Conversation
WalkthroughAdds linting/formatting tooling (ESLint/Prettier) and scripts, updates docs, and reformats source/tests. Introduces VectorStore lifecycle methods (clear, close). Updates runtime flow to handle file-watcher ready event and to skip indexing files exceeding a configured size. Minor logger rotation safety tweak; extensive integration test restructuring. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor User as User/CLI
participant App as App (index.ts)
participant FW as FileWatcher
participant Parser as CodeParser
participant Store as CodeStorage
participant VS as VectorStore
participant SS as SemanticSearch
User->>App: start()
App->>FW: create/start watcher
FW-->>App: ready
Note over App,FW: New: 'ready' event logs readiness and file count
FW-->>App: add/change(file)
App->>App: check file size vs max
alt exceeds limit
App-->>App: skip indexing
App->>User: warn (skipped due to size)
else within limit
App->>Parser: parse(file)
Parser-->>App: chunks
App->>Store: persist(chunks)
Store-->>App: ok
App->>SS: embed+index(chunks)
SS->>VS: upsert embeddings
VS-->>SS: ok
SS-->>App: ok
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~30 minutes Poem
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 10
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (8)
src/code-storage.ts (1)
141-149: Avoid blocking I/O inside an async method. Use fs.promises.writeFile instead of writeFileSync.persistChunkToDisk is async but uses synchronous disk write, which blocks the event loop and can hurt throughput.
Apply this diff:
private async persistChunkToDisk(chunk: CodeChunk): Promise<void> { try { - const fs = require('fs'); + const fs = require('fs').promises; const path = require('path'); const chunkFile = path.join(this.options.storagePath, `${chunk.id}.json`); - fs.writeFileSync(chunkFile, JSON.stringify(chunk)); + await fs.writeFile(chunkFile, JSON.stringify(chunk)); } catch (error) { console.error('Failed to persist chunk to disk:', error); } }tsconfig.json (1)
3-5: Preserve native dynamic import for ESM-only packages (fix tsconfig).
- Found: TypeScript ^5.3.3; tsconfig.json sets "module": "CommonJS"; src/embedding-service.ts:67 uses
await import('@xenova/transformers').- Risk: tsc will downlevel
import()to arequire(...)wrapper when emitting CommonJS, causing ERR_REQUIRE_ESM for ESM-only packages. (github.com)- Node supports dynamic
import()from CommonJS at runtime, so the correct fix is to emit nativeimport()instead of a require wrapper. (nodejs.org)- Action: change tsconfig to emit ESM-aware code (e.g. set
"module": "NodeNext"or"module": "ESNext"and pair with an appropriate"moduleResolution"or use a bundler that preserves dynamic import), rebuild and test the import in src/embedding-service.ts:67. (typescriptlang.org)src/code-parser.ts (1)
849-859: Bug: parseFileIncremental returns non-nullabletreebut may passundefined.
oldTree!can still beundefined. Maketreeoptional or guarantee a valid tree object.Apply:
- ): Promise<{ chunks: CodeChunk[]; tree: ParserType.Tree }> { + ): Promise<{ chunks: CodeChunk[]; tree?: ParserType.Tree }> { @@ - if (!language) { - return { chunks: [], tree: oldTree! }; - } + if (!language) { + return { chunks: [], tree: oldTree }; + } @@ - if (!parser) { - return { chunks: [], tree: oldTree! }; - } + if (!parser) { + return { chunks: [], tree: oldTree }; + } @@ - } catch (error) { + } catch (error) { console.error(`Failed to parse file incrementally ${filePath}:`, error); - return { chunks: [], tree: oldTree! }; + return { chunks: [], tree: oldTree }; }README.md (1)
192-199: Fix incorrect directory name after clone.Repository is
42context, but docs saycd mcp-local-context-engine.Apply:
- git clone https://github.com/AssahBismarkabah/42context.git - cd mcp-local-context-engine + git clone https://github.com/AssahBismarkabah/42context.git + cd 42contextsrc/vector-store.ts (1)
33-40: Type name collision with project-wide VectorSearchResult.This file defines and exports a VectorSearchResult incompatible with src/types.ts’ VectorSearchResult. This will cause confusion and type bugs.
Apply this diff to rename and stop exporting the conflicting type:
-interface VectorSearchResult { +interface InternalVectorSearchResult { chunkId: string; similarity: number; metadata: { chunk: CodeChunk; embedding: EmbeddingResult; }; } @@ -export type { VectorStoreOptions, VectorSearchResult }; +export type { VectorStoreOptions };And update usages:
- ): Promise<VectorSearchResult[]> { + ): Promise<InternalVectorSearchResult[]> {Also applies to: 418-419
src/index.ts (2)
159-162: Size check compares characters, not bytescontent.length counts UTF‑16 code units, not file bytes. This can under‑ or over‑count vs a byte limit.
Use byte length:
- if (content.length > config.security.maxFileSize) { + const bytes = Buffer.byteLength(content, 'utf8'); + if (bytes > config.security.maxFileSize) { this.logger.warn(`File too large, skipping: ${filePath}`); return; }
189-191: Same byte/char size issue on change pathMirror the fix for changed files.
- if (content.length > config.security.maxFileSize) { + const bytes = Buffer.byteLength(content, 'utf8'); + if (bytes > config.security.maxFileSize) { this.logger.warn(`File too large, skipping re-index: ${filePath}`); return; }src/logger.ts (1)
128-141: Log rotation may lose data; end() is asyncYou call end() then renameSync immediately; writes might not be flushed.
Ensure rotation happens after the stream finishes:
- try { - this.fileStream.end(); - - const timestamp = new Date().toISOString().replace(/[:.]/g, '-'); - const rotatedFile = `${this.logFile}.${timestamp}`; - - require('fs').renameSync(this.logFile, rotatedFile); - - // Clean up old rotated files - this.cleanupOldLogs(); - - // Create new log file - this.fileStream = createWriteStream(this.logFile, { flags: 'a' }); - } catch (error) { + try { + const current = this.fileStream; + current.end(() => { + const timestamp = new Date().toISOString().replace(/[:.]/g, '-'); + const rotatedFile = `${this.logFile!}.${timestamp}`; + try { + require('fs').renameSync(this.logFile!, rotatedFile); + this.cleanupOldLogs(); + } catch (e) { + console.error('Failed during log rotation:', e); + } + this.fileStream = createWriteStream(this.logFile!, { flags: 'a' }); + }); + } catch (error) { console.error('Failed to rotate log file:', error); }
🧹 Nitpick comments (37)
src/code-storage.ts (2)
156-172: Minor: tighten candidate filtering with Set.has to cut O(n·m) includes scans.Small perf win without behavior change.
