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport type { ToolDefinition, ToolRenderResultOptions } from \"../extensions/types.js\";\nimport { runSearch } from \"../search/hybrid-search.js\";\nimport type { ResolvedSearchMode } from \"../search/types.js\";\nimport { getTextOutput, invalidArgText, str } from \"./render-utils.js\";\n\n// Default of 5 balances recall against context cost: eval showed limit=5\n// preserving the top-rank hits of limit=10 at roughly half the tokens.\nconst DEFAULT_RESULTS = 5;\nconst MAX_RESULTS = 30;\n\nconst searchSchema = Type.Object({\n\tquery: Type.String({\n\t\tdescription:\n\t\t\t\"What to find: an identifier, error text, or a natural-language description of the code, e.g. 'where sessions are persisted to disk'\",\n\t}),\n\tmode: Type.Optional(\n\t\tType.Union([Type.Literal(\"auto\"), Type.Literal(\"lexical\"), Type.Literal(\"semantic\"), Type.Literal(\"hybrid\")], {\n\t\t\tdescription:\n\t\t\t\t\"Retrieval mode (default: auto). auto picks hybrid when the semantic index is available; lexical = exact-text only; semantic = embedding index only; hybrid = both, fused by rank.\",\n\t\t}),\n\t),\n\tglob: Type.Optional(\n\t\tType.String({\n\t\t\tdescription:\n\t\t\t\t\"Optional glob filter applied to file paths. Only file paths matching the glob are searched. Supports both slashless patterns (match base name anywhere) and slash patterns (match full path).\",\n\t\t}),\n\t),\n\tlimit: Type.Optional(\n\t\tType.Number({ description: `Maximum number of results (default: ${DEFAULT_RESULTS}, max: ${MAX_RESULTS})` }),\n\t),\n});\n\ntype SearchToolInput = Static<typeof searchSchema>;\n\nexport interface SearchToolDetails {\n\tresultCount?: number;\n\tresolvedMode?: ResolvedSearchMode;\n\tindexing?: { done: number; total: number };\n}\n\nexport interface SearchToolOptions {\n\t/**\n\t * Provider for the per-session embsearch service. Resolved lazily at call\n\t * time because the service is created (and its index built) after tool\n\t * registration. When absent/unavailable the tool degrades to lexical mode.\n\t */\n\tgetService?: () => EmbsearchService | undefined;\n}\n\nfunction formatSearchCall(\n\targs: { query?: string; mode?: string; limit?: number } | undefined,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n): string {\n\tconst query = str(args?.query);\n\tconst queryDisplay = query === null ? invalidArgText(theme) : `\"${query}\"`;\n\tlet text = theme.fg(\"toolTitle\", theme.bold(\"search \")) + theme.fg(\"accent\", queryDisplay);\n\tconst extras: string[] = [];\n\tif (args?.mode && args.mode !== \"auto\") extras.push(args.mode);\n\tif (args?.limit !== undefined) extras.push(String(args.limit));\n\tif (extras.length > 0) text += theme.fg(\"muted\", ` (${extras.join(\", \")})`);\n\treturn text;\n}\n\nfunction formatSearchResult(\n\tresult: { content: Array<{ type: string; text?: string }>; details?: SearchToolDetails },\n\toptions: ToolRenderResultOptions,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n\tshowImages: boolean,\n): string {\n\tconst output = getTextOutput(result as any, showImages).trim();\n\tlet text = \"\";\n\tif (output) {\n\t\tconst lines = output.split(\"\\n\");\n\t\tconst maxLines = options.expanded ? lines.length : 15;\n\t\tconst displayLines = lines.slice(0, maxLines);\n\t\tconst remaining = lines.length - maxLines;\n\t\ttext += `\\n${displayLines.map((line) => theme.fg(\"toolOutput\", line)).join(\"\\n\")}`;\n\t\tif (remaining > 0) {\n\t\t\ttext += `${theme.fg(\"muted\", `\\n... (${remaining} more lines,`)} ${keyHint(\"app.tools.expand\", \"to expand\")})`;\n\t\t}\n\t}\n\treturn text;\n}\n\nexport function createSearchToolDefinition(\n\tcwd: string,\n\toptions?: SearchToolOptions,\n): ToolDefinition<typeof searchSchema, SearchToolDetails | undefined> {\n\treturn {\n\t\tname: \"search\",\n\t\tlabel: \"search\",\n\t\t// The search-vs-grep split is stated once, by buildSystemPrompt, whenever both\n\t\t// tools are registered — so it is deliberately absent here and from grep's\n\t\t// description. What stays is what only this tool knows: that the query is not\n\t\t// a regex, and that it degrades to exact-text when the index is missing.\n\t\tdescription:\n\t\t\t\"Find where code lives: ranked file:line-range results from exact-text and semantic (local embedding index) retrieval, fused by rank when both are available. The query is plain text, not a regex — regex metacharacters are matched literally. Falls back to exact-text retrieval automatically when the semantic index is unavailable.\",\n\t\tpromptSnippet: \"Ranked code search (exact + semantic, rank-fused)\",\n\t\tpromptGuidelines: [\n\t\t\t\"search defaults to mode=auto, which is almost always right. Use limit=3 for targeted lookups, 10–20 when exploring a broad topic.\",\n\t\t],\n\t\tparameters: searchSchema,\n\t\tasync execute(_toolCallId, { query, mode, glob, limit }: SearchToolInput, signal?: AbortSignal) {\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst result = await runSearch({\n\t\t\t\tcwd,\n\t\t\t\tquery,\n\t\t\t\tmode,\n\t\t\t\tglob,\n\t\t\t\tlimit: Math.min(MAX_RESULTS, Math.max(1, limit ?? DEFAULT_RESULTS)),\n\t\t\t\tservice: options?.getService?.(),\n\t\t\t\tsignal,\n\t\t\t});\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst details: SearchToolDetails = {\n\t\t\t\tresultCount: result.resultCount,\n\t\t\t\tresolvedMode: result.resolvedMode,\n\t\t\t\tindexing: result.indexing,\n\t\t\t};\n\n\t\t\tconst notices: string[] = [];\n\t\t\tif (result.degradedReason) notices.push(result.degradedReason);\n\t\t\tif (result.indexing) {\n\t\t\t\tnotices.push(\n\t\t\t\t\t`index still building: ${result.indexing.done}/${result.indexing.total} chunks embedded — results may be incomplete`,\n\t\t\t\t);\n\t\t\t}\n\t\t\tconst notice = notices.length > 0 ? `\\n\\n[${notices.join(\". \")}]` : \"\";\n\n\t\t\tif (result.resultCount === 0) {\n\t\t\t\treturn {\n\t\t\t\t\tcontent: [\n\t\t\t\t\t\t{ type: \"text\" as const, text: `No results for \"${query}\" (${result.resolvedMode})${notice}` },\n\t\t\t\t\t],\n\t\t\t\t\tdetails,\n\t\t\t\t};\n\t\t\t}\n\t\t\treturn {\n\t\t\t\tcontent: [{ type: \"text\" as const, text: result.text + notice }],\n\t\t\t\tdetails,\n\t\t\t};\n\t\t},\n\t\trenderCall(args, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchCall(args, theme));\n\t\t\treturn text;\n\t\t},\n\t\trenderResult(result, options, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchResult(result as any, options, theme, context.showImages));\n\t\t\treturn text;\n\t\t},\n\t};\n}\n"]}
