tea-rags 0.9.0

package/build/core/ingest/pipeline/chunker/utils/import-extractor.js

@@ -0,0 +1,47 @@
+/**
+ * Regex-based import/export extraction from source code.
+ */
+export function extractImportsExports(code, language) {
+    const imports = [];
+    const exports = [];
+    if (language === "typescript" || language === "javascript") {
+        // Extract imports
+        const importMatches = code.matchAll(/import\s+.*?\s+from\s+['"]([^'"]+)['"]/g);
+        for (const match of importMatches) {
+            imports.push(match[1]);
+        }
+        // Extract require statements
+        const requireMatches = code.matchAll(/require\s*\(\s*['"]([^'"]+)['"]\s*\)/g);
+        for (const match of requireMatches) {
+            imports.push(match[1]);
+        }
+        // Extract exports - regular declarations
+        const exportMatches = code.matchAll(/export\s+(?:class|function|const|let|var)\s+(\w+)/g);
+        for (const match of exportMatches) {
+            exports.push(match[1]);
+        }
+        // Extract export default
+        if (/export\s+default\b/.test(code)) {
+            exports.push("default");
+        }
+        // Extract named exports from other modules: export { name } from 'module'
+        const reExportMatches = code.matchAll(/export\s+\{\s*(\w+)\s*\}/g);
+        for (const match of reExportMatches) {
+            exports.push(match[1]);
+        }
+    }
+    else if (language === "python") {
+        // Extract imports
+        const importMatches = code.matchAll(/(?:from\s+(\S+)\s+)?import\s+([^;\n]+)/g);
+        for (const match of importMatches) {
+            imports.push(match[1] || match[2]);
+        }
+        // Extract functions/classes (rough)
+        const defMatches = code.matchAll(/^(?:def|class)\s+(\w+)/gm);
+        for (const match of defMatches) {
+            exports.push(match[1]);
+        }
+    }
+    return { imports, exports };
+}
+//# sourceMappingURL=import-extractor.js.map

package/build/core/ingest/pipeline/chunker/utils/import-extractor.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"import-extractor.js","sourceRoot":"","sources":["../../../../../../src/core/ingest/pipeline/chunker/utils/import-extractor.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,MAAM,UAAU,qBAAqB,CACnC,IAAY,EACZ,QAAgB;IAKhB,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,OAAO,GAAa,EAAE,CAAC;IAE7B,IAAI,QAAQ,KAAK,YAAY,IAAI,QAAQ,KAAK,YAAY,EAAE,CAAC;QAC3D,kBAAkB;QAClB,MAAM,aAAa,GAAG,IAAI,CAAC,QAAQ,CAAC,yCAAyC,CAAC,CAAC;QAC/E,KAAK,MAAM,KAAK,IAAI,aAAa,EAAE,CAAC;YAClC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACzB,CAAC;QAED,6BAA6B;QAC7B,MAAM,cAAc,GAAG,IAAI,CAAC,QAAQ,CAAC,uCAAuC,CAAC,CAAC;QAC9E,KAAK,MAAM,KAAK,IAAI,cAAc,EAAE,CAAC;YACnC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACzB,CAAC;QAED,yCAAyC;QACzC,MAAM,aAAa,GAAG,IAAI,CAAC,QAAQ,CAAC,oDAAoD,CAAC,CAAC;QAC1F,KAAK,MAAM,KAAK,IAAI,aAAa,EAAE,CAAC;YAClC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACzB,CAAC;QAED,yBAAyB;QACzB,IAAI,oBAAoB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACpC,OAAO,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAC1B,CAAC;QAED,0EAA0E;QAC1E,MAAM,eAAe,GAAG,IAAI,CAAC,QAAQ,CAAC,2BAA2B,CAAC,CAAC;QACnE,KAAK,MAAM,KAAK,IAAI,eAAe,EAAE,CAAC;YACpC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;SAAM,IAAI,QAAQ,KAAK,QAAQ,EAAE,CAAC;QACjC,kBAAkB;QAClB,MAAM,aAAa,GAAG,IAAI,CAAC,QAAQ,CAAC,yCAAyC,CAAC,CAAC;QAC/E,KAAK,MAAM,KAAK,IAAI,aAAa,EAAE,CAAC;YAClC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACrC,CAAC;QAED,oCAAoC;QACpC,MAAM,UAAU,GAAG,IAAI,CAAC,QAAQ,CAAC,0BAA0B,CAAC,CAAC;QAC7D,KAAK,MAAM,KAAK,IAAI,UAAU,EAAE,CAAC;YAC/B,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;IAED,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,CAAC;AAC9B,CAAC"}