Apply this diff:
- if (query.filePath) { - const fileChunks = this.fileIndex.get(query.filePath) || []; - candidateIds = new Set([...candidateIds].filter(id => fileChunks.includes(id))); - } + if (query.filePath) { + const fileChunks = new Set(this.fileIndex.get(query.filePath) || []); + candidateIds = new Set([...candidateIds].filter(id => fileChunks.has(id))); + } - if (query.language) { - const languageChunks = this.languageIndex.get(query.language) || []; - candidateIds = new Set([...candidateIds].filter(id => languageChunks.includes(id))); - } + if (query.language) { + const languageChunks = new Set(this.languageIndex.get(query.language) || []); + candidateIds = new Set([...candidateIds].filter(id => languageChunks.has(id))); + }
174-178: Apply the same Set.has optimization for type filter.- if (query.chunkType) { - const typeChunks = this.typeIndex.get(query.chunkType) || []; - candidateIds = new Set([...candidateIds].filter(id => typeChunks.includes(id))); - } + if (query.chunkType) { + const typeChunks = new Set(this.typeIndex.get(query.chunkType) || []); + candidateIds = new Set([...candidateIds].filter(id => typeChunks.has(id))); + }test/test-embedding.ts (1)
146-146: Avoid CommonJS export in a TS module; prefer ES export for consistency.Use a named export to align with the rest of the file’s ES imports.
-module.exports = { testEmbeddingService }; +export { testEmbeddingService };src/embedding-service.ts (5)
53-55: Device option note.'gpu' requires a WebGPU-capable runtime for @xenova/transformers; in Node this may fall back to CPU. Consider surfacing actual device used in getModelInfo().
111-112: Model name vs. dimension coupling.getModelInfo/getEmbeddingDimension hardcode 384; if modelName changes, this drifts. Suggest deriving from output shape once and caching.
- getModelInfo(): { + getModelInfo(): { name: string; dimension: number | null; maxSequenceLength: number; device: string; } { - return { - name: this.options.modelName, - dimension: this.isModelLoaded ? 384 : null, // all-MiniLM-L6-v2 outputs 384 dimensions + return { + name: this.options.modelName, + dimension: this.isModelLoaded ? this.getEmbeddingDimension() : null, maxSequenceLength: this.options.maxSequenceLength, device: this.options.device, }; } @@ - getEmbeddingDimension(): number { - return 384; // all-MiniLM-L6-v2 outputs 384 dimensions - } + getEmbeddingDimension(): number { + // Prefer cached inference from a tiny probe string if not known + if (this.embeddingCache.size > 0) { + return this.embeddingCache.values().next().value.dimension; + } + return 384; + }
135-137: Batch logging: keep logs at debug to reduce noise.Consider gating verbose logs behind an env flag.
160-166: Error logging is good; consider attaching chunk id/context.You already include the id; optionally include model name to aid triage.
306-306: Expose actual device used.If transformers chooses a backend automatically, consider capturing it after initialize() and returning it here.
test/test-parser.ts (3)
7-14: Silent catch blocks.The empty catch swallows unexpected errors; at least log a brief message in tests to aid debugging.
-} catch {} +} catch (e) { + console.warn('Setup cleanup warning:', e instanceof Error ? e.message : String(e)); +}
96-99: Prefer consistent module style.Mixing require() and ESM import in TS can confuse linters. Consider converting the require to an import if possible.
153-156: Teardown silent catch.Same as setup: log teardown errors to avoid hiding filesystem issues.
.eslintrc.json (1)
31-31: Reconsider ignoring all “*.js”.If you add any JS sources (CLI shims, config scripts), this will silently exclude them from linting. Consider removing
"*.js"here and rely ondist/ignore instead.src/code-parser.ts (2)
96-103: Replace console logging with an injectable logger or debug flag.Library code spamming
console.*is hard to control in consumers.Also applies to: 206-211
918-921: Remove misleading CommonJS type exports.
module.exports.CodeParserOptions/CodeChunk/ASTNodeInfo = {}export empty objects not tied to real runtime values and can confuse users.Apply:
-module.exports.CodeParserOptions = {}; -module.exports.CodeChunk = {}; -module.exports.ASTNodeInfo = {};src/config.ts (2)
330-335: Guard env parsing against NaN.Current
parseIntcan yieldNaN, silently corrupting config.Apply:
- if (env.DEV_CONTEXT_MAX_CONCURRENT_OPERATIONS) { - config.performance.maxConcurrentOperations = parseInt( - env.DEV_CONTEXT_MAX_CONCURRENT_OPERATIONS - ); - } + if (env.DEV_CONTEXT_MAX_CONCURRENT_OPERATIONS) { + const v = Number(env.DEV_CONTEXT_MAX_CONCURRENT_OPERATIONS); + if (Number.isFinite(v) && v > 0) { + config.performance.maxConcurrentOperations = v; + } + }
134-146: Align parser.supportedLanguages with actual SupportedLanguage.Defaults include languages not supported by
CodeParser(e.g., ruby, php). Either add support or constrain defaults to the union type to avoid confusion.package.json (1)
16-16: Make clean cross‑platform.Use rimraf to work on Windows as well.