+{"version":3,"file":"search.d.ts","sourceRoot":"","sources":["../../../src/core/tools/search.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAGH,OAAO,EAAe,IAAI,EAAE,MAAM,SAAS,CAAC;AAE5C,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,mCAAmC,CAAC;AAC1E,OAAO,KAAK,EAAE,cAAc,EAA2B,MAAM,wBAAwB,CAAC;AAEtF,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,oBAAoB,CAAC;AAU7D,QAAA,MAAM,YAAY;;;;;EAoBhB,CAAC;AAIH,MAAM,WAAW,iBAAiB;IACjC,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,YAAY,CAAC,EAAE,kBAAkB,CAAC;IAClC,QAAQ,CAAC,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAA;KAAE,CAAC;CAC3C;AAED,MAAM,WAAW,iBAAiB;IACjC;;;;OAIG;IACH,UAAU,CAAC,EAAE,MAAM,gBAAgB,GAAG,SAAS,CAAC;CAChD;AAqCD,wBAAgB,0BAA0B,CACzC,GAAG,EAAE,MAAM,EACX,OAAO,CAAC,EAAE,iBAAiB,GACzB,cAAc,CAAC,OAAO,YAAY,EAAE,iBAAiB,GAAG,SAAS,CAAC,CAoEpE","sourcesContent":["/**\n * The unified `search` tool: ranked code discovery in lexical, semantic, or\n * hybrid (rank-fused) mode. Replaces the old `semantic_search` tool — see\n * docs/hybrid-retrieval-design.md, Decision 1. `grep` stays separate for\n * exact line-level mechanics; this tool answers \"find where X lives\".\n */\n\nimport { Text } from \"@kolisachint/hoocode-tui\";\nimport { type Static, Type } from \"typebox\";\nimport { keyHint } from \"../../modes/interactive/components/keybinding-hints.js\";\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport type { ToolDefinition, ToolRenderResultOptions } from \"../extensions/types.js\";\nimport { runSearch } from \"../search/hybrid-search.js\";\nimport type { ResolvedSearchMode } from \"../search/types.js\";\nimport { getTextOutput, invalidArgText, str } from \"./render-utils.js\";\n\n// Default of 5 balances recall against context cost. Measured on the 62-query\n// eval: limit=5 costs ~1000 tokens and puts a gold span in the results 68% of\n// the time; limit=10 costs ~1600 tokens for 74%. The extra 6 points are real\n// but not worth 60% more context on every call, and the model can ask for more.\nconst DEFAULT_RESULTS = 5;\nconst MAX_RESULTS = 30;\n\nconst searchSchema = Type.Object({\n\tquery: Type.String({\n\t\tdescription:\n\t\t\t\"What to find: an identifier, error text, or a natural-language description of the code, e.g. 'where sessions are persisted to disk'\",\n\t}),\n\tmode: Type.Optional(\n\t\tType.Union([Type.Literal(\"auto\"), Type.Literal(\"lexical\"), Type.Literal(\"semantic\"), Type.Literal(\"hybrid\")], {\n\t\t\tdescription:\n\t\t\t\t\"Retrieval mode (default: auto, which is almost always right). auto = hybrid when the index is available, else lexical; hybrid = keyword and meaning over the index plus exact text for anything indexed later; semantic = index only, skipping exact text; lexical = exact text only, the one mode that works with no index.\",\n\t\t}),\n\t),\n\tglob: Type.Optional(\n\t\tType.String({\n\t\t\tdescription:\n\t\t\t\t\"Optional glob filter applied to file paths. Only file paths matching the glob are searched. Supports both slashless patterns (match base name anywhere) and slash patterns (match full path).\",\n\t\t}),\n\t),\n\tlimit: Type.Optional(\n\t\tType.Number({ description: `Maximum number of results (default: ${DEFAULT_RESULTS}, max: ${MAX_RESULTS})` }),\n\t),\n});\n\ntype SearchToolInput = Static<typeof searchSchema>;\n\nexport interface SearchToolDetails {\n\tresultCount?: number;\n\tresolvedMode?: ResolvedSearchMode;\n\tindexing?: { done: number; total: number };\n}\n\nexport interface SearchToolOptions {\n\t/**\n\t * Provider for the per-session embsearch service. Resolved lazily at call\n\t * time because the service is created (and its index built) after tool\n\t * registration. When absent/unavailable the tool degrades to lexical mode.\n\t */\n\tgetService?: () => EmbsearchService | undefined;\n}\n\nfunction formatSearchCall(\n\targs: { query?: string; mode?: string; limit?: number } | undefined,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n): string {\n\tconst query = str(args?.query);\n\tconst queryDisplay = query === null ? invalidArgText(theme) : `\"${query}\"`;\n\tlet text = theme.fg(\"toolTitle\", theme.bold(\"search \")) + theme.fg(\"accent\", queryDisplay);\n\tconst extras: string[] = [];\n\tif (args?.mode && args.mode !== \"auto\") extras.push(args.mode);\n\tif (args?.limit !