codebasesearch 0.1.11 → 0.1.13

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "codebasesearch",
-  "version": "0.1.11",
+  "version": "0.1.13",
   "description": "Ultra-simple code search tool with Jina embeddings, LanceDB, and MCP protocol support",
   "type": "module",
   "bin": {

package/src/cli.js

@@ -74,25 +74,26 @@ export async function run(args) {
     // Always reindex to ensure freshness
     console.log('Generating embeddings and indexing...');
 
-    // Generate embeddings in batches
+    // Generate embeddings in batches and upsert immediately to free memory
     const batchSize = 32;
-    const chunkTexts = chunks.map(c => c.content);
-    const allEmbeddings = [];
+    let processedCount = 0;
 
-    for (let i = 0; i < chunkTexts.length; i += batchSize) {
-      const batchTexts = chunkTexts.slice(i, i + batchSize);
+    for (let i = 0; i < chunks.length; i += batchSize) {
+      const batchChunks = chunks.slice(i, i + batchSize);
+      const batchTexts = batchChunks.map(c => c.content);
       const batchEmbeddings = await generateEmbeddings(batchTexts);
-      allEmbeddings.push(...batchEmbeddings);
-    }
 
-    // Create chunks with embeddings
-    const chunksWithEmbeddings = chunks.map((chunk, idx) => ({
-      ...chunk,
-      vector: allEmbeddings[idx]
-    }));
+      // Create batch with embeddings
+      const batchWithEmbeddings = batchChunks.map((chunk, idx) => ({
+        ...chunk,
+        vector: batchEmbeddings[idx]
+      }));
+
+      // Upsert immediately to free memory
+      await upsertChunks(batchWithEmbeddings);
+      processedCount += batchWithEmbeddings.length;
+    }
 
-    // Upsert to store
-    await upsertChunks(chunksWithEmbeddings);
     console.log('Index created\n');
 
     // Execute search

package/src/ignore-parser.js

@@ -1,9 +1,87 @@
 import { readFileSync, existsSync } from 'fs';
-import { join, dirname } from 'path';
+import { join, dirname, extname } from 'path';
 import { fileURLToPath } from 'url';
 
 const __dirname = dirname(fileURLToPath(import.meta.url));
 
