forge-server 0.1.0

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":""}

package/dist/index.js

@@ -0,0 +1,37 @@
+#!/usr/bin/env node
+// index.ts — B19
+// Entry point for the dk-forge-server MCP server.
+// Starts the server on stdio transport for use with Claude and other MCP clients.
+import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
+import { createForgeServer } from './server.js';
+async function main() {
+    const server = await createForgeServer();
+    const transport = new StdioServerTransport();
+    // Connect the server to the stdio transport
+    await server.connect(transport);
+    // Handle graceful shutdown on SIGINT (Ctrl-C) and SIGTERM
+    const shutdown = async (signal) => {
+        process.stderr.write(`\ndk-forge-server: received ${signal}, shutting down...\n`);
+        try {
+            // Close storage connections if the server exposes a close method
+            const serverWithClose = server;
+            if (typeof serverWithClose._closeConnections === 'function') {
+                await serverWithClose._closeConnections();
+            }
+            await server.close();
+        }
+        catch (err) {
+            process.stderr.write(`dk-forge-server: shutdown error: ${String(err)}\n`);
+        }
+        process.exit(0);
+    };
+    process.on('SIGINT', () => { void shutdown('SIGINT'); });
+    process.on('SIGTERM', () => { void shutdown('SIGTERM'); });
+    // Keep the process alive — stdio transport reads from stdin
+    // Node.js will keep running while stdin is open
+}
+main().catch((err) => {
+    process.stderr.write(`dk-forge-server: fatal error: ${String(err)}\n`);
+    process.exit(1);
+});
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";AACA,iBAAiB;AACjB,kDAAkD;AAClD,kFAAkF;AAElF,OAAO,EAAE,oBAAoB,EAAE,MAAM,2CAA2C,CAAC;AACjF,OAAO,EAAE,iBAAiB,EAAE,MAAM,aAAa,CAAC;AAEhD,KAAK,UAAU,IAAI;IACjB,MAAM,MAAM,GAAG,MAAM,iBAAiB,EAAE,CAAC;IACzC,MAAM,SAAS,GAAG,IAAI,oBAAoB,EAAE,CAAC;IAE7C,4CAA4C;IAC5C,MAAM,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IAEhC,0DAA0D;IAC1D,MAAM,QAAQ,GAAG,KAAK,EAAE,MAAc,EAAE,EAAE;QACxC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,+BAA+B,MAAM,sBAAsB,CAAC,CAAC;QAElF,IAAI,CAAC;YACH,iEAAiE;YACjE,MAAM,eAAe,GAAG,MAEvB,CAAC;YACF,IAAI,OAAO,eAAe,CAAC,iBAAiB,KAAK,UAAU,EAAE,CAAC;gBAC5D,MAAM,eAAe,CAAC,iBAAiB,EAAE,CAAC;YAC5C,CAAC;YAED,MAAM,MAAM,CAAC,KAAK,EAAE,CAAC;QACvB,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,oCAAoC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAC5E,CAAC;QAED,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC,CAAC;IAEF,OAAO,CAAC,EAAE,CAAC,QAAQ,EAAE,GAAG,EAAE,GAAG,KAAK,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACzD,OAAO,CAAC,EAAE,CAAC,SAAS,EAAE,GAAG,EAAE,GAAG,KAAK,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAE3D,4DAA4D;IAC5D,gDAAgD;AAClD,CAAC;AAED,IAAI,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE;IACnB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,iCAAiC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IACvE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC,CAAC,CAAC"}

package/dist/ingestion/chunker.d.ts

@@ -0,0 +1,19 @@
+/**
+ * AST-aware code chunker (ADR-5).
+ * Primary strategy: each top-level entity becomes one chunk (function, class, interface, etc.)
+ * Fallback: fixed-size 512-token chunks with 64-token overlap for non-parseable files.
+ * Large entities (>1024 tokens) are split at method boundaries.
+ */
+import type { ParsedEntity, ParsedImport, CodeChunk, ChunkType } from '../util/types.js';
+/**
+ * Generate chunks from parsed AST entities.
+ * Called when we have a successful AST parse result.
+ */
+export declare function chunkFromEntities(entities: ParsedEntity[], imports: ParsedImport[], fileContent: string, filePath: string, repoId: string, language: string, indexedAt: number): CodeChunk[];
+/**
+ * Generate fixed-size chunks for non-parseable files (markdown, JSON, YAML, config).
+ */
+export declare function chunkFixed(content: string, filePath: string, repoId: string, language: string, indexedAt: number): CodeChunk[];
+export declare function makeChunk(content: string, filePath: string, repoId: string, entityType: ChunkType, startLine: number, endLine: number, language: string, indexedAt: number, entityName?: string): CodeChunk;
+export declare function buildChunkHeader(filePath: string, entityName: string | null, entityType: ChunkType, startLine: number, endLine: number): string;
+//# sourceMappingURL=chunker.d.ts.map

package/dist/ingestion/chunker.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"chunker.d.ts","sourceRoot":"","sources":["../../src/ingestion/chunker.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAKH,OAAO,KAAK,EAAE,YAAY,EAAE,YAAY,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAOzF;;;GAGG;AACH,wBAAgB,iBAAiB,CAC/B,QAAQ,EAAE,YAAY,EAAE,EACxB,OAAO,EAAE,YAAY,EAAE,EACvB,WAAW,EAAE,MAAM,EACnB,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,MAAM,EACd,QAAQ,EAAE,MAAM,EAChB,SAAS,EAAE,MAAM,GAChB,SAAS,EAAE,CAcb;AAED;;GAEG;AACH,wBAAgB,UAAU,CACxB,OAAO,EAAE,MAAM,EACf,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,MAAM,EACd,QAAQ,EAAE,MAAM,EAChB,SAAS,EAAE,MAAM,GAChB,SAAS,EAAE,CAyCb;AAsGD,wBAAgB,SAAS,CACvB,OAAO,EAAE,MAAM,EACf,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,MAAM,EACd,UAAU,EAAE,SAAS,EACrB,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,EACf,QAAQ,EAAE,MAAM,EAChB,SAAS,EAAE,MAAM,EACjB,UAAU,CAAC,EAAE,MAAM,GAClB,SAAS,CAkBX;AAED,wBAAgB,gBAAgB,CAC9B,QAAQ,EAAE,MAAM,EAChB,UAAU,EAAE,MAAM,GAAG,IAAI,EACzB,UAAU,EAAE,SAAS,EACrB,SAAS,EAAE,MAAM,EACjB,OAAO,EAAE,MAAM,GACd,MAAM,CAGR"}