Apply:
- "clean": "rm -rf dist", + "clean": "rimraf dist",And add to devDependencies:
+ "rimraf": "^6.0.0",src/cli.ts (5)
445-474: Wrap switch case declarations in blocks (noSwitchDeclarations).Prevents scope bleed for
const/letinside cases.Apply:
private async configCommand(action: string, key?: string, value?: string): Promise<void> { switch (action) { case 'get': - if (!key) { - console.error('Please specify a configuration key'); - return; - } - const configValue = this.configManager.getConfig(); - console.log(`Configuration: ${key} = ${JSON.stringify(configValue, null, 2)}`); - break; + { + if (!key) { + console.error('Please specify a configuration key'); + return; + } + const configValue = this.configManager.getConfig(); + console.log(`Configuration: ${key} = ${JSON.stringify(configValue, null, 2)}`); + break; + } case 'set': - if (!key || !value) { - console.error('Please specify both key and value'); - return; - } - console.log(`Setting ${key} = ${value}`); - console.log('Config set implementation pending...'); - break; + { + if (!key || !value) { + console.error('Please specify both key and value'); + return; + } + console.log(`Setting ${key} = ${value}`); + console.log('Config set implementation pending...'); + break; + } case 'list': - console.log('Current configuration:'); - console.log(JSON.stringify(this.configManager.getConfig(), null, 2)); - break; + { + console.log('Current configuration:'); + console.log(JSON.stringify(this.configManager.getConfig(), null, 2)); + break; + } case 'reset': - console.log('Resetting configuration to defaults...'); - console.log('Config reset implementation pending...'); - break; + { + console.log('Resetting configuration to defaults...'); + console.log('Config reset implementation pending...'); + break; + } default: console.error(`Unknown action: ${action}`); console.log('Available actions: get, set, list, reset'); } }
606-676: Wrap switch case declarations in blocks (noSwitchDeclarations).Same issue in
debugCommandcases.Apply:
private async debugCommand(action: string, file?: string): Promise<void> { switch (action) { case 'parse': - if (!file) { - console.error('Please specify a file to parse'); - return; - } - if (!existsSync(file)) { - console.error(`File not found: ${file}`); - return; - } - console.log(`Parsing file: ${file}`); - const parser = new CodeParser(); - const content = readFileSync(file, 'utf-8'); - const language = parser.detectLanguage(file); - if (!language) { - console.error(`Unsupported file type: ${file}`); - return; - } - console.log(`Language: ${language}`); - const chunks = await parser.parseFile(file, content); - console.log(`Found ${chunks.length} code chunks:`); - chunks.forEach((chunk, index) => { - console.log( - ` ${index + 1}. ${chunk.type}: ${chunk.name} (${chunk.startLine}-${chunk.endLine})` - ); - }); - break; + { + if (!file) { + console.error('Please specify a file to parse'); + return; + } + if (!existsSync(file)) { + console.error(`File not found: ${file}`); + return; + } + console.log(`Parsing file: ${file}`); + const parser = new CodeParser(); + const content = readFileSync(file, 'utf-8'); + const language = parser.detectLanguage(file); + if (!language) { + console.error(`Unsupported file type: ${file}`); + return; + } + console.log(`Language: ${language}`); + const chunks = await parser.parseFile(file, content); + console.log(`Found ${chunks.length} code chunks:`); + chunks.forEach((chunk, index) => { + console.log( + ` ${index + 1}. ${chunk.type}: ${chunk.name} (${chunk.startLine}-${chunk.endLine})` + ); + }); + break; + } case 'embed': - if (!file) { - console.error('Please specify a file to embed'); - return; - } - console.log(`Generating embeddings for: ${file}`); - const embeddingService = new EmbeddingService(); - await embeddingService.initialize(); - console.log('Embedding debug implementation pending...'); - break; + { + if (!file) { + console.error('Please specify a file to embed'); + return; + } + console.log(`Generating embeddings for: ${file}`); + const embeddingService = new EmbeddingService(); + await embeddingService.initialize(); + console.log('Embedding debug implementation pending...'); + break; + } case 'test-connection': - console.log('Testing connections...'); - try { - const config = this.configManager.getConfig(); - const vectorStore = new ChromaVectorStore( - config.vectorStore.collectionName, - config.vectorStore.host || 'localhost', - config.vectorStore.port || 8000, - config.vectorStore.authToken || 'test-token' - ); - await vectorStore.initialize(); - const stats = await vectorStore.getCollectionStats(); - console.log(`✓ ChromaDB connection successful (${stats.count} vectors)`); - await vectorStore.close(); - } catch (error) { - console.error('✗ ChromaDB connection failed:', error); - } - break; + { + console.log('Testing connections...'); + try { + const config = this.configManager.getConfig(); + const vectorStore = new ChromaVectorStore( + config.vectorStore.collectionName, + config.vectorStore.host || 'localhost', + config.vectorStore.port || 8000, + config.vectorStore.authToken || 'test-token' + ); + await vectorStore.initialize(); + const stats = await vectorStore.getCollectionStats(); + console.log(`✓ ChromaDB connection successful (${stats.count} vectors)`); + await vectorStore.close(); + } catch (error) { + console.error('✗ ChromaDB connection failed:', error); + } + break; + } default: console.error(`Unknown debug action: ${action}`); console.log('Available actions: parse, embed, test-connection'); } }
210-216: Harden number parsing with defaults.Avoid
NaNpropagation from CLI args.Apply:
- const searchOptions = { - topK: parseInt(options.topK), + const topK = Number.parseInt(options.topK, 10); + const minSim = Number.parseFloat(options.minSimilarity); + const searchOptions = { + topK: Number.isFinite(topK) ? topK : 5, language: options.language, chunkType: options.type, filePath: options.file, - minSimilarity: parseFloat(options.minSimilarity), + minSimilarity: Number.isFinite(minSim) ? minSim : 0.2, };
592-594: “Embedding cache cleared” is misleading; nothing is cleared.Either wire through an actual cache instance to clear or adjust the message/flag until implemented.
451-453: Config “get” ignores key path and dumps full config.Implement dot‑path resolution (e.g.,
vectorStore.host) so the command matches README examples.test/test-vector-store.ts (2)
195-195: Avoid mixing CommonJS in a TypeScript module.Prefer ESM export to align with your TS imports and ESLint/Prettier setup.
Apply this diff:
-module.exports = { testVectorStore }; +export { testVectorStore };
75-81: Env var naming: standardize the auth token name.This file uses CHROMA_AUTH_TOKEN while SemanticSearch uses CHROMA_SERVER_AUTH_CREDENTIALS. Pick one to avoid surprise misconfigurations.
Also applies to: 178-181
src/vector-store.ts (2)
315-341: Tighten types: avoidanyfor embeddingService.Use the project’s EmbeddingService interface for compile-time safety.
Apply this diff:
-import { EmbeddingResult } from './embedding-service'; +import { EmbeddingResult, EmbeddingService as IEmbeddingService } from './embedding-service'; @@ - async searchByText( + async searchByText( queryText: string, - embeddingService: any, // EmbeddingService + embeddingService: IEmbeddingService,
90-97: Chroma metadata: 'dimension' is likely ignored.Chroma sets dimension from first insert; storing it in metadata doesn’t enforce schema. Keep it only if you explicitly read it elsewhere.
