open-codebase-index 0.22.4 → 0.23.0

package/dist/pi-extension.cjs

@@ -2342,7 +2342,7 @@ function analyzeQueryIntent(query) {
 }
 function isTestPath(filePath) {
   const normalized = normalizePath(filePath);
-  return /(?:^|\/)(?:test|tests|__tests__|spec|specs)(?:\/|$)/u.test(normalized) || /\.(?:test|spec)\.[^/]+$/u.test(normalized);
+  return /(?:^|\/)(?:test|tests|__tests__|spec|specs)(?:\/|$)/u.test(normalized) || /(?:\.(?:test|spec)|_(?:test|spec))\.[^/]+$/u.test(normalized) || /(?:^|\/)(?:test|spec)_[^/]+\.[^/]+$/u.test(normalized);
 }
 function isFixturePath(filePath) {
   const normalized = normalizePath(filePath);
@@ -6432,85 +6432,6 @@ function createEmbeddingProvider(configuredProviderInfo) {
   }
 }
 
-// src/rerank/index.ts
-function createReranker(config) {
-  if (!config.enabled) {
-    return new NoOpReranker();
-  }
-  return new SiliconFlowReranker(config);
-}
-var NoOpReranker = class {
-  isAvailable() {
-    return false;
-  }
-  async rerank(_query, documents, _topN) {
-    return {
-      results: documents.map((_, index) => ({ index, relevanceScore: 0 }))
-    };
-  }
-};
-var SiliconFlowReranker = class {
-  config;
-  constructor(config) {
-    this.config = config;
-  }
-  isAvailable() {
-    return this.config.enabled && !!this.config.baseUrl && !!this.config.model;
-  }
-  async rerank(query, documents, topN) {
-    if (documents.length === 0) {
-      return { results: [] };
-    }
-    const headers = {
-      "Content-Type": "application/json"
-    };
-    if (this.config.apiKey) {
-      headers["Authorization"] = `Bearer ${this.config.apiKey}`;
-    }
-    const baseUrl = this.config.baseUrl;
-    if (!baseUrl) {
-      throw new Error("Reranker baseUrl is required. Configure reranker.baseUrl in your codebase-index.json.");
-    }
-    const timeoutMs = this.config.timeoutMs ?? 3e4;
-    const controller = new AbortController();
-    const timeout = setTimeout(() => controller.abort(), timeoutMs);
-    try {
-      const response = await fetch(`${baseUrl}/rerank`, {
-        method: "POST",
-        headers,
-        body: JSON.stringify({
-          model: this.config.model,
-          query,
-          documents,
-          top_n: topN ?? this.config.topN ?? 20,
-          return_documents: false
-        }),
-        signal: controller.signal
-      });
-      clearTimeout(timeout);
-      if (!response.ok) {
-        const errorText = await response.text();
-        throw new Error(`Rerank API error: ${response.status} - ${errorText}`);
-      }
-      const data = await response.json();
-      return {
-        results: data.results.map((r) => ({
-          index: r.index,
-          relevanceScore: r.relevance_score,
-          document: r.document?.text
-        })),
-        tokensUsed: data.meta?.tokens?.input_tokens
-      };
-    } catch (error) {
-      clearTimeout(timeout);
-      if (error instanceof Error && error.name === "AbortError") {
-        throw new Error(`Rerank API request timed out after ${timeoutMs}ms`);
-      }
-      throw error;
-    }
-  }
-};
-
 // src/utils/logger.ts
 var LOG_LEVEL_PRIORITY = {
   error: 0,
@@ -8625,7 +8546,8 @@ function rankHybridResults(query, semanticResults, keywordResults, options) {
       return cached;
     }
   }
-  const overfetchLimit = Math.max(options.limit * 4, options.limit);
+  const overfetchFactor = prioritizeSourcePaths ? 12 : 4;
+  const overfetchLimit = Math.max(options.limit * overfetchFactor, options.limit);
   const fused = options.fusionStrategy === "rrf" ? fuseResultsRrf(semanticResults, keywordResults, options.rrfK, overfetchLimit) : fuseResultsWeighted(semanticResults, keywordResults, options.hybridWeight, overfetchLimit);
   const rerankPoolLimit = Math.max(overfetchLimit, options.rerankTopN * 3, options.limit * 6);
   const rerankPool = fused.slice(0, rerankPoolLimit);
@@ -9011,6 +8933,29 @@ function extractPrimaryIdentifierQueryHint(query) {
   const best = codeTerms.find((term) => term.length >= 6);
   return best ?? null;
 }
+function pathSegmentsForAffinityMatch(filePath) {
+  const normalizedPath2 = normalizeRankingText(filePath).replace(/\\/g, "/");
+  const segments = normalizedPath2.split("/").filter((segment) => segment.length > 0);
+  if (segments.length === 0) {
+    return [];
+  }
+  const basename5 = segments[segments.length - 1] ?? "";
+  const basenameWithoutExt = basename5.replace(/\.[^/.]+$/u, "");
+  const normalizedSegments = segments.map((segment) => segment.toLowerCase());
+  return Array.from(/* @__PURE__ */ new Set([
+    ...normalizedSegments,
+    basenameWithoutExt.toLowerCase()
+  ]));
+}
+function hasModuleAffinity(filePath, exactIdentifierVariants) {
+  const haystack = pathSegmentsForAffinityMatch(filePath);