package/dist/ingestion/chunker.js

@@ -0,0 +1,189 @@
+/**
+ * AST-aware code chunker (ADR-5).
+ * Primary strategy: each top-level entity becomes one chunk (function, class, interface, etc.)
+ * Fallback: fixed-size 512-token chunks with 64-token overlap for non-parseable files.
+ * Large entities (>1024 tokens) are split at method boundaries.
+ */
+import { createHash } from 'crypto';
+import { randomUUID } from 'crypto';
+import { extname } from 'path';
+import { estimateTokens } from '../util/token-counter.js';
+const CHUNK_MAX_TOKENS = 1024;
+const FIXED_CHUNK_SIZE_TOKENS = 512;
+const FIXED_CHUNK_OVERLAP_TOKENS = 64;
+/**
+ * Generate chunks from parsed AST entities.
+ * Called when we have a successful AST parse result.
+ */
+export function chunkFromEntities(entities, imports, fileContent, filePath, repoId, language, indexedAt) {
+    const chunks = [];
+    // Generate the file-level chunk (imports + any top-level statements not in entities)
+    const fileLevelChunk = buildFileLevelChunk(imports, fileContent, filePath, repoId, language, indexedAt, entities);
+    if (fileLevelChunk)
+        chunks.push(fileLevelChunk);
+    // Generate one chunk per entity
+    for (const entity of entities) {
+        const entityChunks = buildEntityChunks(entity, filePath, repoId, language, indexedAt);
+        chunks.push(...entityChunks);
+    }
+    return chunks;
+}
+/**
+ * Generate fixed-size chunks for non-parseable files (markdown, JSON, YAML, config).
+ */
+export function chunkFixed(content, filePath, repoId, language, indexedAt) {
+    const ext = extname(filePath).toLowerCase();
+    const chunkType = detectNonCodeType(ext);
+    // Split content into lines for better boundary detection
+    const lines = content.split('\n');
+    const chunks = [];
+    let currentLines = [];
+    let currentTokens = 0;
+    let startLine = 1;
+    const flushChunk = (endLine) => {
+        if (currentLines.length === 0)
+            return;
+        const chunkContent = buildChunkHeader(filePath, null, chunkType, startLine, endLine) +
+            currentLines.join('\n');
+        chunks.push(makeChunk(chunkContent, filePath, repoId, chunkType, startLine, endLine, language, indexedAt));
+    };
+    for (const line of lines) {
+        const lineTokens = estimateTokens(line);
+        if (currentTokens + lineTokens > FIXED_CHUNK_SIZE_TOKENS && currentLines.length > 0) {
+            flushChunk(startLine + currentLines.length - 1);
+            // Overlap: keep last N token-equivalent lines
+            const overlapLines = getOverlapLines(currentLines, FIXED_CHUNK_OVERLAP_TOKENS);
+            startLine = startLine + currentLines.length - overlapLines.length;
+            currentLines = overlapLines;
+            currentTokens = estimateTokensForLines(overlapLines);
+        }
+        currentLines.push(line);
+        currentTokens += lineTokens;
+    }
+    if (currentLines.length > 0) {
+        flushChunk(startLine + currentLines.length - 1);
+    }
+    return chunks;
+}
+function buildFileLevelChunk(imports, fileContent, filePath, repoId, language, indexedAt, entities) {
+    if (imports.length === 0 && entities.length > 0)
+        return null;
+    // Build import block
+    const importLines = imports.map(imp => {
+        if (imp.isDefault) {
+            return `import ${imp.specifiers} from '${imp.fromPath}';`;
+        }
+        return `import ${imp.specifiers} from '${imp.fromPath}';`;
+    });
+    // Find the last line of the last import to get file-level lines
+    const importBlock = importLines.join('\n');
+    if (!importBlock.trim())
+        return null;
+    const header = buildChunkHeader(filePath, null, 'file_level', 1, importLines.length);
+    const content = header + importBlock;
+    if (estimateTokens(content) < 10)
+        return null;
+    return makeChunk(content, filePath, repoId, 'file_level', 1, importLines.length, language, indexedAt);
+}
+function buildEntityChunks(entity, filePath, repoId, language, indexedAt) {
+    const chunkType = entityTypeToChunkType(entity.type);
+    const header = buildChunkHeader(filePath, entity.name, chunkType, entity.startLine, entity.endLine);
+    const content = header + entity.sourceText;
+    const tokens = estimateTokens(content);
+    if (tokens <= CHUNK_MAX_TOKENS) {
+        // Single chunk for this entity
+        return [makeChunk(content, filePath, repoId, chunkType, entity.startLine, entity.endLine, language, indexedAt)];
+    }
+    // Entity is too large - split at method/property boundaries (heuristic line split)
+    return splitLargeEntity(content, entity, filePath, repoId, chunkType, language, indexedAt);
+}
+function splitLargeEntity(content, entity, filePath, repoId, chunkType, language, indexedAt) {
+    const chunks = [];
+    const lines = content.split('\n');
+    let currentLines = [];
+    let currentTokens = 0;
+    let chunkIndex = 0;
+    let chunkStartLine = entity.startLine;
+    const flush = () => {
+        if (currentLines.length === 0)
+            return;
+        const chunkContent = currentLines.join('\n');
+        const chunkEndLine = chunkStartLine + currentLines.length - 1;
+        chunks.push(makeChunk(chunkContent, filePath, repoId, chunkType, chunkStartLine, chunkEndLine, language, indexedAt, `${entity.name}_part${chunkIndex++}`));
+        chunkStartLine = chunkEndLine + 1;
+        currentLines = [];
+        currentTokens = 0;
+    };
+    for (const line of lines) {
+        const lineTokens = estimateTokens(line);
+        if (currentTokens + lineTokens > CHUNK_MAX_TOKENS && currentLines.length > 0) {
+            flush();
+        }
+        currentLines.push(line);
+        currentTokens += lineTokens;
+    }
+    flush();
+    return chunks;
+}
+export function makeChunk(content, filePath, repoId, entityType, startLine, endLine, language, indexedAt, entityName) {
+    const id = randomUUID();
+    const contentHash = createHash('sha256').update(content, 'utf8').digest('hex');
+    return {
+        id,
+        repoId,
+        filePath,
+        entityName: entityName ?? null,
+        entityType,
+        startLine,
+        endLine,
+        language,
+        contentHash,
+        content,
+        contentPreview: content.slice(0, 200),
+        indexedAt,
+    };
+}
+export function buildChunkHeader(filePath, entityName, entityType, startLine, endLine) {
+    const nameStr = entityName ? ` ${entityName}` : '';
+    return `// [dk-forge] file: ${filePath} | entity:${nameStr} (${entityType}) | lines: ${startLine}-${endLine}\n`;
+}
+function entityTypeToChunkType(type) {
+    const map = {
+        function: 'function',
+        class: 'class',
+        interface: 'interface',
+        type_alias: 'type_alias',
+        variable: 'variable',
+        module: 'file_level',
+        file: 'file_level',
+    };
+    return map[type] ?? 'file_level';
+}
+function detectNonCodeType(ext) {
+    if (ext === '.md' || ext === '.mdx')
+        return 'markdown';
+    if (ext === '.json' || ext === '.yaml' || ext === '.yml' || ext === '.toml')
+        return 'config';
+    if (ext === '.graphql' || ext === '.gql')
+        return 'config';
+    if (ext === '.sql')
+        return 'config';
+    return 'config';
+}
+function getOverlapLines(lines, targetTokens) {
+    let tokens = 0;
+    const result = [];
+    for (let i = lines.length - 1; i >= 0; i--) {
+        const line = lines[i] ?? '';
+        const t = estimateTokens(line);
+        if (tokens + t > targetTokens)
+            break;
+        result.unshift(line);
+        tokens += t;
+    }
+    return result;
+}
+function estimateTokensForLines(lines) {
+    return lines.reduce((sum, line) => sum + estimateTokens(line), 0);
+}
+//# sourceMappingURL=chunker.js.map