Also applies to: 374-384
test/test-storage.ts (2)
148-148: Avoid CommonJS export in TS.Use ESM export for consistency.
Apply this diff:
-module.exports = { testCodeStorage }; +export { testCodeStorage };
62-141: This is a script, not an assertion-based test.Consider moving under a scripts/ or examples/ folder or convert to node:test with assertions so it doesn’t accidentally run in CI.
src/semantic-search.ts (2)
36-39: Env var inconsistency with tests.This file uses CHROMA_SERVER_AUTH_CREDENTIALS, tests use CHROMA_AUTH_TOKEN. Standardize to one name and document it.
122-126: Preserve explicit zero values in config defaults.Using
||overrides valid 0. Use??instead.Apply this diff:
- const searchOptions = { - topK: this.config.maxResults || 10, - minSimilarity: this.config.minSimilarity || 0.7, - ...options, - }; + const searchOptions = { + topK: this.config.maxResults ?? 10, + minSimilarity: this.config.minSimilarity ?? 0.7, + ...options, + };src/chroma-vector-store.ts (1)
83-101: Include distances/documents explicitly in query.Some client versions require
includeto return distances/documents reliably.Apply this diff:
-import { ChromaClient, Collection } from 'chromadb'; +import { ChromaClient, Collection, IncludeEnum } from 'chromadb'; @@ - const queryParams: any = { + const queryParams: any = { queryEmbeddings: [queryVector], - nResults: topK, + nResults: topK, + include: [IncludeEnum.Metadatas, IncludeEnum.Documents, IncludeEnum.Distances], };Also applies to: 104-106
test/debug-watcher.ts (1)
51-53: Await watcher readiness instead of sleeping.Waiting on the 'ready' event is more robust than a fixed timeout.
Apply this diff:
+import { once } from 'node:events'; @@ - // Wait for ready - await new Promise(resolve => setTimeout(resolve, 1000)); + // Wait for ready + await once(watcher as any, 'ready');Also applies to: 43-45
src/index.ts (2)
209-214: Unnecessary language field in delete eventhandleFileChange('delete') doesn’t require language; passing 'unknown' may fight stricter type‑guards downstream.
await this.semanticSearch.handleFileChange({ type: 'delete', - filePath, - language: 'unknown', + filePath, });
80-80: Redundant option (nit)persistent defaults to true in FileWatcher; you can omit it for brevity.
test/test-integration.ts (3)
66-69: Handle both add and change eventsNew files may emit add; listening to both removes the need to manually process files created after start.
- fileWatcher.on('fileChanged', handleFileChange); + fileWatcher.on('fileAdded', handleFileChange); + fileWatcher.on('fileChanged', handleFileChange);
109-111: Avoid arbitrary sleeps in testshandleFileChange is awaited; extra timeouts slow CI without benefit.
- // Wait for processing - await new Promise(resolve => setTimeout(resolve, 2000)); + // No sleep needed; processing already awaitedRepeat the same removal after the update.
Also applies to: 181-182
208-212: Duplicate “Test 4” log linesThere are two identical headings back‑to‑back.
- // Test 4: Find similar patterns - console.log('\n7. Test 4: Finding similar patterns...'); - - // Test 4: Find similar patterns (already implemented above) - console.log('\n7. Test 4: Finding similar patterns...'); + // Test 4: Find similar patterns + console.log('\n7. Test 4: Finding similar patterns...');
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
package-lock.jsonis excluded by!**/package-lock.json
📒 Files selected for processing (29)
.eslintrc.json(1 hunks).prettierignore(1 hunks).prettierrc(1 hunks)README.md(10 hunks)docs/todo-phase1.md(1 hunks)package.json(2 hunks)src/chroma-vector-store.ts(9 hunks)src/cli-main.ts(1 hunks)src/cli.ts(31 hunks)src/code-parser.ts(26 hunks)src/code-storage.ts(4 hunks)src/config.ts(6 hunks)src/embedding-service.ts(10 hunks)src/file-watcher.ts(9 hunks)src/index.ts(10 hunks)src/logger.ts(7 hunks)src/semantic-search.ts(10 hunks)src/tree-sitter-types.d.ts(1 hunks)src/types.ts(2 hunks)src/vector-store.ts(12 hunks)test/debug-watcher.ts(3 hunks)test/test-embedding.ts(7 hunks)test/test-file-watcher.ts(12 hunks)test/test-integration.ts(3 hunks)test/test-parser.ts(4 hunks)test/test-semantic-search.ts(9 hunks)test/test-storage.ts(5 hunks)test/test-vector-store.ts(9 hunks)tsconfig.json(1 hunks)
🧰 Additional context used
🧬 Code graph analysis (8)
src/vector-store.ts (1)
src/types.ts (1)
VectorSearchResult(31-41)
src/semantic-search.ts (3)
src/code-parser.ts (1)
CodeParser(65-890)src/embedding-service.ts (1)
EmbeddingService(41-315)src/types.ts (1)
EmbeddingService(58-63)
test/test-file-watcher.ts (1)
src/file-watcher.ts (1)
FileWatcher(95-281)
src/chroma-vector-store.ts (1)
src/types.ts (1)
VectorSearchResult(31-41)
src/cli.ts (5)
src/index.ts (1)
ContextEngine(34-342)src/semantic-search.ts (1)
SemanticSearch(12-228)src/code-parser.ts (1)
CodeParser(65-890)src/embedding-service.ts (1)
EmbeddingService(41-315)src/types.ts (1)
EmbeddingService(58-63)
test/test-integration.ts (3)
src/file-watcher.ts (2)
FileWatcher(95-281)FileWatcherOptions(8-27)src/code-parser.ts (1)
CodeParser(65-890)src/semantic-search.ts (2)
createSemanticSearchService(231-233)handleFileChange(175-195)
src/code-parser.ts (1)
src/types.ts (1)
CodeChunk(1-17)
src/index.ts (1)
src/logger.ts (1)
error(170-172)
🪛 Biome (2.1.2)
src/cli.ts
[error] 619-620: Other switch clauses can erroneously access this declaration.
Wrap the declaration in a block to restrict its access to the switch clause.