+// Whitelist of code file extensions to include
+const CODE_EXTENSIONS = new Set([
+  // JavaScript/TypeScript
+  '.js', '.jsx', '.ts', '.tsx', '.mjs', '.cjs', '.mts', '.cts',
+  // Python
+  '.py', '.pyw', '.pyi',
+  // Java
+  '.java',
+  // C/C++
+  '.c', '.cpp', '.cc', '.cxx', '.h', '.hpp', '.hh', '.hxx',
+  // C#
+  '.cs',
+  // Go
+  '.go',
+  // Rust
+  '.rs',
+  // Ruby
+  '.rb',
+  // PHP
+  '.php', '.phtml',
+  // Swift
+  '.swift',
+  // Kotlin
+  '.kt', '.kts',
+  // Scala
+  '.scala', '.sc',
+  // Perl
+  '.pl', '.pm',
+  // Shell/Bash
+  '.sh', '.bash', '.zsh', '.fish',
+  // PowerShell
+  '.ps1', '.psm1', '.psd1',
+  // Lua
+  '.lua',
+  // R
+  '.r', '.R',
+  // MATLAB/Octave
+  '.m', '.mat',
+  // Julia
+  '.jl',
+  // Dart
+  '.dart',
+  // Elixir
+  '.ex', '.exs',
+  // Erlang
+  '.erl', '.hrl',
+  // Haskell
+  '.hs', '.lhs',
+  // Clojure
+  '.clj', '.cljs', '.cljc',
+  // Lisp
+  '.lisp', '.lsp', '.scm', '.ss', '.rkt',
+  // Fortran
+  '.f', '.for', '.f90', '.f95', '.f03',
+  // Assembly
+  '.asm', '.s', '.S',
+  // Groovy
+  '.groovy', '.gvy',
+  // Visual Basic
+  '.vb', '.vbs',
+  // F#
+  '.fs', '.fsx',
+  // OCaml
+  '.ml', '.mli',
+  // Objective-C
+  '.m', '.mm',
+  // Arduino
+  '.ino',
+  // Vue SFC
+  '.vue',
+  // Svelte
+  '.svelte',
+  // CoffeeScript
+  '.coffee',
+  // Reason
+  '.re', '.rei'
+]);
+
 function loadDefaultIgnores() {
   const ignorePath = join(__dirname, '..', '.thornsignore');
   if (!existsSync(ignorePath)) {
@@ -32,7 +110,97 @@ function getHardcodedIgnores() {
     'temp', 'tmp', '.tmp', '.DS_Store', 'Thumbs.db',
     '.swp', '.swo', '*.swp', '*.swo', '.tern-port',
     'dist-server', 'out-tsc', '.cache', '.parcel-cache',
-    'typings', '.env', '.env.local', '.env.*.local'
+    'typings', '.env', '.env.local', '.env.*.local',
+    // JSON files - PRIMARY PRIORITY for memory reduction
+    '*.json', 'package-lock.json', 'yarn.lock', 'pnpm-lock.yaml',
+    'Gemfile.lock', 'poetry.lock', 'Pipfile.lock',
+    // Lock files
+    '*.lock',
+    // Build outputs
+    'public', 'static', 'site', '_site', '.docusaurus', '.gatsby',
+    // Cache/dependency directories
+    '.rush', '.lerna', '.nx',
+    // IDE/editor configs
+    '.cursor', '.replit', '.sublime-project', '.sublime-workspace',
+    '*.iml', '.project', '.classpath', '.settings', '*.sublime-*',
+    // OS files
+    '.Spotlight-V100', '.Trashes', 'ehthumbs.db', '.fseventsd',
+    '.TemporaryItems', '.AppleDouble', '.LSOverride', 'desktop.ini',
+    // Large data files
+    '*.db', '*.sqlite', '*.sqlite3', '*.bak', '*.dump',
+    '*.backup', '*.data', '*.orig',
+    // Logs and temp
+    '*.log', 'logs', 'npm-debug.log', 'yarn-error.log',
+    // Test coverage and reports
+    'lcov.info', '.coverage', 'test-results',
+    // Database related
+    'storage', 'fixtures',
+    // LLM/Vector related
+    '.llamaindex', '.chroma', '.vectorstore', '.embeddings',
+    '.langchain', '.autogen', '.semantic-kernel', '.openai-cache',
+    '.anthropic-cache', 'embeddings', 'vector-db', 'faiss-index',
+    'chromadb', 'pinecone-cache', 'weaviate-data',
+    // Compiled output
+    '*.min.js', '*.min.css', '*.bundle.js', '*.chunk.js', '*.map',
+    // Generated/build artifacts
+    '.assets', 'out-tsc', 'cmake_build_debug', 'cmake_build_release',
+    // Version managers
+    '.rbenv', '.nvm', '.nvmrc',
+    // Ruby specific
+    '*.gem', '*.rbc', '/pkg', '/spec/reports', '/spec/examples.txt',
+    '/test/tmp', '/test/version_tmp', 'lib/bundler/man', '.ruby-version',
+    // Go specific
+    'go.work',
+    // Rust specific
+    'Cargo.lock', '**/*.rs.bk', '*.pdb',
+    // Java specific
+    '*.class', '*.jar', '*.war', '*.ear', '*.nar', '*.nupkg', '*.snupkg',
+    // C# specific
+    '*.suo', '*.user', '*.userosscache', '*.sln.docstates',
+    'project.lock.json', 'project.fragment.lock.json', 'artifacts',
+    // C/C++ specific
+    '*.o', '*.a', '*.so', '*.exe', '*.obj', '*.dll', '*.dylib',
+    'CMakeFiles', 'CMakeCache.txt', '*.cmake',
+    // Swift/Xcode specific
+    '*.xcodeproj', '*.xcworkspace', '*.moved-aside', '*.pbxuser',
+    '*.mode1v3', '*.mode2v3', '*.perspectivev3',
+    // Scala/SBT specific
+    'lib_managed', 'src_managed', 'project/boot', 'project/plugins/project',
+    '.history', '.lib',
+    // PHP specific
+    'composer.lock', '*.phar',
+    // Docker
+    '.dockerignore', 'docker-compose.override.yml', '.docker',
+    // Documentation build
+    'docs/_build', '.vuepress',
+    // Testing frameworks
+    'jest.config', 'vitest.config', 'pytest.ini', 'tox.ini',
+    '__tests__', '__mocks__', 'spec', 'cypress', 'playwright',
+    // Monorepo workspace patterns (implicit through directory coverage)
+    '.turbo', '.nx',
+    // Python package patterns
+    '*.py[cod]', '*$py.class', '.Python', 'pip-log.txt',
+    'pip-delete-this-directory.txt', '.hypothesis', '.pyre', '.pytype',
+    '*.whl',
+    // Config/metadata that are typically low-value
+    '*.config.js', '*.config.ts', 'webpack.config.js', 'rollup.config.js',
+    'vite.config.js', 'tsconfig.json', 'jsconfig.json', 'babel.config',
+    '.babelrc', '.eslintrc', '.prettierrc', '.stylelintrc', '.editorconfig',
+    '*.local', '*.development', '*.production',
+    // Node specific
+    '.npm', '.node_repl_history', '*.tsbuildinfo', 'yarn-error.log',
+    // Documentation/reference files that don't help with search
+    '*.md', '*.txt', '*.rst', '*.adoc', 'docs', 'documentation', 'wiki',
+    'CHANGELOG', 'HISTORY', 'NEWS', 'UPGRADING', 'FAQ', 'CONTRIBUTING',
+    'SECURITY', 'LICENSE', 'LICENCE', 'COPYRIGHT', 'NOTICE', 'AUTHORS',
+    'THIRDPARTY',
+    // Test and coverage files
+    '*.test', '*.spec', 'test', 'tests', 'htmlcov',
+    // Profiling
+    '*.prof', '*.cpuprofile', '*.heapprofile',
+    // Misc
+    '.tern-port', 'firebase-debug.log', 'firestore-debug.log',
+    'ui-debug.log', '.firebaserc', '.stackdump'
   ]);
 }
 