== undefined) extras.push(String(args.limit));\n\tif (extras.length > 0) text += theme.fg(\"muted\", ` (${extras.join(\", \")})`);\n\treturn text;\n}\n\nfunction formatSearchResult(\n\tresult: { content: Array<{ type: string; text?: string }>; details?: SearchToolDetails },\n\toptions: ToolRenderResultOptions,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n\tshowImages: boolean,\n): string {\n\tconst output = getTextOutput(result as any, showImages).trim();\n\tlet text = \"\";\n\tif (output) {\n\t\tconst lines = output.split(\"\\n\");\n\t\tconst maxLines = options.expanded ? lines.length : 15;\n\t\tconst displayLines = lines.slice(0, maxLines);\n\t\tconst remaining = lines.length - maxLines;\n\t\ttext += `\\n${displayLines.map((line) => theme.fg(\"toolOutput\", line)).join(\"\\n\")}`;\n\t\tif (remaining > 0) {\n\t\t\ttext += `${theme.fg(\"muted\", `\\n... (${remaining} more lines,`)} ${keyHint(\"app.tools.expand\", \"to expand\")})`;\n\t\t}\n\t}\n\treturn text;\n}\n\nexport function createSearchToolDefinition(\n\tcwd: string,\n\toptions?: SearchToolOptions,\n): ToolDefinition<typeof searchSchema, SearchToolDetails | undefined> {\n\treturn {\n\t\tname: \"search\",\n\t\tlabel: \"search\",\n\t\t// The search-vs-grep split is stated once, by buildSystemPrompt, whenever both\n\t\t// tools are registered — so it is deliberately absent here and from grep's\n\t\t// description. What stays is what only this tool knows: that the query is not\n\t\t// a regex, and that it degrades to exact-text when the index is missing.\n\t\tdescription:\n\t\t\t\"Find where code lives: ranked file:line-range results, fusing keyword and semantic retrieval over a local index with exact-text search of files the index has not read yet. The query is plain text, not a regex — regex metacharacters are matched literally. Falls back to exact-text retrieval automatically when the index is unavailable, and still finds code written moments ago that no index has seen.\",\n\t\tpromptSnippet: \"Ranked code search (keyword + semantic, rank-fused)\",\n\t\tpromptGuidelines: [\n\t\t\t\"search defaults to mode=auto, which is almost always right. Use limit=3 for targeted lookups, 10–20 when exploring a broad topic — past ~15 results the deeper ones arrive as ranked file:line-range headers without a snippet, which is still enough to choose what to read.\",\n\t\t],\n\t\tparameters: searchSchema,\n\t\tasync execute(_toolCallId, { query, mode, glob, limit }: SearchToolInput, signal?: AbortSignal) {\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst result = await runSearch({\n\t\t\t\tcwd,\n\t\t\t\tquery,\n\t\t\t\tmode,\n\t\t\t\tglob,\n\t\t\t\tlimit: Math.min(MAX_RESULTS, Math.max(1, limit ?? DEFAULT_RESULTS)),\n\t\t\t\tservice: options?.getService?.(),\n\t\t\t\tsignal,\n\t\t\t});\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst details: SearchToolDetails = {\n\t\t\t\tresultCount: result.resultCount,\n\t\t\t\tresolvedMode: result.resolvedMode,\n\t\t\t\tindexing: result.indexing,\n\t\t\t};\n\n\t\t\tconst notices: string[] = [];\n\t\t\tif (result.degradedReason) notices.push(result.degradedReason);\n\t\t\tif (result.indexing) {\n\t\t\t\tnotices.push(\n\t\t\t\t\t`index still building: ${result.indexing.done}/${result.indexing.total} chunks embedded — results may be incomplete`,\n\t\t\t\t);\n\t\t\t}\n\t\t\tconst notice = notices.length > 0 ? `\\n\\n[${notices.join(\". \")}]` : \"\";\n\n\t\t\tif (result.resultCount === 0) {\n\t\t\t\treturn {\n\t\t\t\t\tcontent: [\n\t\t\t\t\t\t{ type: \"text\" as const, text: `No results for \"${query}\" (${result.resolvedMode})${notice}` },\n\t\t\t\t\t],\n\t\t\t\t\tdetails,\n\t\t\t\t};\n\t\t\t}\n\t\t\treturn {\n\t\t\t\tcontent: [{ type: \"text\" as const, text: result.text + notice }],\n\t\t\t\tdetails,\n\t\t\t};\n\t\t},\n\t\trenderCall(args, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchCall(args, theme));\n\t\t\treturn text;\n\t\t},\n\t\trenderResult(result, options, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchResult(result as any, options, theme, context.showImages));\n\t\t\treturn text;\n\t\t},\n\t};\n}\n"]}

package/dist/core/tools/search.js

@@ -9,8 +9,10 @@ import { Type } from "typebox";
 import { keyHint } from "../../modes/interactive/components/keybinding-hints.js";
 import { runSearch } from "../search/hybrid-search.js";
 import { getTextOutput, invalidArgText, str } from "./render-utils.js";
-// Default of 5 balances recall against context cost: eval showed limit=5
-// preserving the top-rank hits of limit=10 at roughly half the tokens.
+// Default of 5 balances recall against context cost. Measured on the 62-query
+// eval: limit=5 costs ~1000 tokens and puts a gold span in the results 68% of
+// the time; limit=10 costs ~1600 tokens for 74%. The extra 6 points are real
+// but not worth 60% more context on every call, and the model can ask for more.
 const DEFAULT_RESULTS = 5;
 const MAX_RESULTS = 30;
 const searchSchema = Type.Object({
@@ -18,7 +20,7 @@ const searchSchema = Type.Object({
         description: "What to find: an identifier, error text, or a natural-language description of the code, e.g. 'where sessions are persisted to disk'",
     }),
     mode: Type.Optional(Type.Union([Type.Literal("auto"), Type.Literal("lexical"), Type.Literal("semantic"), Type.Literal("hybrid")], {
-        description: "Retrieval mode (default: auto). auto picks hybrid when the semantic index is available; lexical = exact-text only; semantic = embedding index only; hybrid = both, fused by rank.",
+        description: "Retrieval mode (default: auto, which is almost always right). auto = hybrid when the index is available, else lexical; hybrid = keyword and meaning over the index plus exact text for anything indexed later; semantic = index only, skipping exact text; lexical = exact text only, the one mode that works with no index.",
     })),
     glob: Type.Optional(Type.String({
         description: "Optional glob filter applied to file paths. Only file paths matching the glob are searched. Supports both slashless patterns (match base name anywhere) and slash patterns (match full path).",
@@ -61,10 +63,10 @@ export function createSearchToolDefinition(cwd, options) {
         // tools are registered — so it is deliberately absent here and from grep's
         // description. What stays is what only this tool knows: that the query is not
         // a regex, and that it degrades to exact-text when the index is missing.