The declaration is defined in this switch clause:
Safe fix: Wrap the declaration in a block.
(lint/correctness/noSwitchDeclarations)
[error] 620-621: Other switch clauses can erroneously access this declaration.
Wrap the declaration in a block to restrict its access to the switch clause.
The declaration is defined in this switch clause:
Safe fix: Wrap the declaration in a block.
(lint/correctness/noSwitchDeclarations)
[error] 621-623: Other switch clauses can erroneously access this declaration.
Wrap the declaration in a block to restrict its access to the switch clause.
The declaration is defined in this switch clause:
Safe fix: Wrap the declaration in a block.
(lint/correctness/noSwitchDeclarations)
[error] 629-630: Other switch clauses can erroneously access this declaration.
Wrap the declaration in a block to restrict its access to the switch clause.
The declaration is defined in this switch clause:
Safe fix: Wrap the declaration in a block.
(lint/correctness/noSwitchDeclarations)
[error] 646-647: Other switch clauses can erroneously access this declaration.
Wrap the declaration in a block to restrict its access to the switch clause.
The declaration is defined in this switch clause:
Safe fix: Wrap the declaration in a block.
(lint/correctness/noSwitchDeclarations)
🪛 markdownlint-cli2 (0.18.1)
README.md
218-218: Emphasis used instead of a heading
(MD036, no-emphasis-as-heading)
🔇 Additional comments (39)
src/code-storage.ts (2)
60-66: Trailing comma in options object is fine (Prettier‑friendly).No functional change; keeps diffs cleaner. LGTM.
258-264: Trailing comma in getStats return object is fine.Consistent with Prettier; no behavior change.
tsconfig.json (1)
27-28: Formatting-only change LGTM.Include/exclude arrays compressed to single lines; no behavioral impact.
test/test-embedding.ts (1)
25-26: Formatting-only changes look fine.Trailing commas and log wrapping are consistent with the new Prettier rules.
Also applies to: 42-43, 60-61, 71-72, 91-92, 101-103
src/tree-sitter-types.d.ts (1)
35-35: EOF newline fix LGTM.Good for tooling compatibility.
src/embedding-service.ts (4)
65-70: ESM import runtime check.Dynamic import of '@xenova/transformers' must remain native import() after transpile. With CommonJS module output, confirm TS preserves it for your TS version.
See the tsconfig comment for a verification script.
96-101: Normalize=true and cosine similarity are aligned.No issues spotted.
276-276: Similarity calculation usage LGTM.Correct guard for dimension mismatch.
320-320: Factory export unchanged; fine.No action needed.
.prettierrc (1)
1-22: Prettier config looks sane and consistent.Matches common conventions; JSON override is helpful.
test/test-parser.ts (2)
116-119: Console output formatting change LGTM.Purely aesthetic; no behavior change.
136-149: Incremental parsing test: good coverage.Nice touch adding a new function to validate change detection.
.prettierignore (1)
1-7: Ignore list generally good.Covers typical artifacts and large test corpora.
Also applies to: 11-66
docs/todo-phase1.md (1)
36-40: Do not mark "PocketFlow Integration" done — no implementation/tests found.
rg search of src/** returned no occurrences of "PocketFlow", "orchestrat", or "workflow"; only matches were src/config.ts:8 and src/config.ts:94 (retryDelay). docs/todo-phase1.md (lines 36–40) lists the item as completed; add the PocketFlow implementation and tests and ensure CI is green, or revert the checklist entry..eslintrc.json (1)
6-12: Strengthen TS/Prettier integration; prevent rule conflicts.Extend with TS/Prettier presets and use the TS-aware unused-vars rule. Also drop style rules that Prettier will manage to avoid churn.
[ suggest_recommended_refactor ]
Apply:
- "extends": ["eslint:recommended"], + "extends": ["eslint:recommended", "plugin:@typescript-eslint/recommended", "plugin:prettier/recommended"], "plugins": ["@typescript-eslint", "prettier"], "rules": { "prettier/prettier": "warn", "no-console": "off", "no-debugger": "error", "prefer-const": "error", "no-var": "error", - "object-shorthand": "warn", - "prefer-arrow-callback": "warn", - "prefer-template": "warn", - "template-curly-spacing": "error", - "arrow-spacing": "error", - "comma-dangle": ["warn", "always-multiline"], - "quotes": ["error", "single", { "avoidEscape": true }], - "semi": ["error", "always"], - "no-unused-vars": "warn", + "object-shorthand": "warn", + "prefer-arrow-callback": "warn", + "prefer-template": "warn", + "@typescript-eslint/no-unused-vars": ["warn", { "argsIgnorePattern": "^_", "varsIgnorePattern": "^_" }], "no-empty": "warn", "no-case-declarations": "warn" },Also applies to: 13-31
src/cli-main.ts (1)
12-15: LGTM.EOF newline fix and CLI bootstrap look good.
src/types.ts (1)
54-56: No action required — implementers already define clear() and close(). Only implementer found: src/chroma-vector-store.ts — it contains both clear() and close().package.json (1)
53-59: Confirm ESLint/TypeScript/Prettier peer-dependency compatibility.node_modules not present in the sandbox (MODULE_NOT_FOUND); unable to verify peerDependencies. Verify package.json (lines 53–59) against each package's peerDependencies/engines locally (e.g., run npm view @ peerDependencies or npm ls).
test/test-vector-store.ts (2)
71-72: LGTM: sensible embedding options for tests.Using an explicit cacheSize keeps memory bounded during the run.
121-127: LGTM: searchSimilar usage and logging.Signature matches ChromaVectorStore; logging guards are safe.
test/test-storage.ts (2)
24-25: LGTM: deterministic timestamps in fixtures.Keeps test objects uniform and sortable.
Also applies to: 41-42, 58-59
67-70: LGTM: explicit non-persistent storage for tests.Avoids polluting local FS.
src/file-watcher.ts (2)
192-197: LGTM: structured DirectoryAdded event payload.Matches the typed contract and improves readability.
156-157: Confirmed: atomic & awaitWriteFinish forwarded correctly.
Passed to chokidar as this.options.awaitWriteFinish and this.options.atomic; defaults are set before the ...options spread so caller-provided values override.src/semantic-search.ts (1)
138-151: LGTM: result filtering and ordering.Simple and clear; respects minSimilarity then sorts desc.
src/chroma-vector-store.ts (2)
65-67: LGTM: persist timestamp in metadata.Useful for staleness checks and reindexing strategies.