package/build/core/ingest/pipeline/chunker/utils/language-detector.d.ts

@@ -0,0 +1,5 @@
+/**
+ * Maps file extensions to language identifiers.
+ */
+export declare function detectLanguage(filePath: string): string;
+//# sourceMappingURL=language-detector.d.ts.map

package/build/core/ingest/pipeline/chunker/utils/language-detector.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-detector.d.ts","sourceRoot":"","sources":["../../../../../../src/core/ingest/pipeline/chunker/utils/language-detector.ts"],"names":[],"mappings":"AAAA;;GAEG;AAMH,wBAAgB,cAAc,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,CAGvD"}

package/build/core/ingest/pipeline/chunker/utils/language-detector.js

@@ -0,0 +1,10 @@
+/**
+ * Maps file extensions to language identifiers.
+ */
+import { extname } from "node:path";
+import { LANGUAGE_MAP } from "../config.js";
+export function detectLanguage(filePath) {
+    const ext = extname(filePath);
+    return LANGUAGE_MAP[ext] || "unknown";
+}
+//# sourceMappingURL=language-detector.js.map

package/build/core/ingest/pipeline/chunker/utils/language-detector.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-detector.js","sourceRoot":"","sources":["../../../../../../src/core/ingest/pipeline/chunker/utils/language-detector.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAEpC,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAE5C,MAAM,UAAU,cAAc,CAAC,QAAgB;IAC7C,MAAM,GAAG,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC;IAC9B,OAAO,YAAY,CAAC,GAAG,CAAC,IAAI,SAAS,CAAC;AACxC,CAAC"}

package/build/core/ingest/pipeline/chunker/utils/metadata.d.ts

@@ -0,0 +1,15 @@
+/**
+ * MetadataExtractor - Thin facade delegating to focused utilities.
+ */
+import type { CodeChunk } from "../../../../types.js";
+export declare class MetadataExtractor {
+    extractLanguage(filePath: string): string;
+    generateChunkId(chunk: CodeChunk): string;
+    calculateComplexity(code: string): number;
+    containsSecrets(code: string): boolean;
+    extractImportsExports(code: string, language: string): {
+        imports: string[];
+        exports: string[];
+    };
+}
+//# sourceMappingURL=metadata.d.ts.map

package/build/core/ingest/pipeline/chunker/utils/metadata.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"metadata.d.ts","sourceRoot":"","sources":["../../../../../../src/core/ingest/pipeline/chunker/utils/metadata.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,sBAAsB,CAAC;AAOtD,qBAAa,iBAAiB;IAC5B,eAAe,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM;IAIzC,eAAe,CAAC,KAAK,EAAE,SAAS,GAAG,MAAM;IAIzC,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM;IAIzC,eAAe,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO;IAItC,qBAAqB,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,GAAG;QAAE,OAAO,EAAE,MAAM,EAAE,CAAC;QAAC,OAAO,EAAE,MAAM,EAAE,CAAA;KAAE;CAGhG"}