-        description: "Find where code lives: ranked file:line-range results from exact-text and semantic (local embedding index) retrieval, fused by rank when both are available. The query is plain text, not a regex — regex metacharacters are matched literally. Falls back to exact-text retrieval automatically when the semantic index is unavailable.",
-        promptSnippet: "Ranked code search (exact + semantic, rank-fused)",
+        description: "Find where code lives: ranked file:line-range results, fusing keyword and semantic retrieval over a local index with exact-text search of files the index has not read yet. The query is plain text, not a regex — regex metacharacters are matched literally. Falls back to exact-text retrieval automatically when the index is unavailable, and still finds code written moments ago that no index has seen.",
+        promptSnippet: "Ranked code search (keyword + semantic, rank-fused)",
         promptGuidelines: [
-            "search defaults to mode=auto, which is almost always right. Use limit=3 for targeted lookups, 10–20 when exploring a broad topic.",
+            "search defaults to mode=auto, which is almost always right. Use limit=3 for targeted lookups, 10–20 when exploring a broad topic — past ~15 results the deeper ones arrive as ranked file:line-range headers without a snippet, which is still enough to choose what to read.",
         ],
         parameters: searchSchema,
         async execute(_toolCallId, { query, mode, glob, limit }, signal) {

package/dist/core/tools/search.js.map

@@ -1 +1 @@
-{"version":3,"file":"search.js","sourceRoot":"","sources":["../../../src/core/tools/search.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,IAAI,EAAE,MAAM,0BAA0B,CAAC;AAChD,OAAO,EAAe,IAAI,EAAE,MAAM,SAAS,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,wDAAwD,CAAC;AAGjF,OAAO,EAAE,SAAS,EAAE,MAAM,4BAA4B,CAAC;AAEvD,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,GAAG,EAAE,MAAM,mBAAmB,CAAC;AAEvE,yEAAyE;AACzE,uEAAuE;AACvE,MAAM,eAAe,GAAG,CAAC,CAAC;AAC1B,MAAM,WAAW,GAAG,EAAE,CAAC;AAEvB,MAAM,YAAY,GAAG,IAAI,CAAC,MAAM,CAAC;IAChC,KAAK,EAAE,IAAI,CAAC,MAAM,CAAC;QAClB,WAAW,EACV,qIAAqI;KACtI,CAAC;IACF,IAAI,EAAE,IAAI,CAAC,QAAQ,CAClB,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,EAAE;QAC7G,WAAW,EACV,mLAAmL;KACpL,CAAC,CACF;IACD,IAAI,EAAE,IAAI,CAAC,QAAQ,CAClB,IAAI,CAAC,MAAM,CAAC;QACX,WAAW,EACV,+LAA+L;KAChM,CAAC,CACF;IACD,KAAK,EAAE,IAAI,CAAC,QAAQ,CACnB,IAAI,CAAC,MAAM,CAAC,EAAE,WAAW,EAAE,uCAAuC,eAAe,UAAU,WAAW,GAAG,EAAE,CAAC,CAC5G;CACD,CAAC,CAAC;AAmBH,SAAS,gBAAgB,CACxB,IAAmE,EACnE,KAAoE,EAC3D;IACT,MAAM,KAAK,GAAG,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC/B,MAAM,YAAY,GAAG,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,cAAc,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC;IAC3E,IAAI,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,WAAW,EAAE,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,GAAG,KAAK,CAAC,EAAE,CAAC,QAAQ,EAAE,YAAY,CAAC,CAAC;IAC3F,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,IAAI,EAAE,IAAI,IAAI,IAAI,CAAC,IAAI,KAAK,MAAM;QAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC/D,IAAI,IAAI,EAAE,KAAK,KAAK,SAAS;QAAE,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IAC/D,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;QAAE,IAAI,IAAI,KAAK,CAAC,EAAE,CAAC,OAAO,EAAE,KAAK,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC5E,OAAO,IAAI,CAAC;AAAA,CACZ;AAED,SAAS,kBAAkB,CAC1B,MAAwF,EACxF,OAAgC,EAChC,KAAoE,EACpE,UAAmB,EACV;IACT,MAAM,MAAM,GAAG,aAAa,CAAC,MAAa,EAAE,UAAU,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/D,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,IAAI,MAAM,EAAE,CAAC;QACZ,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACjC,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;QACtD,MAAM,YAAY,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC;QAC9C,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,GAAG,QAAQ,CAAC;QAC1C,IAAI,IAAI,KAAK,YAAY,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,CAAC,EAAE,CAAC,YAAY,EAAE,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACnF,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;YACnB,IAAI,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,OAAO,EAAE,UAAU,SAAS,cAAc,CAAC,IAAI,OAAO,CAAC,kBAAkB,EAAE,WAAW,CAAC,GAAG,CAAC;QAChH,CAAC;IACF,CAAC;IACD,OAAO,IAAI,CAAC;AAAA,CACZ;AAED,MAAM,UAAU,0BAA0B,CACzC,GAAW,EACX,OAA2B,EAC0C;IACrE,OAAO;QACN,IAAI,EAAE,QAAQ;QACd,KAAK,EAAE,QAAQ;QACf,+EAA+E;QAC/E,6EAA2E;QAC3E,8EAA8E;QAC9E,yEAAyE;QACzE,WAAW,EACV,4UAA0U;QAC3U,aAAa,EAAE,mDAAmD;QAClE,gBAAgB,EAAE;YACjB,qIAAmI;SACnI;QACD,UAAU,EAAE,YAAY;QACxB,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAmB,EAAE,MAAoB,EAAE;YAC/F,IAAI,MAAM,EAAE,OAAO;gBAAE,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC;YAE1D,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC;gBAC9B,GAAG;gBACH,KAAK;gBACL,IAAI;gBACJ,IAAI;gBACJ,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,IAAI,eAAe,CAAC,CAAC;gBACnE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,EAAE;gBAChC,MAAM;aACN,CAAC,CAAC;YACH,IAAI,MAAM,EAAE,OAAO;gBAAE,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC;YAE1D,MAAM,OAAO,GAAsB;gBAClC,WAAW,EAAE,MAAM,CAAC,WAAW;gBAC/B,YAAY,EAAE,MAAM,CAAC,YAAY;gBACjC,QAAQ,EAAE,MAAM,CAAC,QAAQ;aACzB,CAAC;YAEF,MAAM,OAAO,GAAa,EAAE,CAAC;YAC7B,IAAI,MAAM,CAAC,cAAc;gBAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC;YAC/D,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;gBACrB,OAAO,CAAC,IAAI,CACX,yBAAyB,MAAM,CAAC,QAAQ,CAAC,IAAI,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,gDAA8C,CACpH,CAAC;YACH,CAAC;YACD,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;YAEvE,IAAI,MAAM,CAAC,WAAW,KAAK,CAAC,EAAE,CAAC;gBAC9B,OAAO;oBACN,OAAO,EAAE;wBACR,EAAE,IAAI,EAAE,MAAe,EAAE,IAAI,EAAE,mBAAmB,KAAK,MAAM,MAAM,CAAC,YAAY,IAAI,MAAM,EAAE,EAAE;qBAC9F;oBACD,OAAO;iBACP,CAAC;YACH,CAAC;YACD,OAAO;gBACN,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,MAAe,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,GAAG,MAAM,EAAE,CAAC;gBAChE,OAAO;aACP,CAAC;QAAA,CACF;QACD,UAAU,CAAC,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE;YAChC,MAAM,IAAI,GAAI,OAAO,CAAC,aAAkC,IAAI,IAAI,IAAI,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;YAC/E,IAAI,CAAC,OAAO,CAAC,gBAAgB,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;YAC5C,OAAO,IAAI,CAAC;QAAA,CACZ;QACD,YAAY,CAAC,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE;YAC7C,MAAM,IAAI,GAAI,OAAO,CAAC,aAAkC,IAAI,IAAI,IAAI,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;YAC/E,IAAI,CAAC,OAAO,CAAC,kBAAkB,CAAC,MAAa,EAAE,OAAO,EAAE,KAAK,EAAE,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC;YACpF,OAAO,IAAI,CAAC;QAAA,CACZ;KACD,CAAC;AAAA,CACF","sourcesContent":["/**\n * The unified `search` tool: ranked code discovery in lexical, semantic, or\n * hybrid (rank-fused) mode. Replaces the old `semantic_search` tool — see\n * docs/hybrid-retrieval-design.md, Decision 1. `grep` stays separate for\n * exact line-level mechanics; this tool answers \"find where X lives\".\n */\n\nimport { Text } from \"@kolisachint/hoocode-tui\";\nimport { type Static, Type } from \"typebox\";\nimport { keyHint } from \"../../modes/interactive/components/keybinding-hints.js\";\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport type { ToolDefinition, ToolRenderResultOptions } from \"../extensions/types.js\";\nimport { runSearch } from \"../search/hybrid-search.js\";\nimport type { ResolvedSearchMode } from \"../search/types.js\";\nimport { getTextOutput, invalidArgText, str } from \"./render-utils.js\";\n\n// Default of 5 balances recall against context cost: eval showed limit=5\n// preserving the top-rank hits of limit=10 at roughly half the tokens.\nconst DEFAULT_RESULTS = 5;\nconst MAX_RESULTS = 30;\n\nconst searchSchema = Type.Object({\n\tquery: Type.String({\n\t\tdescription:\n\t\t\t\"What to find: an identifier, error text, or a natural-language description of the code, e.g. 'where sessions are persisted to disk'\",\n\t}),\n\tmode: Type.Optional(\n\t\tType.Union([Type.Literal(\"auto\"), Type.Literal(\"lexical\"), Type.Literal(\"semantic\"), Type.Literal(\"hybrid\")], {\n\t\t\tdescription:\n\t\t\t\t\"Retrieval mode (default: auto). auto picks hybrid when the semantic index is available; lexical = exact-text only; semantic = embedding index only; hybrid = both, fused by rank.\",\n\t\t}),\n\t),\n\tglob: Type.Optional(\n\t\tType.String({\n\t\t\tdescription:\n\t\t\t\t\"Optional glob filter applied to file paths. Only file paths matching the glob are searched. Supports both slashless patterns (match base name anywhere) and slash patterns (match full path).\",\n\t\t}),\n\t),\n\tlimit: Type.Optional(\n\t\tType.Number({ description: `Maximum number of results (default: ${DEFAULT_RESULTS}, max: ${MAX_RESULTS})` }),\n\t),\n});\n\ntype SearchToolInput = Static<typeof searchSchema>;\n\nexport interface SearchToolDetails {\n\tresultCount?: number;\n\tresolvedMode?: ResolvedSearchMode;\n\tindexing?: { done: number; total: number };\n}\n\nexport interface SearchToolOptions {\n\t/**\n\t * Provider for the per-session embsearch service. Resolved lazily at call\n\t * time because the service is created (and its index built) after tool\n\t * registration. When absent/unavailable the tool degrades to lexical mode.\n\t */\n\tgetService?: () => EmbsearchService | undefined;\n}\n\nfunction formatSearchCall(\n\targs: { query?: string; mode?: string; limit?: number } | undefined,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n): string {\n\tconst query = str(args?.query);\n\tconst queryDisplay = query === null ? invalidArgText(theme) : `\"${query}\"`;\n\tlet text = theme.fg(\"toolTitle\", theme.bold(\"search \")) + theme.fg(\"accent\", queryDisplay);\n\tconst extras: string[] = [];\n\tif (args?.mode && args.mode !== \"auto\") extras.push(args.mode);\n\tif (args?.limit !== undefined) extras.push(String(args.limit));\n\tif (extras.length > 0) text += theme.fg(\"muted\", ` (${extras.join(\", \")})`);\n\treturn text;\n}\n\nfunction formatSearchResult(\n\tresult: { content: Array<{ type: string; text?: string }>; details?: SearchToolDetails },\n\toptions: ToolRenderResultOptions,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n\tshowImages: boolean,\n): string {\n\tconst output = getTextOutput(result as any, showImages).trim();\n\tlet text = \"\";\n\tif (output) {\n\t\tconst lines = output.split(\"\\n\");\n\t\tconst maxLines = options.expanded ? lines.length : 15;\n\t\tconst displayLines = lines.slice(0, maxLines);\n\t\tconst remaining = lines.length - maxLines;\n\t\ttext += `\\n${displayLines.map((line) => theme.fg(\"toolOutput\", line)).join(\"\\n\")}`;\n\t\tif (remaining > 0) {\n\t\t\ttext += `${theme.fg(\"muted\", `\\n... (${remaining} more lines,`)} ${keyHint(\"app.tools.expand\", \"to expand\")})`;\n\t\t}\n\t}\n\treturn text;\n}\n\nexport function createSearchToolDefinition(\n\tcwd: string,\n\toptions?: SearchToolOptions,\n): ToolDefinition<typeof searchSchema, SearchToolDetails | undefined> {\n\treturn {\n\t\tname: \"search\",\n\t\tlabel: \"search\",\n\t\t// The search-vs-grep split is stated once, by buildSystemPrompt, whenever both\n\t\t// tools are registered — so it is deliberately absent here and from grep's\n\t\t// description. What stays is what only this tool knows: that the query is not\n\t\t// a regex, and that it degrades to exact-text when the index is missing.