@ulpi/codemap 0.3.0 → 0.3.2

package/README.md

@@ -33,7 +33,7 @@ codemap search "authentication middleware"
 |----------|--------|------|------|----------|
 | **Ollama** | `nomic-embed-text` (768), `mxbai-embed-large` (1024), `all-minilm` (384) | varies | Free | Ollama running locally |
 | **OpenAI** | `text-embedding-3-small` (1536), `text-embedding-3-large` (3072) | varies | Paid | `OPENAI_API_KEY` |
-| **ULPI Cloud** | `code` (1024) | 1024 | Managed | `ULPI_API_KEY` |
+| **ULPI Cloud** | `qwen3-embedding:4b` (2048) | 2048 | Managed | `ULPI_API_KEY` |
 
 `codemap init` walks you through provider and model selection interactively.
 

package/dist/index.js

@@ -4098,6 +4098,59 @@ function extractSymbolsFromChunks(chunks) {
   }
   return allSymbols;
 }
+var TAG_TYPE_MAP = {
+  class: "class",
+  function: "function",
+  method: "method",
+  interface: "interface",
+  type: "type",
+  enum: "enum",
+  module: "class",
+  trait: "interface",
+  struct: "class",
+  property: "variable",
+  constant: "const",
+  variable: "variable",
+  mixin: "class",
+  extension: "class",
+  union: "enum"
+};
+function symbolsFromTags(allTags, chunks) {
+  const chunksByFile = /* @__PURE__ */ new Map();
+  for (const chunk of chunks) {
+    let arr = chunksByFile.get(chunk.filePath);
+    if (!arr) {
+      arr = [];
+      chunksByFile.set(chunk.filePath, arr);
+    }
+    arr.push(chunk);
+  }
+  const seen = /* @__PURE__ */ new Set();
+  const symbols = [];
+  for (const [filePath, tags] of allTags) {
+    const fileChunks = chunksByFile.get(filePath);
+    for (const tag of tags) {
+      if (tag.kind !== "def") continue;
+      if (!tag.name || tag.name.length < 2) continue;
+      const symbolType = TAG_TYPE_MAP[tag.symbolType ?? ""] ?? "unknown";
+      const key = `${tag.name}:${filePath}:${tag.line}`;
+      if (seen.has(key)) continue;
+      seen.add(key);
+      const chunk = fileChunks?.find(
+        (c2) => tag.line >= c2.startLine && tag.line <= c2.endLine
+      );
+      symbols.push({
+        name: tag.name,
+        symbolType,
+        filePath,
+        startLine: tag.line,
+        endLine: chunk?.endLine ?? tag.line,
+        chunkId: chunk?.id ?? ""
+      });
+    }
+  }
+  return symbols;
+}
 function searchSymbols(symbols, query, topK = 20) {
   const q = query.toLowerCase();
   const results = [];
@@ -4512,11 +4565,13 @@ async function runInitPipelineInner(projectDir, config, branch, onProgress, data
   await store.dropTable();
   await embedAndStore(embedder, store, allChunks, onProgress);
   onProgress?.({ phase: "symbols", message: "Extracting symbols..." });
-  const symbols = extractSymbolsFromChunks(allChunks);
+  const regexSymbols = extractSymbolsFromChunks(allChunks);
+  const tagSymbols = allTags.size > 0 ? symbolsFromTags(allTags, allChunks) : [];
+  const symbols = mergeSymbols(regexSymbols, tagSymbols);
   saveSymbolIndex(metadataDir, symbols);
   onProgress?.({
     phase: "symbols",
-    message: `Extracted ${symbols.length} symbols`,
+    message: `Extracted ${symbols.length} symbols (${tagSymbols.length} from tags, ${regexSymbols.length} from regex)`,
     current: symbols.length,
     total: symbols.length
   });
@@ -4631,7 +4686,10 @@ async function runIncrementalInner(projectDir, config, branch, manifest, changed
     saveBM25Index(metadataDir, bm25Index);
     const allFilePaths2 = scannedFiles.map((f) => f.filePath);
     const allChunks = chunkFiles(projectDir, allFilePaths2, config.chunking);
-    const symbols = extractSymbolsFromChunks(allChunks);
+    const regexSymbols = extractSymbolsFromChunks(allChunks);
+    const allTags2 = await extractAllTags(projectDir, allFilePaths2, onProgress);
+    const tagSymbols = allTags2.size > 0 ? symbolsFromTags(allTags2, allChunks) : [];
+    const symbols = mergeSymbols(regexSymbols, tagSymbols);
     saveSymbolIndex(metadataDir, symbols);
   } else {
     const bm25Index = loadBM25Index(metadataDir);
@@ -4703,6 +4761,23 @@ function buildEmbeddingText(chunk) {
   return `${header}
 ${chunk.content}`;
 }
+function mergeSymbols(regexSymbols, tagSymbols) {
+  const seen = /* @__PURE__ */ new Set();
+  const merged = [];