package/dist/ingestion/chunker.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"chunker.js","sourceRoot":"","sources":["../../src/ingestion/chunker.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AACpC,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AACpC,OAAO,EAAE,OAAO,EAAE,MAAM,MAAM,CAAC;AAE/B,OAAO,EAAE,cAAc,EAAE,MAAM,0BAA0B,CAAC;AAE1D,MAAM,gBAAgB,GAAG,IAAI,CAAC;AAC9B,MAAM,uBAAuB,GAAG,GAAG,CAAC;AACpC,MAAM,0BAA0B,GAAG,EAAE,CAAC;AAEtC;;;GAGG;AACH,MAAM,UAAU,iBAAiB,CAC/B,QAAwB,EACxB,OAAuB,EACvB,WAAmB,EACnB,QAAgB,EAChB,MAAc,EACd,QAAgB,EAChB,SAAiB;IAEjB,MAAM,MAAM,GAAgB,EAAE,CAAC;IAE/B,qFAAqF;IACrF,MAAM,cAAc,GAAG,mBAAmB,CAAC,OAAO,EAAE,WAAW,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,SAAS,EAAE,QAAQ,CAAC,CAAC;IAClH,IAAI,cAAc;QAAE,MAAM,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;IAEhD,gCAAgC;IAChC,KAAK,MAAM,MAAM,IAAI,QAAQ,EAAE,CAAC;QAC9B,MAAM,YAAY,GAAG,iBAAiB,CAAC,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC;QACtF,MAAM,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC;IAC/B,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CACxB,OAAe,EACf,QAAgB,EAChB,MAAc,EACd,QAAgB,EAChB,SAAiB;IAEjB,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IAC5C,MAAM,SAAS,GAAG,iBAAiB,CAAC,GAAG,CAAC,CAAC;IAEzC,yDAAyD;IACzD,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,MAAM,GAAgB,EAAE,CAAC;IAE/B,IAAI,YAAY,GAAa,EAAE,CAAC;IAChC,IAAI,aAAa,GAAG,CAAC,CAAC;IACtB,IAAI,SAAS,GAAG,CAAC,CAAC;IAElB,MAAM,UAAU,GAAG,CAAC,OAAe,EAAE,EAAE;QACrC,IAAI,YAAY,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QACtC,MAAM,YAAY,GAAG,gBAAgB,CAAC,QAAQ,EAAE,IAAI,EAAE,SAAS,EAAE,SAAS,EAAE,OAAO,CAAC;YAClF,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC1B,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,YAAY,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,SAAS,EAAE,OAAO,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC,CAAC;IAC7G,CAAC,CAAC;IAEF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,UAAU,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;QAExC,IAAI,aAAa,GAAG,UAAU,GAAG,uBAAuB,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACpF,UAAU,CAAC,SAAS,GAAG,YAAY,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;YAEhD,8CAA8C;YAC9C,MAAM,YAAY,GAAG,eAAe,CAAC,YAAY,EAAE,0BAA0B,CAAC,CAAC;YAC/E,SAAS,GAAG,SAAS,GAAG,YAAY,CAAC,MAAM,GAAG,YAAY,CAAC,MAAM,CAAC;YAClE,YAAY,GAAG,YAAY,CAAC;YAC5B,aAAa,GAAG,sBAAsB,CAAC,YAAY,CAAC,CAAC;QACvD,CAAC;QAED,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACxB,aAAa,IAAI,UAAU,CAAC;IAC9B,CAAC;IAED,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,UAAU,CAAC,SAAS,GAAG,YAAY,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAClD,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,mBAAmB,CAC1B,OAAuB,EACvB,WAAmB,EACnB,QAAgB,EAChB,MAAc,EACd,QAAgB,EAChB,SAAiB,EACjB,QAAwB;IAExB,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAE7D,qBAAqB;IACrB,MAAM,WAAW,GAAG,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE;QACpC,IAAI,GAAG,CAAC,SAAS,EAAE,CAAC;YAClB,OAAO,UAAU,GAAG,CAAC,UAAU,UAAU,GAAG,CAAC,QAAQ,IAAI,CAAC;QAC5D,CAAC;QACD,OAAO,UAAU,GAAG,CAAC,UAAU,UAAU,GAAG,CAAC,QAAQ,IAAI,CAAC;IAC5D,CAAC,CAAC,CAAC;IAEH,gEAAgE;IAChE,MAAM,WAAW,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC3C,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE;QAAE,OAAO,IAAI,CAAC;IAErC,MAAM,MAAM,GAAG,gBAAgB,CAAC,QAAQ,EAAE,IAAI,EAAE,YAAY,EAAE,CAAC,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;IACrF,MAAM,OAAO,GAAG,MAAM,GAAG,WAAW,CAAC;IAErC,IAAI,cAAc,CAAC,OAAO,CAAC,GAAG,EAAE;QAAE,OAAO,IAAI,CAAC;IAE9C,OAAO,SAAS,CAAC,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,YAAY,EAAE,CAAC,EAAE,WAAW,CAAC,MAAM,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC;AACxG,CAAC;AAED,SAAS,iBAAiB,CACxB,MAAoB,EACpB,QAAgB,EAChB,MAAc,EACd,QAAgB,EAChB,SAAiB;IAEjB,MAAM,SAAS,GAAG,qBAAqB,