\n\t\tdescription:\n\t\t\t\"Find where code lives: ranked file:line-range results from exact-text and semantic (local embedding index) retrieval, fused by rank when both are available. The query is plain text, not a regex — regex metacharacters are matched literally. Falls back to exact-text retrieval automatically when the semantic index is unavailable.\",\n\t\tpromptSnippet: \"Ranked code search (exact + semantic, rank-fused)\",\n\t\tpromptGuidelines: [\n\t\t\t\"search defaults to mode=auto, which is almost always right. Use limit=3 for targeted lookups, 10–20 when exploring a broad topic.\",\n\t\t],\n\t\tparameters: searchSchema,\n\t\tasync execute(_toolCallId, { query, mode, glob, limit }: SearchToolInput, signal?: AbortSignal) {\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst result = await runSearch({\n\t\t\t\tcwd,\n\t\t\t\tquery,\n\t\t\t\tmode,\n\t\t\t\tglob,\n\t\t\t\tlimit: Math.min(MAX_RESULTS, Math.max(1, limit ?? DEFAULT_RESULTS)),\n\t\t\t\tservice: options?.getService?.(),\n\t\t\t\tsignal,\n\t\t\t});\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst details: SearchToolDetails = {\n\t\t\t\tresultCount: result.resultCount,\n\t\t\t\tresolvedMode: result.resolvedMode,\n\t\t\t\tindexing: result.indexing,\n\t\t\t};\n\n\t\t\tconst notices: string[] = [];\n\t\t\tif (result.degradedReason) notices.push(result.degradedReason);\n\t\t\tif (result.indexing) {\n\t\t\t\tnotices.push(\n\t\t\t\t\t`index still building: ${result.indexing.done}/${result.indexing.total} chunks embedded — results may be incomplete`,\n\t\t\t\t);\n\t\t\t}\n\t\t\tconst notice = notices.length > 0 ? `\\n\\n[${notices.join(\". \")}]` : \"\";\n\n\t\t\tif (result.resultCount === 0) {\n\t\t\t\treturn {\n\t\t\t\t\tcontent: [\n\t\t\t\t\t\t{ type: \"text\" as const, text: `No results for \"${query}\" (${result.resolvedMode})${notice}` },\n\t\t\t\t\t],\n\t\t\t\t\tdetails,\n\t\t\t\t};\n\t\t\t}\n\t\t\treturn {\n\t\t\t\tcontent: [{ type: \"text\" as const, text: result.text + notice }],\n\t\t\t\tdetails,\n\t\t\t};\n\t\t},\n\t\trenderCall(args, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchCall(args, theme));\n\t\t\treturn text;\n\t\t},\n\t\trenderResult(result, options, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchResult(result as any, options, theme, context.showImages));\n\t\t\treturn text;\n\t\t},\n\t};\n}\n"]}
+{"version":3,"file":"search.js","sourceRoot":"","sources":["../../../src/core/tools/search.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,IAAI,EAAE,MAAM,0BAA0B,CAAC;AAChD,OAAO,EAAe,IAAI,EAAE,MAAM,SAAS,CAAC;AAC5C,OAAO,EAAE,OAAO,EAAE,MAAM,wDAAwD,CAAC;AAGjF,OAAO,EAAE,SAAS,EAAE,MAAM,4BAA4B,CAAC;AAEvD,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,GAAG,EAAE,MAAM,mBAAmB,CAAC;AAEvE,8EAA8E;AAC9E,8EAA8E;AAC9E,6EAA6E;AAC7E,gFAAgF;AAChF,MAAM,eAAe,GAAG,CAAC,CAAC;AAC1B,MAAM,WAAW,GAAG,EAAE,CAAC;AAEvB,MAAM,YAAY,GAAG,IAAI,CAAC,MAAM,CAAC;IAChC,KAAK,EAAE,IAAI,CAAC,MAAM,CAAC;QAClB,WAAW,EACV,qIAAqI;KACtI,CAAC;IACF,IAAI,EAAE,IAAI,CAAC,QAAQ,CAClB,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,EAAE;QAC7G,WAAW,EACV,8TAA8T;KAC/T,CAAC,CACF;IACD,IAAI,EAAE,IAAI,CAAC,QAAQ,CAClB,IAAI,CAAC,MAAM,CAAC;QACX,WAAW,EACV,+LAA+L;KAChM,CAAC,CACF;IACD,KAAK,EAAE,IAAI,CAAC,QAAQ,CACnB,IAAI,CAAC,MAAM,CAAC,EAAE,WAAW,EAAE,uCAAuC,eAAe,UAAU,WAAW,GAAG,EAAE,CAAC,CAC5G;CACD,CAAC,CAAC;AAmBH,SAAS,gBAAgB,CACxB,IAAmE,EACnE,KAAoE,EAC3D;IACT,MAAM,KAAK,GAAG,GAAG,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC/B,MAAM,YAAY,GAAG,KAAK,KAAK,IAAI,CAAC,CAAC,CAAC,cAAc,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC;IAC3E,IAAI,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,WAAW,EAAE,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,GAAG,KAAK,CAAC,EAAE,CAAC,QAAQ,EAAE,YAAY,CAAC,CAAC;IAC3F,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,IAAI,EAAE,IAAI,IAAI,IAAI,CAAC,IAAI,KAAK,MAAM;QAAE,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC/D,IAAI,IAAI,EAAE,KAAK,KAAK,SAAS;QAAE,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IAC/D,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC;QAAE,IAAI,IAAI,KAAK,CAAC,EAAE,CAAC,OAAO,EAAE,KAAK,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC5E,OAAO,IAAI,CAAC;AAAA,CACZ;AAED,SAAS,kBAAkB,CAC1B,MAAwF,EACxF,OAAgC,EAChC,KAAoE,EACpE,UAAmB,EACV;IACT,MAAM,MAAM,GAAG,aAAa,CAAC,MAAa,EAAE,UAAU,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/D,IAAI,IAAI,GAAG,EAAE,CAAC;IACd,IAAI,MAAM,EAAE,CAAC;QACZ,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACjC,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;QACtD,MAAM,YAAY,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC;QAC9C,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,GAAG,QAAQ,CAAC;QAC1C,IAAI,IAAI,KAAK,YAAY,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,CAAC,EAAE,CAAC,YAAY,EAAE,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACnF,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;YACnB,IAAI,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,OAAO,EAAE,UAAU,SAAS,cAAc,CAAC,IAAI,OAAO,CAAC,kBAAkB,EAAE,WAAW,CAAC,GAAG,CAAC;QAChH,CAAC;IACF,CAAC;IACD,OAAO,IAAI,CAAC;AAAA,CACZ;AAED,MAAM,UAAU,0BAA0B,CACzC,GAAW,EACX,OAA2B,EAC0C;IACrE,OAAO;QACN,IAAI,EAAE,QAAQ;QACd,KAAK,EAAE,QAAQ;QACf,+EAA+E;QAC/E,6EAA2E;QAC3E,8EAA8E;QAC9E,yEAAyE;QACzE,WAAW,EACV,mZAAiZ;QAClZ,aAAa,EAAE,qDAAqD;QACpE,gBAAgB,EAAE;YACjB,mRAA+Q;SAC/Q;QACD,UAAU,EAAE,YAAY;QACxB,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAmB,EAAE,MAAoB,EAAE;YAC/F,IAAI,MAAM,EAAE,OAAO;gBAAE,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC;YAE1D,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC;gBAC9B,GAAG;gBACH,KAAK;gBACL,IAAI;gBACJ,IAAI;gBACJ,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