package/build/core/ingest/pipeline/chunker/utils/metadata.js

@@ -0,0 +1,26 @@
+/**
+ * MetadataExtractor - Thin facade delegating to focused utilities.
+ */
+import { generateChunkId } from "./chunk-id.js";
+import { calculateComplexity } from "./complexity.js";
+import { extractImportsExports } from "./import-extractor.js";
+import { detectLanguage } from "./language-detector.js";
+import { containsSecrets } from "./secrets-detector.js";
+export class MetadataExtractor {
+    extractLanguage(filePath) {
+        return detectLanguage(filePath);
+    }
+    generateChunkId(chunk) {
+        return generateChunkId(chunk);
+    }
+    calculateComplexity(code) {
+        return calculateComplexity(code);
+    }
+    containsSecrets(code) {
+        return containsSecrets(code);
+    }
+    extractImportsExports(code, language) {
+        return extractImportsExports(code, language);
+    }
+}
+//# sourceMappingURL=metadata.js.map

package/build/core/ingest/pipeline/chunker/utils/metadata.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"metadata.js","sourceRoot":"","sources":["../../../../../../src/core/ingest/pipeline/chunker/utils/metadata.ts"],"names":[],"mappings":"AAAA;;GAEG;AAGH,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AAChD,OAAO,EAAE,mBAAmB,EAAE,MAAM,iBAAiB,CAAC;AACtD,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAC9D,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AAExD,MAAM,OAAO,iBAAiB;IAC5B,eAAe,CAAC,QAAgB;QAC9B,OAAO,cAAc,CAAC,QAAQ,CAAC,CAAC;IAClC,CAAC;IAED,eAAe,CAAC,KAAgB;QAC9B,OAAO,eAAe,CAAC,KAAK,CAAC,CAAC;IAChC,CAAC;IAED,mBAAmB,CAAC,IAAY;QAC9B,OAAO,mBAAmB,CAAC,IAAI,CAAC,CAAC;IACnC,CAAC;IAED,eAAe,CAAC,IAAY;QAC1B,OAAO,eAAe,CAAC,IAAI,CAAC,CAAC;IAC/B,CAAC;IAED,qBAAqB,CAAC,IAAY,EAAE,QAAgB;QAClD,OAAO,qBAAqB,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;IAC/C,CAAC;CACF"}

package/build/core/ingest/pipeline/chunker/utils/secrets-detector.d.ts

@@ -0,0 +1,5 @@
+/**
+ * Basic secret detection in source code via pattern matching.
+ */
+export declare function containsSecrets(code: string): boolean;
+//# sourceMappingURL=secrets-detector.d.ts.map

package/build/core/ingest/pipeline/chunker/utils/secrets-detector.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"secrets-detector.d.ts","sourceRoot":"","sources":["../../../../../../src/core/ingest/pipeline/chunker/utils/secrets-detector.ts"],"names":[],"mappings":"AAAA;;GAEG;AAcH,wBAAgB,eAAe,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAOrD"}

package/build/core/ingest/pipeline/chunker/utils/secrets-detector.js

@@ -0,0 +1,23 @@
+/**
+ * Basic secret detection in source code via pattern matching.
+ */
+const SECRET_PATTERNS = [
+    /(?:api[-_]?key|apikey)\s*=\s*['"][^'"]{20,}['"]/i,
+    /(?:secret|SECRET)\s*=\s*['"][^'"]{20,}['"]/i,
+    /(?:password|PASSWORD)\s*=\s*['"][^'"]{8,}['"]/i,
+    /(?:token|TOKEN)\s*=\s*['"][^'"]{20,}['"]/i,
+    /-----BEGIN\s+(RSA\s+)?PRIVATE\s+KEY-----/,
+    /sk_live_[a-zA-Z0-9]{24,}/, // Stripe secret key
+    /ghp_[a-zA-Z0-9]{36,}/, // GitHub personal access token
+    /AIza[0-9A-Za-z\\-_]{35}/, // Google API key
+    /AKIA[0-9A-Z]{16}/, // AWS access key
+];
+export function containsSecrets(code) {
+    for (const pattern of SECRET_PATTERNS) {
+        if (pattern.test(code)) {
+            return true;
+        }
+    }
+    return false;
+}
+//# sourceMappingURL=secrets-detector.js.map