Also applies to: 173-175
24-30: Auth shape verified — supported by Chroma JS client.ChromaClient accepts auth: { provider: 'token', credentials: '...' } (tokenHeaderType is optional); no change required.
test/debug-watcher.ts (1)
18-22: LGTM: awaitWriteFinish tuned for debug runs.Reasonable values to reduce event flapping.
test/test-file-watcher.ts (5)
29-35: LGTM: instance creation with ignoreInitial.Matches typical test baseline.
65-73: Tests expect atomic/awaitWriteFinish to be configurable.These rely on FileWatcher honoring
awaitWriteFinish: falseandatomic: false. Ensure constructor fix lands (see src/file-watcher.ts).
75-86: LGTM: event listener registration before start.Prevents missed events during watcher boot.
Also applies to: 131-141, 232-241
249-258: LGTM: tracked files assertions are resilient.Substring match handles platform-specific separators.
291-316: LGTM: error-handling test is pragmatic.Accepts either emitted error or running watcher on nonexistent path.
test/test-semantic-search.ts (3)
65-66: Formatting reflow looks goodWrapping the long string literal improves readability without changing behavior.
168-174: Multi‑line call formatting is fineNo functional changes; call site remains equivalent.
215-215: Do not convert this file to ESM — repo targets CommonJStsconfig.json sets "module": "CommonJS" and package.json has no "type" field; multiple files use module.exports, so replacing this single export with
export { ... }would create mixed module systems and likely break build/lint.
Occurrences: test/test-embedding.ts:146, test/test-vector-store.ts:195, test/test-semantic-search.ts:215, test/test-storage.ts:148, src/code-parser.ts:902-920.Likely an incorrect or invalid review comment.
src/index.ts (1)
135-139: Nice: watcher ready visibilityReady event + watched count improve observability at startup.
src/logger.ts (2)
88-99: Console method selection is fineBehavior preserved; reads clean and satisfies typical ESLint style.
245-245: Default export retainedNo issues with the default export naming.
| *.js | ||
| *.js.map |
There was a problem hiding this comment.
Do not ignore all .js files.
This will prevent Prettier from formatting any JS sources (including config scripts). Narrow the pattern to build outputs only.
-*.js
-*.js.map
+# Built JS maps (if any) in dist/build only
+dist/**/*.js
+dist/**/*.js.map
+build/**/*.js
+build/**/*.js.map📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| *.js | |
| *.js.map | |
| # Built JS maps (if any) in dist/build only | |
| dist/**/*.js | |
| dist/**/*.js.map | |
| build/**/*.js | |
| build/**/*.js.map |
🤖 Prompt for AI Agents
In .prettierignore around lines 8-9, the entries "*.js" and "*.js.map" are too
broad and prevent Prettier from formatting all JavaScript sources; replace them
with patterns that only ignore build/artifact outputs (for example
/dist/**/*.js, /build/**/*.js, *.min.js or specific generated folders) so source
JS and config scripts remain formatted; remove the global "*.js" and "*.js.map"
lines and add one or more targeted paths that match only compiled output.
| "@types/chokidar": "^2.1.3", | ||
| "@types/node": "^20.11.0", |
There was a problem hiding this comment.
Remove @types/chokidar (v3 already ships types) to avoid conflicts.
chokidar@^3 includes its own types; @types/chokidar@^2 can cause type mismatches.
Apply:
- "@types/chokidar": "^2.1.3",📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| "@types/chokidar": "^2.1.3", | |
| "@types/node": "^20.11.0", | |
| "@types/node": "^20.11.0", |
🤖 Prompt for AI Agents
In package.json around lines 51-52, remove the "@types/chokidar": "^2.1.3"
dependency because chokidar v3 includes its own types and the separate @types
package (v2) causes type conflicts; delete that line from
dependencies/devDependencies, run your package manager (npm/yarn/pnpm) to update
node_modules and the lockfile, and then rebuild/tsc to verify no type errors
remain.
| - [CI/CD Setup Guide](CI_CD_SETUP.md) - Quick start and configuration | ||
| - [CI/CD Documentation](CI_CD_DOCUMENTATION.md) - Technical details | ||
|
|
There was a problem hiding this comment.
🧩 Analysis chain
Verify referenced CI/CD docs exist (and link paths).
Ensure CI_CD_SETUP.md and CI_CD_DOCUMENTATION.md are present or adjust links.
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
fd -a 'CI*_*.md' -E 'node_modules' -E 'dist' || true
test -f CI_CD_SETUP.md && echo "OK: CI_CD_SETUP.md" || echo "Missing: CI_CD_SETUP.md"
test -f CI_CD_DOCUMENTATION.md && echo "OK: CI_CD_DOCUMENTATION.md" || echo "Missing: CI_CD_DOCUMENTATION.md"Length of output: 55
Fix README links: missing CI/CD docs
README.md (lines 420–422) links to CI_CD_SETUP.md and CI_CD_DOCUMENTATION.md but both files are missing from the repo (verification: Missing: CI_CD_SETUP.md; Missing: CI_CD_DOCUMENTATION.md). Add the two files at the referenced paths or update the links to the correct locations.
🤖 Prompt for AI Agents
In README.md around lines 420 to 422 the links point to CI_CD_SETUP.md and
CI_CD_DOCUMENTATION.md which are missing from the repo; either add new files at
those exact paths containing the CI/CD quick start and technical docs, or update
the README links to point to the correct existing files or locations (e.g.,
docs/ci/CI_CD_SETUP.md and docs/ci/CI_CD_DOCUMENTATION.md) and verify relative
paths and filenames; after adding or updating, run a repo link check to ensure
no broken references remain.
| awaitWriteFinish: options.awaitWriteFinish || { | ||
| stabilityThreshold: 2000, | ||
| pollInterval: 100 | ||
| pollInterval: 100, | ||
| }, | ||
| ignorePermissionErrors: options.ignorePermissionErrors !== false, | ||
| atomic: options.atomic || true, | ||
| ...options | ||
| ...options, | ||
| } as Required<FileWatcherOptions>; |
There was a problem hiding this comment.