,WAAW,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,IAAI,eAAe,CAAC,CAAC;gBACnE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,EAAE;gBAChC,MAAM;aACN,CAAC,CAAC;YACH,IAAI,MAAM,EAAE,OAAO;gBAAE,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC;YAE1D,MAAM,OAAO,GAAsB;gBAClC,WAAW,EAAE,MAAM,CAAC,WAAW;gBAC/B,YAAY,EAAE,MAAM,CAAC,YAAY;gBACjC,QAAQ,EAAE,MAAM,CAAC,QAAQ;aACzB,CAAC;YAEF,MAAM,OAAO,GAAa,EAAE,CAAC;YAC7B,IAAI,MAAM,CAAC,cAAc;gBAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC;YAC/D,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;gBACrB,OAAO,CAAC,IAAI,CACX,yBAAyB,MAAM,CAAC,QAAQ,CAAC,IAAI,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,gDAA8C,CACpH,CAAC;YACH,CAAC;YACD,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;YAEvE,IAAI,MAAM,CAAC,WAAW,KAAK,CAAC,EAAE,CAAC;gBAC9B,OAAO;oBACN,OAAO,EAAE;wBACR,EAAE,IAAI,EAAE,MAAe,EAAE,IAAI,EAAE,mBAAmB,KAAK,MAAM,MAAM,CAAC,YAAY,IAAI,MAAM,EAAE,EAAE;qBAC9F;oBACD,OAAO;iBACP,CAAC;YACH,CAAC;YACD,OAAO;gBACN,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,MAAe,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,GAAG,MAAM,EAAE,CAAC;gBAChE,OAAO;aACP,CAAC;QAAA,CACF;QACD,UAAU,CAAC,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE;YAChC,MAAM,IAAI,GAAI,OAAO,CAAC,aAAkC,IAAI,IAAI,IAAI,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;YAC/E,IAAI,CAAC,OAAO,CAAC,gBAAgB,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;YAC5C,OAAO,IAAI,CAAC;QAAA,CACZ;QACD,YAAY,CAAC,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE;YAC7C,MAAM,IAAI,GAAI,OAAO,CAAC,aAAkC,IAAI,IAAI,IAAI,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;YAC/E,IAAI,CAAC,OAAO,CAAC,kBAAkB,CAAC,MAAa,EAAE,OAAO,EAAE,KAAK,EAAE,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC;YACpF,OAAO,IAAI,CAAC;QAAA,CACZ;KACD,CAAC;AAAA,CACF","sourcesContent":["/**\n * The unified `search` tool: ranked code discovery in lexical, semantic, or\n * hybrid (rank-fused) mode. Replaces the old `semantic_search` tool — see\n * docs/hybrid-retrieval-design.md, Decision 1. `grep` stays separate for\n * exact line-level mechanics; this tool answers \"find where X lives\".\n */\n\nimport { Text } from \"@kolisachint/hoocode-tui\";\nimport { type Static, Type } from \"typebox\";\nimport { keyHint } from \"../../modes/interactive/components/keybinding-hints.js\";\nimport type { EmbsearchService } from \"../embsearch/embsearch-service.js\";\nimport type { ToolDefinition, ToolRenderResultOptions } from \"../extensions/types.js\";\nimport { runSearch } from \"../search/hybrid-search.js\";\nimport type { ResolvedSearchMode } from \"../search/types.js\";\nimport { getTextOutput, invalidArgText, str } from \"./render-utils.js\";\n\n// Default of 5 balances recall against context cost. Measured on the 62-query\n// eval: limit=5 costs ~1000 tokens and puts a gold span in the results 68% of\n// the time; limit=10 costs ~1600 tokens for 74%. The extra 6 points are real\n// but not worth 60% more context on every call, and the model can ask for more.\nconst DEFAULT_RESULTS = 5;\nconst MAX_RESULTS = 30;\n\nconst searchSchema = Type.Object({\n\tquery: Type.String({\n\t\tdescription:\n\t\t\t\"What to find: an identifier, error text, or a natural-language description of the code, e.g. 'where sessions are persisted to disk'\",\n\t}),\n\tmode: Type.Optional(\n\t\tType.Union([Type.Literal(\"auto\"), Type.Literal(\"lexical\"), Type.Literal(\"semantic\"), Type.Literal(\"hybrid\")], {\n\t\t\tdescription:\n\t\t\t\t\"Retrieval mode (default: auto, which is almost always right). auto = hybrid when the index is available, else lexical; hybrid = keyword and meaning over the index plus exact text for anything indexed later; semantic = index only, skipping exact text; lexical = exact text only, the one mode that works with no index.\",\n\t\t}),\n\t),\n\tglob: Type.Optional(\n\t\tType.String({\n\t\t\tdescription:\n\t\t\t\t\"Optional glob filter applied to file paths. Only file paths matching the glob are searched. Supports both slashless patterns (match base name anywhere) and slash patterns (match full path).\",\n\t\t}),\n\t),\n\tlimit: Type.Optional(\n\t\tType.Number({ description: `Maximum number of results (default: ${DEFAULT_RESULTS}, max: ${MAX_RESULTS})` }),\n\t),\n});\n\ntype SearchToolInput = Static<typeof searchSchema>;\n\nexport interface SearchToolDetails {\n\tresultCount?: number;\n\tresolvedMode?: ResolvedSearchMode;\n\tindexing?: { done: number; total: number };\n}\n\nexport interface SearchToolOptions {\n\t/**\n\t * Provider for the per-session embsearch service. Resolved lazily at call\n\t * time because the service is created (and its index built) after tool\n\t * registration. When absent/unavailable the tool degrades to lexical mode.\n\t */\n\tgetService?: () => EmbsearchService | undefined;\n}\n\nfunction formatSearchCall(\n\targs: { query?: string; mode?: string; limit?: number } | undefined,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n): string {\n\tconst query = str(args?.query);\n\tconst queryDisplay = query === null ? invalidArgText(theme) : `\"${query}\"`;\n\tlet text = theme.fg(\"toolTitle\", theme.bold(\"search \")) + theme.fg(\"accent\", queryDisplay);\n\tconst extras: string[] = [];\n\tif (args?.mode && args.mode !== \"auto\") extras.push(args.mode);\n\tif (args?.limit !