package/build/core/ingest/pipeline/chunker/utils/secrets-detector.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"secrets-detector.js","sourceRoot":"","sources":["../../../../../../src/core/ingest/pipeline/chunker/utils/secrets-detector.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,MAAM,eAAe,GAAG;IACtB,kDAAkD;IAClD,6CAA6C;IAC7C,gDAAgD;IAChD,2CAA2C;IAC3C,0CAA0C;IAC1C,0BAA0B,EAAE,oBAAoB;IAChD,sBAAsB,EAAE,+BAA+B;IACvD,yBAAyB,EAAE,iBAAiB;IAC5C,kBAAkB,EAAE,iBAAiB;CACtC,CAAC;AAEF,MAAM,UAAU,eAAe,CAAC,IAAY;IAC1C,KAAK,MAAM,OAAO,IAAI,eAAe,EAAE,CAAC;QACtC,IAAI,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACvB,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC"}

package/build/core/ingest/pipeline/enrichment/applier.d.ts

@@ -0,0 +1,37 @@
+/**
+ * EnrichmentApplier — provider-agnostic payload writer for Qdrant.
+ *
+ * Writes enrichment data under nested structure:
+ * - File-level: { [providerKey]: { file: data } }
+ * - Chunk-level: { [providerKey]: { chunk: overlay } }
+ *
+ * Replaces git-specific MetadataApplier + chunk-churn write logic.
+ */
+import type { QdrantManager } from "../../../adapters/qdrant/client.js";
+import type { ChunkSignalOverlay, FileSignalOverlay, FileSignalTransform } from "../../../contracts/types/provider.js";
+import type { ChunkItem } from "../types.js";
+export declare class EnrichmentApplier {
+    private readonly qdrant;
+    matchedFiles: number;
+    missedFiles: number;
+    readonly missedPathSamples: string[];
+    readonly missedFileChunks: Map<string, {
+        chunkId: string;
+        endLine: number;
+    }[]>;
+    constructor(qdrant: QdrantManager);
+    /**
+     * Apply file-level signals to a batch of chunks.
+     * Payload written as { [providerKey]: { file: data } }.
+     *
+     * @param transform Optional per-file transform called with (rawData, maxEndLine).
+     *   Git uses this for computeFileSignals(churnData, maxEndLine).
+     */
+    applyFileSignals(collectionName: string, providerKey: string, fileMetadata: Map<string, FileSignalOverlay>, pathBase: string, items: ChunkItem[], transform?: FileSignalTransform): Promise<void>;
+    /**
+     * Apply chunk-level signal overlays.
+     * Payload written as { [providerKey]: { chunk: overlay } }.
+     */
+    applyChunkSignals(collectionName: string, providerKey: string, chunkMetadata: Map<string, Map<string, ChunkSignalOverlay>>): Promise<number>;
+}
+//# sourceMappingURL=applier.d.ts.map

package/build/core/ingest/pipeline/enrichment/applier.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"applier.d.ts","sourceRoot":"","sources":["../../../../../src/core/ingest/pipeline/enrichment/applier.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAIH,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,oCAAoC,CAAC;AACxE,OAAO,KAAK,EAAE,kBAAkB,EAAE,iBAAiB,EAAE,mBAAmB,EAAE,MAAM,sCAAsC,CAAC;AAGvH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,aAAa,CAAC;AAI7C,qBAAa,iBAAiB;IAMhB,OAAO,CAAC,QAAQ,CAAC,MAAM;IALnC,YAAY,SAAK;IACjB,WAAW,SAAK;IAChB,QAAQ,CAAC,iBAAiB,EAAE,MAAM,EAAE,CAAM;IAC1C,QAAQ,CAAC,gBAAgB;iBAA8B,MAAM;iBAAW,MAAM;SAAQ;gBAEzD,MAAM,EAAE,aAAa;IAElD;;;;;;OAMG;IACG,gBAAgB,CACpB,cAAc,EAAE,MAAM,EACtB,WAAW,EAAE,MAAM,EACnB,YAAY,EAAE,GAAG,CAAC,MAAM,EAAE,iBAAiB,CAAC,EAC5C,QAAQ,EAAE,MAAM,EAChB,KAAK,EAAE,SAAS,EAAE,EAClB,SAAS,CAAC,EAAE,mBAAmB,GAC9B,OAAO,CAAC,IAAI,CAAC;IA2DhB;;;OAGG;IACG,iBAAiB,CACrB,cAAc,EAAE,MAAM,EACtB,WAAW,EAAE,MAAM,EACnB,aAAa,EAAE,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,kBAAkB,CAAC,CAAC,GAC1D,OAAO,CAAC,MAAM,CAAC;CAyCnB"}