Boolean options incorrectly override user-specified false.
awaitWriteFinish: options.awaitWriteFinish || {…} and atomic: options.atomic || true force defaults when false is passed. Use nullish coalescing.
Apply this diff:
- awaitWriteFinish: options.awaitWriteFinish || {
+ awaitWriteFinish: options.awaitWriteFinish ?? {
stabilityThreshold: 2000,
pollInterval: 100,
},
- ignorePermissionErrors: options.ignorePermissionErrors !== false,
- atomic: options.atomic || true,
+ ignorePermissionErrors: options.ignorePermissionErrors !== false,
+ atomic: options.atomic ?? true,
...options,This also aligns with tests that pass awaitWriteFinish: false and atomic: false.
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| awaitWriteFinish: options.awaitWriteFinish || { | |
| stabilityThreshold: 2000, | |
| pollInterval: 100 | |
| pollInterval: 100, | |
| }, | |
| ignorePermissionErrors: options.ignorePermissionErrors !== false, | |
| atomic: options.atomic || true, | |
| ...options | |
| ...options, | |
| } as Required<FileWatcherOptions>; | |
| awaitWriteFinish: options.awaitWriteFinish ?? { | |
| stabilityThreshold: 2000, | |
| pollInterval: 100, | |
| }, | |
| ignorePermissionErrors: options.ignorePermissionErrors !== false, | |
| atomic: options.atomic ?? true, | |
| ...options, | |
| } as Required<FileWatcherOptions>; |
🤖 Prompt for AI Agents
In src/file-watcher.ts around lines 122 to 129, the boolean defaults use logical
OR which treats explicit false as falsy and thus overrides user-specified false;
replace `options.awaitWriteFinish || { ... }` with `options.awaitWriteFinish ??
{ ... }` and `options.atomic || true` with `options.atomic ?? true` so only null
or undefined fall back to defaults, keeping the rest of the options
spread/casting as-is.
| this.logger.debug( | ||
| `Indexing batch ${Math.floor(i / batchSize) + 1}/${Math.ceil(supportedFiles.length / batchSize)}` | ||
| ); | ||
|
|
||
| await Promise.all( | ||
| batch.map(async filePath => { | ||
| try { | ||
| const content = readFileSync(filePath, 'utf-8'); | ||
| await this.semanticSearch!.indexFile(filePath, content); | ||
| } catch (error) { | ||
| this.logger.error(`Failed to index file ${filePath} during initial indexing:`, error); | ||
| } | ||
| }) | ||
| ); |
There was a problem hiding this comment.
Initial indexing bypasses size limit and may read huge files
The initial batch path reads content before any size check and skips the configured limit entirely.
Add a stat‑based check before reading:
await Promise.all(
batch.map(async filePath => {
try {
- const content = readFileSync(filePath, 'utf-8');
- await this.semanticSearch!.indexFile(filePath, content);
+ const { size } = require('fs').statSync(filePath);
+ const config = this.configManager.getConfig();
+ if (size > config.security.maxFileSize) {
+ this.logger.warn(`File too large, skipping during initial indexing: ${filePath}`);
+ return;
+ }
+ const content = readFileSync(filePath, 'utf-8');
+ await this.semanticSearch!.indexFile(filePath, content);
} catch (error) {
this.logger.error(`Failed to index file ${filePath} during initial indexing:`, error);
}
})
);📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| this.logger.debug( | |
| `Indexing batch ${Math.floor(i / batchSize) + 1}/${Math.ceil(supportedFiles.length / batchSize)}` | |
| ); | |
| await Promise.all( | |
| batch.map(async filePath => { | |
| try { | |
| const content = readFileSync(filePath, 'utf-8'); | |
| await this.semanticSearch!.indexFile(filePath, content); | |
| } catch (error) { | |
| this.logger.error(`Failed to index file ${filePath} during initial indexing:`, error); | |
| } | |
| }) | |
| ); | |
| this.logger.debug( | |
| `Indexing batch ${Math.floor(i / batchSize) + 1}/${Math.ceil(supportedFiles.length / batchSize)}` | |
| ); | |
| await Promise.all( | |
| batch.map(async filePath => { | |
| try { | |
| const { size } = require('fs').statSync(filePath); | |
| const config = this.configManager.getConfig(); | |
| if (size > config.security.maxFileSize) { | |
| this.logger.warn(`File too large, skipping during initial indexing: ${filePath}`); | |
| return; | |
| } | |
| const content = readFileSync(filePath, 'utf-8'); | |
| await this.semanticSearch!.indexFile(filePath, content); | |
| } catch (error) { | |
| this.logger.error(`Failed to index file ${filePath} during initial indexing:`, error); | |
| } | |
| }) | |
| ); |
| uptime: this.isRunning ? Date.now() - this.startTime : 0, | ||
| }; |
There was a problem hiding this comment.
🧩 Analysis chain
uptime never increases; startTime is never set
getStats uses startTime, but start() doesn’t assign it.
Add after the watcher starts (before logging “started successfully”):
// in start(), after: await this.fileWatcher.start();
this.startTime = Date.now();🏁 Script executed:
#!/bin/bash
rg -nP -C2 '\bstartTime\b' --type=tsLength of output: 2095
Set startTime in start() so uptime increments
startTime is declared (src/index.ts:41) and used in getStats (src/index.ts:339) but never assigned — uptime will remain 0. Add after the watcher starts (after await this.fileWatcher.start(), before logging "started successfully"):
this.startTime = Date.now();🤖 Prompt for AI Agents
In src/index.ts around lines 339-340, startTime is declared but never assigned
so getStats uptime stays 0; after the watcher starts (immediately after the
await this.fileWatcher.start() call and before logging "started successfully")
assign the start timestamp to this.startTime using the current time so uptime
increments correctly; ensure this assignment only happens on successful start
and consider clearing or resetting startTime when stopping the watcher.
| const formattedArgs = | ||
| args.length > 0 | ||
| ? ` ${args | ||
| .map(arg => (typeof arg === 'object' ? JSON.stringify(arg, null, 2) : String(arg))) | ||
| .join(' ')}` | ||
| : ''; | ||
|
|
There was a problem hiding this comment.
Safer arg formatting; include Error stacks
JSON.stringify drops Error details and can throw on cycles. Prefer util.inspect and handle Error specially.