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IACrD,MAAM,MAAM,GAAG,gBAAgB,CAAC,QAAQ,EAAE,MAAM,CAAC,IAAI,EAAE,SAAS,EAAE,MAAM,CAAC,SAAS,EAAE,MAAM,CAAC,OAAO,CAAC,CAAC;IACpG,MAAM,OAAO,GAAG,MAAM,GAAG,MAAM,CAAC,UAAU,CAAC;IAE3C,MAAM,MAAM,GAAG,cAAc,CAAC,OAAO,CAAC,CAAC;IAEvC,IAAI,MAAM,IAAI,gBAAgB,EAAE,CAAC;QAC/B,+BAA+B;QAC/B,OAAO,CAAC,SAAS,CAAC,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,CAAC,SAAS,EAAE,MAAM,CAAC,OAAO,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC,CAAC;IAClH,CAAC;IAED,mFAAmF;IACnF,OAAO,gBAAgB,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC;AAC7F,CAAC;AAED,SAAS,gBAAgB,CACvB,OAAe,EACf,MAAoB,EACpB,QAAgB,EAChB,MAAc,EACd,SAAoB,EACpB,QAAgB,EAChB,SAAiB;IAEjB,MAAM,MAAM,GAAgB,EAAE,CAAC;IAC/B,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAClC,IAAI,YAAY,GAAa,EAAE,CAAC;IAChC,IAAI,aAAa,GAAG,CAAC,CAAC;IACtB,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,IAAI,cAAc,GAAG,MAAM,CAAC,SAAS,CAAC;IAEtC,MAAM,KAAK,GAAG,GAAG,EAAE;QACjB,IAAI,YAAY,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QACtC,MAAM,YAAY,GAAG,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7C,MAAM,YAAY,GAAG,cAAc,GAAG,YAAY,CAAC,MAAM,GAAG,CAAC,CAAC;QAC9D,MAAM,CAAC,IAAI,CAAC,SAAS,CACnB,YAAY,EACZ,QAAQ,EAAE,MAAM,EAAE,SAAS,EAC3B,cAAc,EAAE,YAAY,EAC5B,QAAQ,EAAE,SAAS,EACnB,GAAG,MAAM,CAAC,IAAI,QAAQ,UAAU,EAAE,EAAE,CACrC,CAAC,CAAC;QACH,cAAc,GAAG,YAAY,GAAG,CAAC,CAAC;QAClC,YAAY,GAAG,EAAE,CAAC;QAClB,aAAa,GAAG,CAAC,CAAC;IACpB,CAAC,CAAC;IAEF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,UAAU,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;QAExC,IAAI,aAAa,GAAG,UAAU,GAAG,gBAAgB,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC7E,KAAK,EAAE,CAAC;QACV,CAAC;QAED,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACxB,aAAa,IAAI,UAAU,CAAC;IAC9B,CAAC;IAED,KAAK,EAAE,CAAC;IACR,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,MAAM,UAAU,SAAS,CACvB,OAAe,EACf,QAAgB,EAChB,MAAc,EACd,UAAqB,EACrB,SAAiB,EACjB,OAAe,EACf,QAAgB,EAChB,SAAiB,EACjB,UAAmB;IAEnB,MAAM,EAAE,GAAG,UAAU,EAAE,CAAC;IACxB,MAAM,WAAW,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IAE/E,OAAO;QACL,EAAE;QACF,MAAM;QACN,QAAQ;QACR,UAAU,EAAE,UAAU,IAAI,IAAI;QAC9B,UAAU;QACV,SAAS;QACT,OAAO;QACP,QAAQ;QACR,WAAW;QACX,OAAO;QACP,cAAc,EAAE,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC;QACrC,SAAS;KACV,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,gBAAgB,CAC9B,QAAgB,EAChB,UAAyB,EACzB,UAAqB,EACrB,SAAiB,EACjB,OAAe;IAEf,MAAM,OAAO,GAAG,UAAU,CAAC,CAAC,CAAC,IAAI,UAAU,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;IACnD,OAAO,uBAAuB,QAAQ,aAAa,OAAO,KAAK,UAAU,cAAc,SAAS,IAAI,OAAO,IAAI,CAAC;AAClH,CAAC;AAED,SAAS,qBAAqB,CAAC,IAAY;IACzC,MAAM,GAAG,GAA8B;QACrC,QAAQ,EAAE,UAAU;QACpB,KAAK,EAAE,OAAO;QACd,SAAS,EAAE,WAAW;QACtB,UAAU,EAAE,YAAY;QACxB,QAAQ,EAAE,UAAU;QACpB,MAAM,EAAE,YAAY;QACpB,IAAI,EAAE,YAAY;KACnB,CAAC;IACF,OAAO,GAAG,CAAC,IAAI,CAAC,IAAI,YAAY,CAAC;AACnC,CAAC;AAED,SAAS,iBAAiB,CAAC,GAAW;IACpC,IAAI,GAAG,KAAK,KAAK,IAAI,GAAG,KAAK,MAAM;QAAE,OAAO,UAAU,CAAC;IACvD,IAAI,GAAG,KAAK,OAAO,IAAI,GAAG,KAAK,OAAO,IAAI,GAAG,KAAK,MAAM,IAAI,GAAG,KAAK,OAAO;QAAE,OAAO,QAAQ,CAAC;IAC7F,IAAI,GAAG,KAAK,UAAU,IAAI,GAAG,KAAK,MAAM;QAAE,OAAO,QAAQ,CAAC;IAC1D,IAAI,GAAG,KAAK,MAAM;QAAE,OAAO,QAAQ,CAAC;IACpC,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,eAAe,CAAC,KAAe,EAAE,YAAoB;IAC5D,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,MAAM,MAAM,GAAa,EAAE,CAAC;IAE5B,KAAK,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5B,MAAM,CAAC,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;QAC/B,IAAI,MAAM,GAAG,CAAC,GAAG,YAAY;YAAE,MAAM;QACrC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACrB,MAAM,IAAI,CAAC,CAAC;IACd,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,sBAAsB,CAAC,KAAe;IAC7C,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,CAAC,GAAG,GAAG,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;AACpE,CAAC"}