package/build/core/ingest/pipeline/enrichment/applier.js

@@ -0,0 +1,120 @@
+/**
+ * EnrichmentApplier — provider-agnostic payload writer for Qdrant.
+ *
+ * Writes enrichment data under nested structure:
+ * - File-level: { [providerKey]: { file: data } }
+ * - Chunk-level: { [providerKey]: { chunk: overlay } }
+ *
+ * Replaces git-specific MetadataApplier + chunk-churn write logic.
+ */
+import { relative } from "node:path";
+import { pipelineLog } from "../infra/debug-logger.js";
+import { isDebug } from "../infra/runtime.js";
+const BATCH_SIZE = 100;
+export class EnrichmentApplier {
+    qdrant;
+    matchedFiles = 0;
+    missedFiles = 0;
+    missedPathSamples = [];
+    missedFileChunks = new Map();
+    constructor(qdrant) {
+        this.qdrant = qdrant;
+    }
+    /**
+     * Apply file-level signals to a batch of chunks.
+     * Payload written as { [providerKey]: { file: data } }.
+     *
+     * @param transform Optional per-file transform called with (rawData, maxEndLine).
+     *   Git uses this for computeFileSignals(churnData, maxEndLine).
+     */
+    async applyFileSignals(collectionName, providerKey, fileMetadata, pathBase, items, transform) {
+        const applyStart = Date.now();
+        // Group items by filePath
+        const byFile = new Map();
+        for (const item of items) {
+            const fp = item.chunk.metadata.filePath;
+            const existing = byFile.get(fp) || [];
+            existing.push(item);
+            byFile.set(fp, existing);
+        }
+        const operations = [];
+        for (const [filePath, fileItems] of byFile) {
+            const relativePath = relative(pathBase, filePath);
+            const data = fileMetadata.get(relativePath);
+            if (!data) {
+                this.missedFiles++;
+                if (this.missedPathSamples.length < 10) {
+                    this.missedPathSamples.push(relativePath);
+                }
+                const existing = this.missedFileChunks.get(relativePath) || [];
+                for (const item of fileItems) {
+                    existing.push({ chunkId: item.chunkId, endLine: item.chunk.endLine });
+                }
+                this.missedFileChunks.set(relativePath, existing);
+                continue;
+            }
+            this.matchedFiles++;
+            const maxEndLine = fileItems.reduce((max, item) => Math.max(max, item.chunk.endLine), 0);
+            const finalData = transform ? transform(data, maxEndLine) : data;
+            const payload = { [providerKey]: { file: finalData } };
+            for (const item of fileItems) {
+                operations.push({ payload, points: [item.chunkId] });
+            }
+        }
+        if (operations.length === 0)
+            return;
+        for (let i = 0; i < operations.length; i += BATCH_SIZE) {
+            const batch = operations.slice(i, i + BATCH_SIZE);
+            try {
+                await this.qdrant.batchSetPayload(collectionName, batch);
+            }
+            catch (error) {
+                if (isDebug()) {
+                    console.error("[EnrichmentApplier] batchSetPayload failed:", error);
+                }
+            }
+        }
+        pipelineLog.addStageTime("enrichApply", Date.now() - applyStart);
+    }
+    /**
+     * Apply chunk-level signal overlays.
+     * Payload written as { [providerKey]: { chunk: overlay } }.
+     */
+    async applyChunkSignals(collectionName, providerKey, chunkMetadata) {
+        let batch = [];
+        let applied = 0;
+        for (const [, overlayMap] of chunkMetadata) {
+            for (const [chunkId, overlay] of overlayMap) {
+                batch.push({
+                    payload: { [providerKey]: { chunk: overlay } },
+                    points: [chunkId],
+                });
+                if (batch.length >= BATCH_SIZE) {
+                    try {
+                        await this.qdrant.batchSetPayload(collectionName, batch);
+                        applied += batch.length;
+                    }
+                    catch (error) {
+                        if (isDebug()) {
+                            console.error("[EnrichmentApplier] chunk batch failed:", error);
+                        }
+                    }
+                    batch = [];
+                }
+            }
+        }
+        if (batch.length > 0) {
+            try {
+                await this.qdrant.batchSetPayload(collectionName, batch);
+                applied += batch.length;
+            }
+            catch (error) {
+                if (isDebug()) {
+                    console.error("[EnrichmentApplier] final chunk batch failed:", error);
+                }
+            }
+        }
+        return applied;
+    }
+}
+//# sourceMappingURL=applier.js.map