@@ -107,18 +275,139 @@ export function loadIgnorePatterns(rootPath) {
   return merged;
 }
 
+// Directories to always ignore
+const IGNORED_DIRECTORIES = new Set([
+  // Dependencies - NEVER include
+  'node_modules', 'bower_components', 'jspm_packages', 'web_modules',
+  // Version control
+  '.git', '.svn', '.hg', '.bzr', '.vscode', '.idea', '.vs', '.atom', '.sublime-project',
+  // Build outputs - comprehensive list
+  'dist', 'dist-server', 'dist-ssr', 'dist-client', 'dist-server',
+  'build', 'built', 'Build', 'BUILD',
+  'out', 'output', 'Output', 'OUT', 'release', 'Release', 'RELEASE',
+  'target', 'Target', 'TARGET',
+  'bin', 'Bin', 'BIN', 'obj', 'Obj', 'OBJ',
+  'public', 'static', 'assets', 'www', 'wwwroot',
+  'site', '_site', '.site', '.docusaurus', '.gatsby', '.vuepress',
+  'storybook-static', '.nuxt', 'nuxt', '.next', 'next',
+  'out-tsc', 'tsc', '.tsc',
+  // Cache directories
+  '.cache', 'cache', '.parcel-cache', '.vite', 'vite', '.turbo', 'turbo',
+  '.npm', '.yarn', '.pnp', '.pnpm-store', '.rush', '.lerna', '.nx',
+  // Testing
+  'coverage', '.nyc_output', '.coverage', 'htmlcov', 'test-results',
+  'test', 'tests', 'Test', 'Tests', 'TEST', 'TESTS',
+  '__tests__', '__mocks__', '__snapshots__', '__fixtures__',
+  'cypress', 'playwright', 'e2e', 'integration', 'spec', 'specs',
+  '.tox', '.eggs', '.hypothesis', '.pyre', '.pytype',
+  // Python
+  '__pycache__', '.pytest_cache', '.mypy_cache', '.venv', 'venv', 'env',
+  'env.bak', 'venv.bak', '.Python', 'pip-wheel-metadata', '*.egg-info',
+  // Java/Gradle/Maven
+  '.gradle', '.mvn', 'gradle', 'mvn', '.settings', '.project', '.classpath',
+  // iOS/Android
+  'Pods', 'DerivedData', 'build', '.bundle', 'xcuserdata', '.xcodeproj', '.xcworkspace',
+  // Ruby
+  'vendor', '.bundle', '.ruby-version', 'pkg',
+  // Rust
+  'target', 'Cargo.lock',
+  // Go
+  'vendor', 'Godeps',
+  // PHP
+  'vendor', 'composer',
+  // Infrastructure
+  '.terraform', '.terragrunt-cache', '.pulumi', '.serverless', '.firebase',
+  '.aws', '.azure', '.gcloud', '.vercel', '.netlify', '.now',
+  // Docker
+  '.docker', 'docker', '.dockerignore',
+  // Temp files
+  'temp', 'tmp', '.tmp', '.temp', 'tmpfs', 'scratch', '.scratch',
+  // Documentation
+  'docs', 'doc', 'documentation', 'wiki', 'guides', 'examples', 'demo', 'demos',
+  'CHANGELOG', 'HISTORY', 'NEWS', 'LICENSE', 'LICENCE', 'COPYING', 'AUTHORS',
+  // IDE/Editor
+  '.vs', '.vscode', '.idea', '.eclipse', '.settings', '.classpath', '.project',
+  // Logs
+  'logs', 'log', '*.log',
+  // Data/Storage
+  'storage', 'data', 'database', 'db', 'fixtures', 'seeds',
+  'uploads', 'files', 'media', 'resources', 'assets', 'images', 'img',
+  // LLM/AI
+  '.llamaindex', '.chroma', '.vectorstore', '.embeddings',
+  '.langchain', '.autogen', '.semantic-kernel', '.openai-cache',
+  '.anthropic-cache', 'embeddings', 'vector-db', 'faiss-index',
+  'chromadb', 'pinecone-cache', 'weaviate-data',
+  // Package managers
+  '.yarn', '.pnpm', '.npm', '.bun',
+  // Compiled outputs
+  'typings', 'types', '@types', 'type-definitions',
+  // Misc
+  'public', 'static', 'site', '_site',
+  'cmake_build_debug', 'cmake_build_release', 'CMakeFiles', 'CMakeCache.txt',
+  'out-tsc', 'dist-server', 'server', 'client', 'browser', 'esm', 'cjs', 'umd', 'lib', 'es'
+]);
+
+export function isCodeFile(filePath) {
+  const normalizedPath = filePath.replace(/\\/g, '/');
+  const pathParts = normalizedPath.split('/');
+  const fileName = pathParts[pathParts.length - 1];
+  
+  // Get file extension
+  const lastDotIndex = fileName.lastIndexOf('.');
+  if (lastDotIndex === -1 || lastDotIndex === 0) {
+    return false; // No extension or hidden file without extension
+  }
+  
+  const ext = fileName.slice(lastDotIndex).toLowerCase();
+  return CODE_EXTENSIONS.has(ext);
+}
+
+export function shouldIgnoreDirectory(dirPath) {
+  const normalizedPath = dirPath.replace(/\\/g, '/');
+  const pathParts = normalizedPath.split('/');
+  
+  for (const part of pathParts) {
+    if (IGNORED_DIRECTORIES.has(part)) {
+      return true;
+    }
+  }
+  
+  return false;
+}
+
 export function shouldIgnore(filePath, ignorePatterns) {
   const normalizedPath = filePath.replace(/\\/g, '/');
   const pathParts = normalizedPath.split('/');
+  const fileName = pathParts[pathParts.length - 1];
+  
+  // Check if any directory in path should be ignored
+  for (const part of pathParts.slice(0, -1)) {
+    if (IGNORED_DIRECTORIES.has(part)) {
+      return true;
+    }
+  }
+  
+  // Check if it's a code file using whitelist
+  if (!isCodeFile(filePath)) {
+    return true;
+  }
 