package/dist/ingestion/embedder.d.ts

@@ -0,0 +1,45 @@
+/**
+ * Embedding pipeline using Transformers.js with ONNX backend (ADR-2).
+ * Model: all-MiniLM-L6-v2 (384-dim, ~23 MB quantized, ~5ms/embed).
+ * Singleton pattern - expensive to initialize, reuse across all calls.
+ * Batch embedding: adaptive batch size (32 for long, 64 for short texts).
+ *
+ * Phase 5 changes:
+ *   - Lazy initialization: model loads on first real query, not at startup.
+ *   - isReady() lets callers surface "model loading..." status.
+ *   - Adaptive batching: short texts (<100 chars) use batch size 64.
+ *   - Throughput logging at info level during indexing.
+ */
+/**
+ * Returns true when the embedding pipeline has been initialized and is ready.
+ * False means the first embed call will trigger model load (adds ~3-5 seconds).
+ */
+export declare function isReady(): boolean;
+/**
+ * Embed a single text string.
+ * Returns a 384-dimensional float array.
+ */
+export declare function embedText(text: string): Promise<number[]>;
+/**
+ * Embed multiple texts in adaptive batches.
+ * Short texts (< SHORT_TEXT_THRESHOLD chars average) use batch size 64.
+ * Long texts use batch size 32.
+ * Returns arrays in the same order as input.
+ * Logs embedding throughput (texts/sec) at info level.
+ */
+export declare function embedBatch(texts: string[]): Promise<number[][]>;
+/**
+ * Get the embedding dimension.
+ */
+export declare function getEmbeddingDimension(): number;
+/**
+ * Optionally pre-warm the embedding pipeline in the background.
+ *
+ * Phase 5: This is now a no-op at the call site — initialization is lazy.
+ * The model will load on the first real embed call instead.
+ * This function is kept for backward compatibility; callers that previously
+ * awaited it will continue to work. The model loads asynchronously without
+ * blocking the MCP server startup.
+ */
+export declare function warmupEmbedder(): Promise<void>;
+//# sourceMappingURL=embedder.d.ts.map

package/dist/ingestion/embedder.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"embedder.d.ts","sourceRoot":"","sources":["../../src/ingestion/embedder.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAeH;;;GAGG;AACH,wBAAgB,OAAO,IAAI,OAAO,CAEjC;AAkDD;;;GAGG;AACH,wBAAsB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,CAI/D;AAED;;;;;;GAMG;AACH,wBAAsB,UAAU,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CA6CrE;AAED;;GAEG;AACH,wBAAgB,qBAAqB,IAAI,MAAM,CAE9C;AAED;;;;;;;;GAQG;AACH,wBAAsB,cAAc,IAAI,OAAO,CAAC,IAAI,CAAC,CAIpD"}