== undefined) extras.push(String(args.limit));\n\tif (extras.length > 0) text += theme.fg(\"muted\", ` (${extras.join(\", \")})`);\n\treturn text;\n}\n\nfunction formatSearchResult(\n\tresult: { content: Array<{ type: string; text?: string }>; details?: SearchToolDetails },\n\toptions: ToolRenderResultOptions,\n\ttheme: typeof import(\"../../modes/interactive/theme/theme.js\").theme,\n\tshowImages: boolean,\n): string {\n\tconst output = getTextOutput(result as any, showImages).trim();\n\tlet text = \"\";\n\tif (output) {\n\t\tconst lines = output.split(\"\\n\");\n\t\tconst maxLines = options.expanded ? lines.length : 15;\n\t\tconst displayLines = lines.slice(0, maxLines);\n\t\tconst remaining = lines.length - maxLines;\n\t\ttext += `\\n${displayLines.map((line) => theme.fg(\"toolOutput\", line)).join(\"\\n\")}`;\n\t\tif (remaining > 0) {\n\t\t\ttext += `${theme.fg(\"muted\", `\\n... (${remaining} more lines,`)} ${keyHint(\"app.tools.expand\", \"to expand\")})`;\n\t\t}\n\t}\n\treturn text;\n}\n\nexport function createSearchToolDefinition(\n\tcwd: string,\n\toptions?: SearchToolOptions,\n): ToolDefinition<typeof searchSchema, SearchToolDetails | undefined> {\n\treturn {\n\t\tname: \"search\",\n\t\tlabel: \"search\",\n\t\t// The search-vs-grep split is stated once, by buildSystemPrompt, whenever both\n\t\t// tools are registered — so it is deliberately absent here and from grep's\n\t\t// description. What stays is what only this tool knows: that the query is not\n\t\t// a regex, and that it degrades to exact-text when the index is missing.\n\t\tdescription:\n\t\t\t\"Find where code lives: ranked file:line-range results, fusing keyword and semantic retrieval over a local index with exact-text search of files the index has not read yet. The query is plain text, not a regex — regex metacharacters are matched literally. Falls back to exact-text retrieval automatically when the index is unavailable, and still finds code written moments ago that no index has seen.\",\n\t\tpromptSnippet: \"Ranked code search (keyword + semantic, rank-fused)\",\n\t\tpromptGuidelines: [\n\t\t\t\"search defaults to mode=auto, which is almost always right. Use limit=3 for targeted lookups, 10–20 when exploring a broad topic — past ~15 results the deeper ones arrive as ranked file:line-range headers without a snippet, which is still enough to choose what to read.\",\n\t\t],\n\t\tparameters: searchSchema,\n\t\tasync execute(_toolCallId, { query, mode, glob, limit }: SearchToolInput, signal?: AbortSignal) {\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst result = await runSearch({\n\t\t\t\tcwd,\n\t\t\t\tquery,\n\t\t\t\tmode,\n\t\t\t\tglob,\n\t\t\t\tlimit: Math.min(MAX_RESULTS, Math.max(1, limit ?? DEFAULT_RESULTS)),\n\t\t\t\tservice: options?.getService?.(),\n\t\t\t\tsignal,\n\t\t\t});\n\t\t\tif (signal?.aborted) throw new Error(\"Operation aborted\");\n\n\t\t\tconst details: SearchToolDetails = {\n\t\t\t\tresultCount: result.resultCount,\n\t\t\t\tresolvedMode: result.resolvedMode,\n\t\t\t\tindexing: result.indexing,\n\t\t\t};\n\n\t\t\tconst notices: string[] = [];\n\t\t\tif (result.degradedReason) notices.push(result.degradedReason);\n\t\t\tif (result.indexing) {\n\t\t\t\tnotices.push(\n\t\t\t\t\t`index still building: ${result.indexing.done}/${result.indexing.total} chunks embedded — results may be incomplete`,\n\t\t\t\t);\n\t\t\t}\n\t\t\tconst notice = notices.length > 0 ? `\\n\\n[${notices.join(\". \")}]` : \"\";\n\n\t\t\tif (result.resultCount === 0) {\n\t\t\t\treturn {\n\t\t\t\t\tcontent: [\n\t\t\t\t\t\t{ type: \"text\" as const, text: `No results for \"${query}\" (${result.resolvedMode})${notice}` },\n\t\t\t\t\t],\n\t\t\t\t\tdetails,\n\t\t\t\t};\n\t\t\t}\n\t\t\treturn {\n\t\t\t\tcontent: [{ type: \"text\" as const, text: result.text + notice }],\n\t\t\t\tdetails,\n\t\t\t};\n\t\t},\n\t\trenderCall(args, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchCall(args, theme));\n\t\t\treturn text;\n\t\t},\n\t\trenderResult(result, options, theme, context) {\n\t\t\tconst text = (context.lastComponent as Text | undefined) ?? new Text(\"\", 0, 0);\n\t\t\ttext.setText(formatSearchResult(result as any, options, theme, context.showImages));\n\t\t\treturn text;\n\t\t},\n\t};\n}\n"]}

package/examples/extensions/custom-provider-anthropic/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "@kolisachint/hoocode-extension-custom-provider-anthropic",
 	"private": true,
-	"version": "0.3.0",
+	"version": "0.3.2",
 	"type": "module",
 	"engines": {
 		"bun": ">=1.0.0"

package/examples/extensions/custom-provider-gitlab-duo/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "@kolisachint/hoocode-extension-custom-provider-gitlab-duo",
 	"private": true,
-	"version": "0.3.0",
+	"version": "0.3.2",
 	"type": "module",
 	"engines": {
 		"bun": ">=1.0.0"

package/examples/extensions/sandbox/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "@kolisachint/hoocode-extension-sandbox",
 	"private": true,
-	"version": "0.3.0",
+	"version": "0.3.2",
 	"type": "module",
 	"engines": {
 		"bun": ">=1.0.0"

package/examples/extensions/with-deps/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "@kolisachint/hoocode-extension-with-deps",
 	"private": true,
-	"version": "0.3.0",
+	"version": "0.3.2",
 	"type": "module",
 	"engines": {
 		"bun": ">=1.0.0"

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@kolisachint/hoocode-agent",
-	"version": "0.5.0",
+	"version": "0.5.2",
 	"description": "Coding agent CLI with read, bash, edit, write tools and session management",
 	"type": "module",
 	"hoocodeConfig": {
@@ -48,9 +48,9 @@
 		"prepublishOnly": "npm run clean && npm run build"
 	},
 	"dependencies": {
-		"@kolisachint/hoocode-agent-core": "^0.5.0",
-		"@kolisachint/hoocode-ai": "^0.5.0",
-		"@kolisachint/hoocode-tui": "^0.5.0",
+		"@kolisachint/hoocode-agent-core": "^0.5.2",
+		"@kolisachint/hoocode-ai": "^0.5.2",
+		"@kolisachint/hoocode-tui": "^0.5.2",
 		"@silvia-odwyer/photon-node": "^0.3.4",
 		"chalk": "^5.5.0",
 		"cli-highlight": "^2.1.11",