+  // Check against additional ignore patterns
   for (const pattern of ignorePatterns) {
+    // Handle path patterns (contain /)
     if (pattern.includes('/')) {
       if (normalizedPath.includes(pattern)) {
         return true;
       }
-    } else {
+    }
+    // Handle exact file name patterns
+    else if (fileName === pattern) {
+      return true;
+    }
+    // Handle directory name patterns (match any path part)
+    else {
       for (const part of pathParts) {
-        if (part === pattern) {
+        if (part === pattern || part.startsWith(pattern + '/')) {
           return true;
         }
       }

package/src/scanner.js

@@ -23,7 +23,6 @@ const SUPPORTED_EXTENSIONS = new Set([
   '.groovy',
   '.gradle',
   '.xml', '.xsd',
-  '.json', '.jsonc',
   '.yaml', '.yml',
   '.toml',
   '.html', '.htm',
@@ -95,7 +94,7 @@ function walkDirectory(dirPath, ignorePatterns, relativePath = '') {
   return files;
 }
 
-function chunkContent(content, chunkSize = 1000, overlapSize = 200) {
+function chunkContent(content, chunkSize = 1000, overlapSize = 100) {
   const lines = content.split('\n');
   const chunks = [];
 

package/src/store.js

@@ -4,6 +4,7 @@ import { mkdirSync, existsSync } from 'fs';
 
 let dbConnection = null;
 let tableRef = null;
+let isFirstBatch = true;
 
 export async function initStore(dbPath) {
   // Ensure directory exists
@@ -19,6 +20,7 @@ export async function initStore(dbPath) {
       uri: dbDir,
       mode: 'overwrite'
     });
+    isFirstBatch = true;
     console.error('Vector store initialized');
     return true;
   } catch (e) {
@@ -68,14 +70,24 @@ export async function upsertChunks(chunks) {
   try {
     let table = null;
 
-    // Try to open existing table
-    try {
-      table = await dbConnection.openTable(tableName);
-      // Overwrite existing table with new data
-      await table.overwrite(data);
-    } catch (e) {
-      // Table doesn't exist, create new one
-      table = await dbConnection.createTable(tableName, data);
+    if (isFirstBatch) {
+      // First batch: try to open existing table, or create new one
+      try {
+        table = await dbConnection.openTable(tableName);
+        await table.overwrite(data);
+      } catch (e) {
+        table = await dbConnection.createTable(tableName, data);
+      }
+      isFirstBatch = false;
+    } else {
+      // Subsequent batches: add to existing table
+      try {
+        table = await dbConnection.openTable(tableName);
+        await table.add(data);
+      } catch (e) {
+        console.error('Failed to add to table:', e.message);
+        throw e;
+      }
     }
 
     tableRef = table;