package/dist/ingestion/embedder.js

@@ -0,0 +1,152 @@
+/**
+ * Embedding pipeline using Transformers.js with ONNX backend (ADR-2).
+ * Model: all-MiniLM-L6-v2 (384-dim, ~23 MB quantized, ~5ms/embed).
+ * Singleton pattern - expensive to initialize, reuse across all calls.
+ * Batch embedding: adaptive batch size (32 for long, 64 for short texts).
+ *
+ * Phase 5 changes:
+ *   - Lazy initialization: model loads on first real query, not at startup.
+ *   - isReady() lets callers surface "model loading..." status.
+ *   - Adaptive batching: short texts (<100 chars) use batch size 64.
+ *   - Throughput logging at info level during indexing.
+ */
+import { logger } from '../util/logger.js';
+const BASE_BATCH_SIZE = 32;
+const SHORT_TEXT_BATCH_SIZE = 64;
+const SHORT_TEXT_THRESHOLD = 100; // chars
+const MODEL_NAME = 'Xenova/all-MiniLM-L6-v2';
+const EMBEDDING_DIMENSION = 384;
+// eslint-disable-next-line @typescript-eslint/no-explicit-any
+let pipelineInstance = null;
+let initPromise = null;
+let modelReady = false;
+/**
+ * Returns true when the embedding pipeline has been initialized and is ready.
+ * False means the first embed call will trigger model load (adds ~3-5 seconds).
+ */
+export function isReady() {
+    return modelReady;
+}
+/**
+ * Initialize the embedding pipeline (singleton).
+ * Lazy: only called on first embed request, not at server startup.
+ * Downloads model on first use, uses cached copy thereafter.
+ */
+async function ensurePipeline() {
+    if (pipelineInstance)
+        return;
+    if (initPromise)
+        return initPromise;
+    initPromise = (async () => {
+        try {
+            logger.info('Initializing embedding model (lazy load)', { model: MODEL_NAME });
+            const { pipeline, env } = await import('@huggingface/transformers');
+            // Configure cache location
+            const cacheDir = process.env.TRANSFORMERS_CACHE ?? undefined;
+            if (cacheDir) {
+                env.cacheDir = cacheDir;
+            }
+            // Use ONNX backend (default in Transformers.js)
+            // eslint-disable-next-line @typescript-eslint/no-explicit-any
+            pipelineInstance = await pipeline('feature-extraction', MODEL_NAME, {
+                dtype: 'q8', // Quantized INT8 for smaller memory footprint
+            });
+            modelReady = true;
+            logger.info('Embedding model initialized', { model: MODEL_NAME, dimension: EMBEDDING_DIMENSION });
+        }
+        catch (err) {
+            logger.error('Failed to initialize embedding model', { error: String(err) });
+            initPromise = null;
+            throw err;
+        }
+    })();
+    return initPromise;
+}
+/**
+ * Choose optimal batch size based on average text length.
+ * Short texts (<100 chars on average) pack more efficiently at batch 64.
+ */
+function chooseBatchSize(texts) {
+    if (texts.length === 0)
+        return BASE_BATCH_SIZE;
+    const avgLen = texts.reduce((s, t) => s + t.length, 0) / texts.length;
+    return avgLen < SHORT_TEXT_THRESHOLD ? SHORT_TEXT_BATCH_SIZE : BASE_BATCH_SIZE;
+}
+/**
+ * Embed a single text string.
+ * Returns a 384-dimensional float array.
+ */
+export async function embedText(text) {
+    await ensurePipeline();
+    const result = await pipelineInstance(text, { pooling: 'mean', normalize: true });
+    return Array.from(result.data);
+}
+/**
+ * Embed multiple texts in adaptive batches.
+ * Short texts (< SHORT_TEXT_THRESHOLD chars average) use batch size 64.
+ * Long texts use batch size 32.
+ * Returns arrays in the same order as input.
+ * Logs embedding throughput (texts/sec) at info level.
+ */
+export async function embedBatch(texts) {
+    if (texts.length === 0)
+        return [];
+    await ensurePipeline();
+    const batchSize = chooseBatchSize(texts);
+    const results = [];
+    const startTime = Date.now();
+    for (let i = 0; i < texts.length; i += batchSize) {
+        const batch = texts.slice(i, i + batchSize);
+        try {
+            const output = await pipelineInstance(batch, { pooling: 'mean', normalize: true });
+            // output.data is a flat Float32Array: [batch_size * dim]
+            const data = output.data;
+            const dim = data.length / batch.length;
+            for (let j = 0; j < batch.length; j++) {
+                const vector = Array.from(data.slice(j * dim, (j + 1) * dim));
+                results.push(vector);
+            }
+        }
+        catch (err) {
+            logger.warn('Embedding batch failed, using zero vectors', {
+                batchStart: i,
+                batchSize: batch.length,
+                error: String(err),
+            });
+            // Fallback: push zero vectors for failed batch
+            for (let j = 0; j < batch.length; j++) {
+                results.push(new Array(EMBEDDING_DIMENSION).fill(0));
+            }
+        }
+    }
+    const elapsedMs = Date.now() - startTime;
+    const throughput = elapsedMs > 0 ? Math.round((texts.length / elapsedMs) * 1000) : texts.length;
+    logger.info('Embedding batch complete', {
+        totalTexts: texts.length,
+        batchSize,
+        elapsedMs,
+        throughputTextsPerSec: throughput,
+    });
+    return results;
+}
+/**
+ * Get the embedding dimension.
+ */
+export function getEmbeddingDimension() {
+    return EMBEDDING_DIMENSION;
+}
+/**
+ * Optionally pre-warm the embedding pipeline in the background.
+ *
+ * Phase 5: This is now a no-op at the call site — initialization is lazy.
+ * The model will load on the first real embed call instead.
+ * This function is kept for backward compatibility; callers that previously
+ * awaited it will continue to work. The model loads asynchronously without
+ * blocking the MCP server startup.
+ */
+export async function warmupEmbedder() {
+    // Do not eagerly load — lazy initialization is handled by ensurePipeline().
+    // Log a message so callers know the model will warm on first use.
+    logger.info('Embedding pipeline will warm on first use (lazy mode)', { model: MODEL_NAME });
+}
+//# sourceMappingURL=embedder.js.map