package/build/core/ingest/pipeline/enrichment/applier.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"applier.js","sourceRoot":"","sources":["../../../../../src/core/ingest/pipeline/enrichment/applier.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,QAAQ,EAAE,MAAM,WAAW,CAAC;AAIrC,OAAO,EAAE,WAAW,EAAE,MAAM,0BAA0B,CAAC;AACvD,OAAO,EAAE,OAAO,EAAE,MAAM,qBAAqB,CAAC;AAG9C,MAAM,UAAU,GAAG,GAAG,CAAC;AAEvB,MAAM,OAAO,iBAAiB;IAMC;IAL7B,YAAY,GAAG,CAAC,CAAC;IACjB,WAAW,GAAG,CAAC,CAAC;IACP,iBAAiB,GAAa,EAAE,CAAC;IACjC,gBAAgB,GAAG,IAAI,GAAG,EAAkD,CAAC;IAEtF,YAA6B,MAAqB;QAArB,WAAM,GAAN,MAAM,CAAe;IAAG,CAAC;IAEtD;;;;;;OAMG;IACH,KAAK,CAAC,gBAAgB,CACpB,cAAsB,EACtB,WAAmB,EACnB,YAA4C,EAC5C,QAAgB,EAChB,KAAkB,EAClB,SAA+B;QAE/B,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QAE9B,0BAA0B;QAC1B,MAAM,MAAM,GAAG,IAAI,GAAG,EAAuB,CAAC;QAC9C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,EAAE,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC;YACxC,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC;YACtC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACpB,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAC;QAC3B,CAAC;QAED,MAAM,UAAU,GAGV,EAAE,CAAC;QAET,KAAK,MAAM,CAAC,QAAQ,EAAE,SAAS,CAAC,IAAI,MAAM,EAAE,CAAC;YAC3C,MAAM,YAAY,GAAG,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;YAClD,MAAM,IAAI,GAAG,YAAY,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC;YAC5C,IAAI,CAAC,IAAI,EAAE,CAAC;gBACV,IAAI,CAAC,WAAW,EAAE,CAAC;gBACnB,IAAI,IAAI,CAAC,iBAAiB,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;oBACvC,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;gBAC5C,CAAC;gBACD,MAAM,QAAQ,GAAG,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC;gBAC/D,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE,CAAC;oBAC7B,QAAQ,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,IAAI,CAAC,OAAO,EAAE,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;gBACxE,CAAC;gBACD,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,YAAY,EAAE,QAAQ,CAAC,CAAC;gBAClD,SAAS;YACX,CAAC;YACD,IAAI,CAAC,YAAY,EAAE,CAAC;YAEpB,MAAM,UAAU,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC;YACzF,MAAM,SAAS,GAAG,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;YACjE,MAAM,OAAO,GAAG,EAAE,CAAC,WAAW,CAAC,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,EAAE,CAAC;YAEvD,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE,CAAC;gBAC7B,UAAU,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;YACvD,CAAC;QACH,CAAC;QAED,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO;QAEpC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,IAAI,UAAU,EAAE,CAAC;YACvD,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,CAAC;YAClD,IAAI,CAAC;gBACH,MAAM,IAAI,CAAC,MAAM,CAAC,eAAe,CAAC,cAAc,EAAE,KAAK,CAAC,CAAC;YAC3D,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,OAAO,EAAE,EAAE,CAAC;oBACd,OAAO,CAAC,KAAK,CAAC,6CAA6C,EAAE,KAAK,CAAC,CAAC;gBACtE,CAAC;YACH,CAAC;QACH,CAAC;QAED,WAAW,CAAC,YAAY,CAAC,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,UAAU,CAAC,CAAC;IACnE,CAAC;IAED;;;OAGG;IACH,KAAK,CAAC,iBAAiB,CACrB,cAAsB,EACtB,WAAmB,EACnB,aAA2D;QAE3D,IAAI,KAAK,GAGH,EAAE,CAAC;QACT,IAAI,OAAO,GAAG,CAAC,CAAC;QAEhB,KAAK,MAAM,CAAC,EAAE,UAAU,CAAC,IAAI,aAAa,EAAE,CAAC;YAC3C,KAAK,MAAM,CAAC,OAAO,EAAE,OAAO,CAAC,IAAI,UAAU,EAAE,CAAC;gBAC5C,KAAK,CAAC,IAAI,CAAC;oBACT,OAAO,EAAE,EAAE,CAAC,WAAW,CAAC,EAAE,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE;oBAC9C,MAAM,EAAE,CAAC,OAAO,CAAC;iBAClB,CAAC,CAAC;gBAEH,IAAI,KAAK,CAAC,MAAM,IAAI,UAAU,EAAE,CAAC;oBAC/B,IAAI,CAAC;wBACH,MAAM,IAAI,CAAC,MAAM,CAAC,eAAe,CAAC,cAAc,EAAE,KAAK,CAAC,CAAC;wBACzD,OAAO,IAAI,KAAK,CAAC,MAAM,CAAC;oBAC1B,CAAC;oBAAC,OAAO,KAAK,EAAE,CAAC;wBACf,IAAI,OAAO,EAAE,EAAE,CAAC;4BACd,OAAO,CAAC,KAAK,CAAC,yCAAyC,EAAE,KAAK,CAAC,CAAC;wBAClE,CAAC;oBACH,CAAC;oBACD,KAAK,GAAG,EAAE,CAAC;gBACb,CAAC;YACH,CAAC;QACH,CAAC;QAED,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrB,IAAI,CAAC;gBACH,MAAM,IAAI,CAAC,MAAM,CAAC,eAAe,CAAC,cAAc,EAAE,KAAK,CAAC,CAAC;gBACzD,OAAO,IAAI,KAAK,CAAC,MAAM,CAAC;YAC1B,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,OAAO,EAAE,EAAE,CAAC;oBACd,OAAO,CAAC,KAAK,CAAC,+CAA+C,EAAE,KAAK,CAAC,CAAC;gBACxE,CAAC;YACH,CAAC;QACH,CAAC;QAED,OAAO,OAAO,CAAC;IACjB,CAAC;CACF"}