+  for (const sym of tagSymbols) {
+    const key = `${sym.name}:${sym.filePath}:${sym.startLine}`;
+    if (seen.has(key)) continue;
+    seen.add(key);
+    merged.push(sym);
+  }
+  for (const sym of regexSymbols) {
+    const key = `${sym.name}:${sym.filePath}:${sym.startLine}`;
+    if (seen.has(key)) continue;
+    seen.add(key);
+    merged.push(sym);
+  }
+  return merged;
+}
 async function extractAllTags(projectDir, filePaths, onProgress) {
   const allTags = /* @__PURE__ */ new Map();
   try {
@@ -5000,6 +5075,7 @@ async function searchCode(projectDir, query, options) {
     const hybridConfig = config.hybrid;
     if (hybridConfig.enabled) {
       return await hybridSearch(
+        projectDir,
         store,
         metadataDir,
         queryVector,
@@ -5027,7 +5103,7 @@ async function searchCode(projectDir, query, options) {
     await store.close();
   }
 }
-async function hybridSearch(store, metadataDir, queryVector, query, weights, vectorK, bm25K, limit, threshold, pathPrefix, includeTests, includeDocs) {
+async function hybridSearch(projectDir, store, metadataDir, queryVector, query, weights, vectorK, bm25K, limit, threshold, pathPrefix, includeTests, includeDocs) {
   const vectorResults = await store.query(queryVector, vectorK);
   const bm25Index = loadBM25Index(metadataDir);
   const bm25Results = bm25Index ? queryBM25(bm25Index, query, bm25K) : [];
@@ -5051,6 +5127,7 @@ async function hybridSearch(store, metadataDir, queryVector, query, weights, vec
     weights,
     Math.max(limit * 3, 80)
   );
+  enrichBM25Candidates(projectDir, fused);
   return applyFilters(fused, limit, threshold, pathPrefix, includeTests, includeDocs);
 }
 async function vectorOnlySearch(store, queryVector, limit, threshold, pathPrefix, includeTests, includeDocs) {
@@ -5089,6 +5166,28 @@ function applyFilters(candidates, limit, threshold, pathPrefix, includeTests, in
     symbolScore: c2.symbolScore
   }));
 }
+function enrichBM25Candidates(projectDir, candidates) {
+  for (const c2 of candidates) {
+    if (c2.snippet !== "" || c2.startLine !== 0) continue;
+    const lastColon = c2.id.lastIndexOf(":");
+    if (lastColon === -1) continue;
+    const beforeHash = c2.id.slice(0, lastColon);
+    const secondLastColon = beforeHash.lastIndexOf(":");
+    if (secondLastColon === -1) continue;
+    const lineRange = beforeHash.slice(secondLastColon + 1);
+    const match = lineRange.match(/^(\d+)-(\d+)$/);
+    if (!match) continue;
+    c2.startLine = parseInt(match[1], 10);
+    c2.endLine = parseInt(match[2], 10);
+    try {
+      const absPath = path10.join(projectDir, c2.filePath);
+      const content = fs10.readFileSync(absPath, "utf-8");
+      const lines = content.split("\n");
+      c2.snippet = lines.slice(c2.startLine - 1, c2.endLine).join("\n").slice(0, 200);
+    } catch {
+    }
+  }
+}
 function loadGraphRanks(metadataDir) {
   try {
     const file = path10.join(metadataDir, "depgraph", "pagerank.json");
@@ -6038,7 +6137,7 @@ function loadMetricsSafe(projectDir, branch) {
 function createCodemapMcpServer() {
   const server = new McpServer({
     name: "codemap",
-    version: "0.3.0"
+    version: "0.3.2"
   });
   const dataDir = getDataDir2();
   server.tool(
@@ -6614,7 +6713,7 @@ var __dirname = path13.dirname(fileURLToPath(import.meta.url));
 var grammarsDir = path13.join(__dirname, "grammars");
 setGrammarDir(grammarsDir);
 var program = new Command();
-program.name("codemap").description("Code intelligence CLI \u2014 hybrid search, dependency analysis, PageRank").version("0.3.0").option("--cwd <dir>", "Project directory (default: cwd)");
+program.name("codemap").description("Code intelligence CLI \u2014 hybrid search, dependency analysis, PageRank").version("0.3.2").option("--cwd <dir>", "Project directory (default: cwd)");
 registerSearch(program);
 registerSymbols(program);
 registerIndex(program);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ulpi/codemap",
-  "version": "0.3.0",
+  "version": "0.3.2",
   "type": "module",
   "description": "Standalone code intelligence CLI — hybrid vector + BM25 search, dependency analysis, PageRank",
   "bin": {