package/dist/ingestion/embedder.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"embedder.js","sourceRoot":"","sources":["../../src/ingestion/embedder.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C,MAAM,eAAe,GAAG,EAAE,CAAC;AAC3B,MAAM,qBAAqB,GAAG,EAAE,CAAC;AACjC,MAAM,oBAAoB,GAAG,GAAG,CAAC,CAAC,QAAQ;AAC1C,MAAM,UAAU,GAAG,yBAAyB,CAAC;AAC7C,MAAM,mBAAmB,GAAG,GAAG,CAAC;AAEhC,8DAA8D;AAC9D,IAAI,gBAAgB,GAAQ,IAAI,CAAC;AACjC,IAAI,WAAW,GAAyB,IAAI,CAAC;AAC7C,IAAI,UAAU,GAAG,KAAK,CAAC;AAEvB;;;GAGG;AACH,MAAM,UAAU,OAAO;IACrB,OAAO,UAAU,CAAC;AACpB,CAAC;AAED;;;;GAIG;AACH,KAAK,UAAU,cAAc;IAC3B,IAAI,gBAAgB;QAAE,OAAO;IAC7B,IAAI,WAAW;QAAE,OAAO,WAAW,CAAC;IAEpC,WAAW,GAAG,CAAC,KAAK,IAAI,EAAE;QACxB,IAAI,CAAC;YACH,MAAM,CAAC,IAAI,CAAC,0CAA0C,EAAE,EAAE,KAAK,EAAE,UAAU,EAAE,CAAC,CAAC;YAC/E,MAAM,EAAE,QAAQ,EAAE,GAAG,EAAE,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAC;YAEpE,2BAA2B;YAC3B,MAAM,QAAQ,GAAG,OAAO,CAAC,GAAG,CAAC,kBAAkB,IAAI,SAAS,CAAC;YAC7D,IAAI,QAAQ,EAAE,CAAC;gBACb,GAAG,CAAC,QAAQ,GAAG,QAAQ,CAAC;YAC1B,CAAC;YAED,gDAAgD;YAChD,8DAA8D;YAC9D,gBAAgB,GAAG,MAAO,QAAgB,CAAC,oBAAoB,EAAE,UAAU,EAAE;gBAC3E,KAAK,EAAE,IAAI,EAAG,8CAA8C;aAC7D,CAAC,CAAC;YAEH,UAAU,GAAG,IAAI,CAAC;YAClB,MAAM,CAAC,IAAI,CAAC,6BAA6B,EAAE,EAAE,KAAK,EAAE,UAAU,EAAE,SAAS,EAAE,mBAAmB,EAAE,CAAC,CAAC;QACpG,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,CAAC,KAAK,CAAC,sCAAsC,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;YAC7E,WAAW,GAAG,IAAI,CAAC;YACnB,MAAM,GAAG,CAAC;QACZ,CAAC;IACH,CAAC,CAAC,EAAE,CAAC;IAEL,OAAO,WAAW,CAAC;AACrB,CAAC;AAED;;;GAGG;AACH,SAAS,eAAe,CAAC,KAAe;IACtC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,eAAe,CAAC;IAC/C,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC;IACtE,OAAO,MAAM,GAAG,oBAAoB,CAAC,CAAC,CAAC,qBAAqB,CAAC,CAAC,CAAC,eAAe,CAAC;AACjF,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,IAAY;IAC1C,MAAM,cAAc,EAAE,CAAC;IACvB,MAAM,MAAM,GAAG,MAAM,gBAAgB,CAAC,IAAI,EAAE,EAAE,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAClF,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAoB,CAAC,CAAC;AACjD,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,KAAe;IAC9C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAClC,MAAM,cAAc,EAAE,CAAC;IAEvB,MAAM,SAAS,GAAG,eAAe,CAAC,KAAK,CAAC,CAAC;IACzC,MAAM,OAAO,GAAe,EAAE,CAAC;IAC/B,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAE7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;QACjD,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC;QAE5C,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,MAAM,gBAAgB,CAAC,KAAK,EAAE,EAAE,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;YAEnF,yDAAyD;YACzD,MAAM,IAAI,GAAG,MAAM,CAAC,IAAoB,CAAC;YACzC,MAAM,GAAG,GAAG,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC;YAEvC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACtC,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;gBAC9D,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACvB,CAAC;QACH,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,CAAC,IAAI,CAAC,4CAA4C,EAAE;gBACxD,UAAU,EAAE,CAAC;gBACb,SAAS,EAAE,KAAK,CAAC,MAAM;gBACvB,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC;aACnB,CAAC,CAAC;YACH,+CAA+C;YAC/C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACtC,OAAO,CAAC,IAAI,CAAC,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;YACvD,CAAC;QACH,CAAC;IACH,CAAC;IAED,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,SAAS,CAAC;IACzC,MAAM,UAAU,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,SAAS,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC;IAChG,MAAM,CAAC,IAAI,CAAC,0BAA0B,EAAE;QACtC,UAAU,EAAE,KAAK,CAAC,MAAM;QACxB,SAAS;QACT,SAAS;QACT,qBAAqB,EAAE,UAAU;KAClC,CAAC,CAAC;IAEH,OAAO,OAAO,CAAC;AACjB,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,qBAAqB;IACnC,OAAO,mBAAmB,CAAC;AAC7B,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,CAAC,KAAK,UAAU,cAAc;IAClC,4EAA4E;IAC5E,kEAAkE;IAClE,MAAM,CAAC,IAAI,CAAC,uDAAuD,EAAE,EAAE,KAAK,EAAE,UAAU,EAAE,CAAC,CAAC;AAC9F,CAAC"}