package/build/core/ingest/pipeline/enrichment/coordinator.d.ts

@@ -0,0 +1,53 @@
+/**
+ * EnrichmentCoordinator — generic timing orchestrator for enrichment providers.
+ *
+ * Coordinates three phases per provider:
+ * 1. Prefetch: provider.buildFileSignals (fire-and-forget at T=0)
+ * 2. Per-batch: apply file signals as chunks arrive
+ * 3. Post-flush: provider.buildChunkSignals overlays
+ *
+ * Supports multiple providers in parallel — each with independent state.
+ * Provider-agnostic — works with any EnrichmentProvider implementation.
+ */
+import type { Ignore } from "ignore";
+import type { QdrantManager } from "../../../adapters/qdrant/client.js";
+import type { ChunkLookupEntry, EnrichmentInfo, EnrichmentMetrics } from "../../../types.js";
+import type { ChunkItem } from "../types.js";
+import type { EnrichmentProvider } from "./types.js";
+export declare class EnrichmentCoordinator {
+    private readonly qdrant;
+    private readonly states;
+    private startTime;
+    private readonly applier;
+    /** All provider keys managed by this coordinator. */
+    get providerKeys(): string[];
+    /** @deprecated Use providerKeys instead. Returns the first provider key for backward compat. */
+    get providerKey(): string;
+    constructor(qdrant: QdrantManager, providers: EnrichmentProvider | EnrichmentProvider[]);
+    /**
+     * Start file-level metadata prefetch at T=0. Non-blocking.
+     * Call before pipeline.start() to maximize overlap.
+     * All providers prefetch in parallel.
+     */
+    prefetch(absolutePath: string, collectionName?: string, ignoreFilter?: Ignore): void;
+    /**
+     * Called per-batch by pipeline callback after chunks are stored in Qdrant.
+     */
+    onChunksStored(collectionName: string, absolutePath: string, items: ChunkItem[]): void;
+    /**
+     * Start chunk-level enrichment (Phase 2b). Fire-and-forget, tracked internally.
+     * Each provider runs independently.
+     */
+    startChunkEnrichment(collectionName: string, absolutePath: string, chunkMap: Map<string, ChunkLookupEntry[]>): void;
+    /**
+     * Wait for all in-flight enrichment work to complete across all providers.
+     */
+    awaitCompletion(collectionName: string): Promise<EnrichmentMetrics>;
+    /**
+     * Update enrichment progress marker in Qdrant.
+     */
+    updateEnrichmentMarker(collectionName: string, info: Partial<EnrichmentInfo>): Promise<void>;
+    private flushPendingBatches;
+    private backfillMissedFiles;
+}
+//# sourceMappingURL=coordinator.d.ts.map

package/build/core/ingest/pipeline/enrichment/coordinator.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"coordinator.d.ts","sourceRoot":"","sources":["../../../../../src/core/ingest/pipeline/enrichment/coordinator.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAIH,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,QAAQ,CAAC;AAErC,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,oCAAoC,CAAC;AAExE,OAAO,KAAK,EAAE,gBAAgB,EAAE,cAAc,EAAE,iBAAiB,EAAE,MAAM,mBAAmB,CAAC;AAI7F,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,aAAa,CAAC;AAE7C,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,YAAY,CAAC;AAgDrD,qBAAa,qBAAqB;IAmB9B,OAAO,CAAC,QAAQ,CAAC,MAAM;IAlBzB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAA6B;IACpD,OAAO,CAAC,SAAS,CAAK;IAGtB,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAoB;IAE5C,qDAAqD;IACrD,IAAI,YAAY,IAAI,MAAM,EAAE,CAE3B;IAED,gGAAgG;IAChG,IAAI,WAAW,IAAI,MAAM,CAGxB;gBAGkB,MAAM,EAAE,aAAa,EACtC,SAAS,EAAE,kBAAkB,GAAG,kBAAkB,EAAE;IAOtD;;;;OAIG;IACH,QAAQ,CAAC,YAAY,EAAE,MAAM,EAAE,cAAc,CAAC,EAAE,MAAM,EAAE,YAAY,CAAC,EAAE,MAAM,GAAG,IAAI;IAoFpF;;OAEG;IACH,cAAc,CAAC,cAAc,EAAE,MAAM,EAAE,YAAY,EAAE,MAAM,EAAE,KAAK,EAAE,SAAS,EAAE,GAAG,IAAI;IA4BtF;;;OAGG;IACH,oBAAoB,CAAC,cAAc,EAAE,MAAM,EAAE,YAAY,EAAE,MAAM,EAAE,QAAQ,EAAE,GAAG,CAAC,MAAM,EAAE,gBAAgB,EAAE,CAAC,GAAG,IAAI;IA4EnH;;OAEG;IACG,eAAe,CAAC,cAAc,EAAE,MAAM,GAAG,OAAO,CAAC,iBAAiB,CAAC;IAqGzE;;OAEG;IACG,sBAAsB,CAAC,cAAc,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,CAAC,cAAc,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC;IAgBlG,OAAO,CAAC,mBAAmB;YA4Bb,mBAAmB;CAsElC"}