- const formattedArgs =
- args.length > 0
- ? ` ${args
- .map(arg => (typeof arg === 'object' ? JSON.stringify(arg, null, 2) : String(arg)))
- .join(' ')}`
- : '';
+ const util = require('util');
+ const formattedArgs =
+ args.length > 0
+ ? ' ' +
+ args
+ .map(arg =>
+ arg instanceof Error ? arg.stack || arg.message : util.inspect(arg, { depth: 6, colors: false })
+ )
+ .join(' ')
+ : '';📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| const formattedArgs = | |
| args.length > 0 | |
| ? ` ${args | |
| .map(arg => (typeof arg === 'object' ? JSON.stringify(arg, null, 2) : String(arg))) | |
| .join(' ')}` | |
| : ''; | |
| const util = require('util'); | |
| const formattedArgs = | |
| args.length > 0 | |
| ? ' ' + | |
| args | |
| .map(arg => | |
| arg instanceof Error ? arg.stack || arg.message : util.inspect(arg, { depth: 6, colors: false }) | |
| ) | |
| .join(' ') | |
| : ''; |
🤖 Prompt for AI Agents
In src/logger.ts around lines 71 to 77, the current args formatting uses
JSON.stringify which drops Error stacks and can throw on cyclic structures;
replace that logic to use Node's util.inspect and special-case Error instances
to include their stack (err.stack || err.message) so errors are fully logged,
and call util.inspect(obj, { depth: null, maxArrayLength: null }) for other
objects to avoid cyclic-structure exceptions and preserve readable output;
ensure util is imported and trim/format the resulting string similarly to the
existing code.
| path: this.options.persistDirectory, | ||
| }); |
There was a problem hiding this comment.
ChromaClient path likely incorrect (points to a local folder).
path should be the server URL (e.g., http://localhost:8000). Passing a filesystem directory here will not connect to a server.
Example:
- this.client = new ChromaClient({
- path: this.options.persistDirectory,
- });
+ this.client = new ChromaClient({
+ path: process.env.CHROMA_URL || 'http://localhost:8000',
+ });If you intended true local persistence, confirm JS client supports it; otherwise drop persistDirectory from here and document server config.
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| path: this.options.persistDirectory, | |
| }); | |
| this.client = new ChromaClient({ | |
| path: process.env.CHROMA_URL || 'http://localhost:8000', | |
| }); |
🤖 Prompt for AI Agents
In src/vector-store.ts around lines 61 to 62, the ChromaClient is being
constructed with path: this.options.persistDirectory which points at a local
filesystem folder; change this to the Chroma server URL (e.g.,
http://localhost:8000) when connecting to a running Chroma server, or if you
truly intend local on-disk persistence, verify the JS client supports a
filesystem-backed mode and switch to the correct client/config option for local
persistence; if server mode is used, remove persistDirectory from the client
config and instead document/accept a serverUrl option and pass that to
ChromaClient.
| const { CodeParser } = require('../src/code-parser'); | ||
| const parser = new CodeParser(); | ||
| const storage = new CodeStorage(); | ||
| const semanticSearch = createSemanticSearchService(); |
There was a problem hiding this comment.
Prefer ESM import over require to satisfy ESLint
Avoid require in TS; dynamic import works inside async function and keeps tests ESM‑friendly.
- const { CodeParser } = require('../src/code-parser');
- const parser = new CodeParser();
+ const { CodeParser } = await import('../src/code-parser');
+ const parser = new CodeParser();Committable suggestion skipped: line range outside the PR's diff.
🤖 Prompt for AI Agents
In test/test-integration.ts around lines 24-27, replace the CommonJS require
usage with ESM imports to satisfy ESLint: remove the require calls and use
top-level ES module imports for CodeParser, CodeStorage and
createSemanticSearchService (or, if the test runner executes tests as CommonJS,
perform dynamic import() inside an async beforeAll/describe setup), then
instantiate the classes as before; ensure the file is treated as ESM by the test
runner config or by using dynamic import in async setup.
| // Clean up | ||
| console.log('\n🧹 Cleaning up...'); | ||
| if (fileWatcher) { | ||
| await fileWatcher.stop(); | ||
| } | ||
|
|
||
| if (fs.existsSync(testDir)) { | ||
| fs.rmSync(testDir, { recursive: true, force: true }); | ||
| } | ||
|
|
||
| console.log(' Cleanup completed'); | ||
| } |
There was a problem hiding this comment.
🛠️ Refactor suggestion
🧩 Analysis chain
Close semanticSearch in cleanup (resource hygiene)
fileWatcher is stopped; also close the semantic search service to release resources.
Add (outside the shown range) a variable hoist and close call:
// near line 12, hoist:
let semanticSearch: ReturnType<typeof createSemanticSearchService> | null = null;
// near line 27, assign:
semanticSearch = createSemanticSearchService();
// in finally, before removing dir:
if (semanticSearch) {
await semanticSearch.close();
}To find other tests needing this:
🏁 Script executed:
#!/bin/bash
rg -nP --type=ts 'createSemanticSearchService\(' -C2Length of output: 1029
Close semanticSearch in cleanup (resource hygiene)
-
Hoist semanticSearch to the test scope and close it in the finally before removing testDir. In test/test-integration.ts semanticSearch is created as
constat line 27 and cleanup is at lines 256–267 — change to:let semanticSearch: ReturnType<typeof createSemanticSearchService> | null = null;semanticSearch = createSemanticSearchService();- in finally (before
fs.rmSync):if (semanticSearch) await semanticSearch.close();
-
Also update test/test-semantic-search.ts (semanticSearch created at line 104 inside
try) to the same hoist-and-close pattern.
🤖 Prompt for AI Agents
In test/test-integration.ts around lines 256–267 and
test/test-semantic-search.ts around line 104, semanticSearch is created inside
narrow scopes and never explicitly closed before removing the testDir; hoist the
variable to test scope as: let semanticSearch: ReturnType<typeof
createSemanticSearchService> | null = null, assign it where currently created
(semanticSearch = createSemanticSearchService()), and in the finally block
before fs.rmSync call check if (semanticSearch) await semanticSearch.close(); to
ensure the service is closed prior to deleting test artifacts.
1 participant
This PR aims at setting up eslinting and prettier for code formating,. No closes for this yet.
I am used in pushing the package-lock file 😆 since our project requires you push it, hope it no big deal.
Summary by CodeRabbit
New Features
Documentation
Chores