package/dist/ingestion/git-analyzer.d.ts

@@ -0,0 +1,77 @@
+/**
+ * Git history analyzer for co-modification edges (Phase 2 full implementation).
+ * Runs `git log --name-only` to find files changed together frequently,
+ * creates CO_MODIFIED edges, and detects test file relationships.
+ *
+ * Phase 3 additions:
+ *   - analyzeFileStats(): per-file commit count, stability, velocity
+ *   - extractCommitRecords(): parse git log into CommitRecord[]
+ */
+import type { GraphStore } from '../storage/interfaces.js';
+import type { FileChangeStats, CommitRecord } from '../util/types.js';
+export interface CoModificationPair {
+    fileA: string;
+    fileB: string;
+    count: number;
+    lastCommit: string;
+}
+export interface GitAnalysisResult {
+    coModifications: CoModificationPair[];
+    testPairs: Array<{
+        testFile: string;
+        sourceFile: string;
+    }>;
+    totalCommitsAnalyzed: number;
+    errors: string[];
+}
+/**
+ * Analyze git history to find files frequently changed together.
+ * Parses `git log --name-only` output for the last `historyDepth` commits,
+ * counts co-occurrence pairs, and returns pairs with count >= minComodCount.
+ */
+export declare function analyzeCoModifications(repoPath: string, historyDepth?: number, minComodCount?: number): Promise<CoModificationPair[]>;
+/**
+ * Parse `git log --name-only --format="%H"` output into commit groups.
+ * Each group contains a commit hash and the list of files changed.
+ */
+export declare function parseGitLog(stdout: string): Array<{
+    hash: string;
+    files: string[];
+}>;
+/**
+ * Detect test files and their corresponding source files.
+ * Returns pairs: { testFile, sourceFile }
+ *
+ * Handles patterns:
+ *   foo.spec.ts  -> foo.ts
+ *   foo.test.ts  -> foo.ts
+ *   __tests__/foo.ts -> ../foo.ts (sibling detection)
+ */
+export declare function detectTestFiles(files: string[]): Promise<Array<{
+    testFile: string;
+    sourceFile: string;
+}>>;
+/**
+ * Run full git analysis for a repo: co-modifications + test file detection.
+ * Then write the edges to FalkorDB.
+ */
+export declare function runGitAnalysis(repoPath: string, repoId: string, allFiles: string[], graphStore: GraphStore, historyDepth?: number, minComodCount?: number): Promise<GitAnalysisResult>;
+/**
+ * Analyze per-file change statistics from git history.
+ * Runs a second git log with `%H|%s|%an|%aI` format (hash, subject, author, ISO date)
+ * plus `--name-only` to map commits to files.
+ *
+ * Computes per-file:
+ *   - commitCount: total commits touching this file
+ *   - stabilityScore: 0-1, higher = fewer recent changes relative to history
+ *   - changeVelocity: -1 to +1, positive = accelerating changes
+ *   - lastCommitHash, lastCommitTs, recentMessages (last 5)
+ */
+export declare function analyzeFileStats(repoPath: string, historyDepth?: number): Promise<Map<string, FileChangeStats>>;
+/**
+ * Extract commit records from git history.
+ * Parses git log into CommitRecord[], filtering merge commits and
+ * bulk reformats (commits touching >50 files).
+ */
+export declare function extractCommitRecords(repoPath: string, historyDepth?: number): Promise<CommitRecord[]>;
+//# sourceMappingURL=git-analyzer.d.ts.map

package/dist/ingestion/git-analyzer.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"git-analyzer.d.ts","sourceRoot":"","sources":["../../src/ingestion/git-analyzer.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAKH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,0BAA0B,CAAC;AAC3D,OAAO,KAAK,EAAE,eAAe,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAKtE,MAAM,WAAW,kBAAkB;IACjC,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,iBAAiB;IAChC,eAAe,EAAE,kBAAkB,EAAE,CAAC;IACtC,SAAS,EAAE,KAAK,CAAC;QAAE,QAAQ,EAAE,MAAM,CAAC;QAAC,UAAU,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAC3D,oBAAoB,EAAE,MAAM,CAAC;IAC7B,MAAM,EAAE,MAAM,EAAE,CAAC;CAClB;AAED;;;;GAIG;AACH,wBAAsB,sBAAsB,CAC1C,QAAQ,EAAE,MAAM,EAChB,YAAY,GAAE,MAAY,EAC1B,aAAa,GAAE,MAAU,GACxB,OAAO,CAAC,kBAAkB,EAAE,CAAC,CA2D/B;AAED;;;GAGG;AACH,wBAAgB,WAAW,CACzB,MAAM,EAAE,MAAM,GACb,KAAK,CAAC;IAAE,IAAI,EAAE,MAAM,CAAC;IAAC,KAAK,EAAE,MAAM,EAAE,CAAA;CAAE,CAAC,CAwC1C;AAED;;;;;;;;GAQG;AACH,wBAAsB,eAAe,CACnC,KAAK,EAAE,MAAM,EAAE,GACd,OAAO,CAAC,KAAK,CAAC;IAAE,QAAQ,EAAE,MAAM,CAAC;IAAC,UAAU,EAAE,MAAM,CAAA;CAAE,CAAC,CAAC,CAuC1D;AAED;;;GAGG;AACH,wBAAsB,cAAc,CAClC,QAAQ,EAAE,MAAM,EAChB,MAAM,EAAE,MAAM,EACd,QAAQ,EAAE,MAAM,EAAE,EAClB,UAAU,EAAE,UAAU,EACtB,YAAY,GAAE,MAAY,EAC1B,aAAa,GAAE,MAAU,GACxB,OAAO,CAAC,iBAAiB,CAAC,CAoE5B;AAMD;;;;;;;;;;GAUG;AACH,wBAAsB,gBAAgB,CACpC,QAAQ,EAAE,MAAM,EAChB,YAAY,GAAE,MAAY,GACzB,OAAO,CAAC,GAAG,CAAC,MAAM,EAAE,eAAe,CAAC,CAAC,CAiFvC;AAED;;;;GAIG;AACH,wBAAsB,oBAAoB,CACxC,QAAQ,EAAE,MAAM,EAChB,YAAY,GAAE,MAAY,GACzB,OAAO,CAAC,YAAY,EAAE,CAAC,CAwCzB"}