+  return exactIdentifierVariants.some((variant) => {
+    if (!variant || variant.length < 2) {
+      return false;
+    }
+    return haystack.includes(variant);
+  });
+}
 var FILE_PATH_HINT_EXTENSIONS = [
   "ts",
   "tsx",
@@ -9086,10 +9031,13 @@ function buildDeterministicIdentifierPass(query, candidates, limit, prioritizeSo
   ).map((candidate) => {
     const nameLower = (candidate.metadata.name ?? "").toLowerCase();
     const pathLower = candidate.metadata.filePath.toLowerCase();
-    let maxMatch = 0;
-    const nameMatchesPrimary = primaryVariants.some(
+    const exactIdentifierVariants = primaryVariants.filter((value) => value.length >= 2);
+    const exactMatch = exactIdentifierVariants.some(
       (variant) => nameLower === variant || nameLower.replace(/[^a-z0-9]/g, "") === variant.replace(/[^a-z0-9]/g, "")
     );
+    let maxMatch = 0;
+    const nameMatchesPrimary = exactMatch;
+    const pathAffinity = exactMatch ? hasModuleAffinity(candidate.metadata.filePath, exactIdentifierVariants) : false;
     const pathMatchesFileHint = filePathHint ? pathMatchesHint(candidate.metadata.filePath, filePathHint) : false;
     for (const hint of hints) {
       const variants = normalizeIdentifierVariants(hint);
@@ -9110,12 +9058,17 @@ function buildDeterministicIdentifierPass(query, candidates, limit, prioritizeSo
       candidate,
       maxMatch,
       pathMatchesFileHint,
-      nameMatchesPrimary
+      nameMatchesPrimary,
+      pathAffinity
     };
   }).filter((entry) => entry.maxMatch >= 0.7).sort((a, b) => {
     const aAnchored = a.pathMatchesFileHint && a.nameMatchesPrimary ? 1 : 0;
     const bAnchored = b.pathMatchesFileHint && b.nameMatchesPrimary ? 1 : 0;
     if (aAnchored !== bAnchored) return bAnchored - aAnchored;
+    if (a.nameMatchesPrimary !== b.nameMatchesPrimary) {
+      return b.nameMatchesPrimary ? 1 : -1;
+    }
+    if (a.pathAffinity !== b.pathAffinity) return b.pathAffinity ? 1 : -1;
     if (b.maxMatch !== a.maxMatch) return b.maxMatch - a.maxMatch;
     if (b.candidate.score !== a.candidate.score) return b.candidate.score - a.candidate.score;
     return a.candidate.id.localeCompare(b.candidate.id);
@@ -10105,7 +10058,6 @@ var Indexer = class _Indexer {
   database = null;
   provider = null;
   configuredProviderInfo = null;
-  reranker = null;
   fileHashCache = /* @__PURE__ */ new Map();
   fileHashCachePath = "";
   failedBatchesPath = "";
@@ -10265,7 +10217,6 @@ var Indexer = class _Indexer {
     this.database = null;
     this.provider = null;
     this.configuredProviderInfo = null;
-    this.reranker = null;
     this.indexCompatibility = null;
     this.initializationMode = "none";
     this.readIssues = [];
@@ -11562,15 +11513,6 @@ var Indexer = class _Indexer {
       rerankerEnabled: this.config.reranker?.enabled ?? false
     });
     this.provider = createEmbeddingProvider(this.configuredProviderInfo);
-    if (this.config.reranker?.enabled) {
-      this.reranker = createReranker(this.config.reranker);
-      if (this.reranker.isAvailable()) {
-        this.logger.info("Reranker initialized", {
-          model: this.config.reranker.model,
-          baseUrl: this.config.reranker.baseUrl
-        });
-      }
-    }
     const dimensions = this.configuredProviderInfo.modelInfo.dimensions;
     const storePath = path19.join(this.indexPath, "vectors");
     const vectorMetadataPath = `${storePath}.meta.json`;
@@ -12975,6 +12917,7 @@ var Indexer = class _Indexer {
     const filterByBranch = options?.filterByBranch ?? true;
     const sourceIntent = options?.definitionIntent === true || classifyQueryIntentRaw(query) === "source";
     const identifierHints = extractIdentifierHints(query);
+    const candidateLimit = maxResults * (sourceIntent ? 12 : 4);
     this.logger.search("debug", "Starting search", {
       query,
       maxResults,
@@ -13011,7 +12954,7 @@ var Indexer = class _Indexer {
     const semanticCandidates = embedding ? this.searchSemanticCandidates(
       store,
       embedding,
-      maxResults * 4,
+      candidateLimit,
       branchChunkIds,
       shouldPrefilterByBranch
     ) : [];
@@ -13019,7 +12962,7 @@ var Indexer = class _Indexer {
     const keywordStartTime = import_perf_hooks.performance.now();
     const keywordCandidates = await this.keywordSearch(
       query,
-      maxResults * 4,
+      candidateLimit,
       store,
       invertedIndex,
       branchChunkIds,
@@ -14181,7 +14124,6 @@ var Indexer = class _Indexer {
     this.store = null;
     this.invertedIndex = null;
     this.provider = null;
-    this.reranker = null;
     this.configuredProviderInfo = null;
     this.indexCompatibility = null;
     this.initializationMode = "none";