rag-lite-ts 1.0.1 → 2.0.0

package/dist/core/search.js

@@ -0,0 +1,296 @@
+/**
+ * CORE MODULE — Shared between text-only (rag-lite-ts) and future multimodal (rag-lite-mm)
+ * Model-agnostic. No transformer or modality-specific logic.
+ */
+import { getChunksByEmbeddingIds } from './db.js';
+import { config } from './config.js';
+import { createMissingDependencyError } from './actionable-error-messages.js';
+/**
+ * Search engine that provides semantic search capabilities
+ * Implements the core search pipeline: query embedding → vector search → metadata retrieval → optional reranking
+ * Uses explicit dependency injection for clean architecture
+ */
+export class SearchEngine {
+    embedFn;
+    indexManager;
+    db;
+    rerankFn;
+    contentResolver;
+    /**
+     * Creates a new SearchEngine with explicit dependency injection
+     *
+     * DEPENDENCY INJECTION PATTERN:
+     * This constructor requires all dependencies to be explicitly provided, enabling:
+     * - Clean separation between core logic and implementation-specific components
+     * - Support for different embedding models (text-only, multimodal, custom)
+     * - Testability through mock injection
+     * - Future extensibility without core changes
+     *
+     * @param embedFn - Function to embed queries into vectors
+     *   - Signature: (query: string, contentType?: string) => Promise<EmbeddingResult>
+     *   - Examples:
+     *     - Text: const embedFn = (query) => textEmbedder.embedSingle(query)
+     *     - Multimodal: const embedFn = (query, type) => type === 'image' ? clipEmbedder.embedImage(query) : clipEmbedder.embedText(query)
+     *     - Custom: const embedFn = (query) => customModel.embed(query)
+     *
+     * @param indexManager - Vector index manager for similarity search
+     *   - Handles vector storage and retrieval operations
+     *   - Works with any embedding dimensions (384, 512, 768, etc.)
+     *   - Example: new IndexManager('./index.bin')
+     *
+     * @param db - Database connection for metadata retrieval
+     *   - Provides access to document and chunk metadata
+     *   - Supports different content types through metadata fields
+     *   - Example: await openDatabase('./db.sqlite')
+     *
+     * @param rerankFn - Optional function to rerank search results
+     *   - Signature: (query: string, results: SearchResult[], contentType?: string) => Promise<SearchResult[]>
+     *   - Examples:
+     *     - Text: const rerankFn = (query, results) => textReranker.rerank(query, results)
+     *     - Custom: const rerankFn = (query, results) => customReranker.rerank(query, results)
+     *     - Disabled: undefined (no reranking)
+     *
+     * USAGE EXAMPLES:
+     * ```typescript
+     * // Text-only search engine
+     * const textEmbedFn = await createTextEmbedder();
+     * const textRerankFn = await createTextReranker();
+     * const indexManager = new IndexManager('./index.bin');
+     * const db = await openDatabase('./db.sqlite');
+     * const search = new SearchEngine(textEmbedFn, indexManager, db, textRerankFn);
+     *
+     * // Search engine without reranking
+     * const search = new SearchEngine(textEmbedFn, indexManager, db);
+     *
+     * // Custom embedding implementation
+     * const customEmbedFn = async (query) => ({
+     *   embedding_id: generateId(),
+     *   vector: await myCustomModel.embed(query)
+     * });
+     * const search = new SearchEngine(customEmbedFn, indexManager, db);
+     * ```
+     */
+    constructor(embedFn, indexManager, db, rerankFn, contentResolver) {
+        this.embedFn = embedFn;
+        this.indexManager = indexManager;
+        this.db = db;
+        this.rerankFn = rerankFn;
+        // Validate required dependencies
+        if (!embedFn || typeof embedFn !== 'function') {
+            throw createMissingDependencyError('embedFn', 'function', {
+                operationContext: 'SearchEngine constructor'
+            });
+        }
+        if (!indexManager) {
+            throw createMissingDependencyError('indexManager', 'object', {
+                operationContext: 'SearchEngine constructor'
+            });
+        }
+        if (!db) {
+            throw createMissingDependencyError('db', 'object', {
+                operationContext: 'SearchEngine constructor'
+            });
+        }
+        // Initialize ContentResolver if provided, or create lazily when needed
+        this.contentResolver = contentResolver;
+    }
+    /**
+     * Perform semantic search on the indexed documents
+     * Implements the core search pipeline: query embedding → vector search → metadata retrieval → optional reranking
+     * @param query - Search query string
+     * @param options - Search options including top_k and rerank settings
+     * @returns Promise resolving to array of search results
+     */
+    async search(query, options = {}) {
+        if (!query || query.trim().length === 0) {
+            return [];
+        }
+        const startTime = performance.now();
+        const topK = options.top_k || config.top_k || 10;
+        const shouldRerank = options.rerank !== undefined ? options.rerank : (this.rerankFn !== undefined);
+        try {
+            // Step 1: Build query embedding using injected embed function
+            const embeddingStartTime = performance.now();
+            const queryEmbedding = await this.embedFn(query);
+            const embeddingTime = performance.now() - embeddingStartTime;
+            // Step 2: Search using IndexManager (which handles hash mapping properly)
+            const searchStartTime = performance.now();
+            let searchResult;
+            try {
+                searchResult = this.indexManager.search(queryEmbedding.vector, topK);
+            }
+            catch (error) {
+                if (error instanceof Error && error.message.includes('No embedding ID found for hash')) {
+                    console.warn(`Hash mapping issue detected: ${error.message}`);
+                    console.warn('This may indicate index/database synchronization issues. Consider running: raglite rebuild');
+                    return [];
+                }
+                throw error;
+            }
+            const vectorSearchTime = performance.now() - searchStartTime;
+            if (searchResult.embeddingIds.length === 0) {
+                const totalTime = performance.now() - startTime;
+                console.log(`No similar documents found (${totalTime.toFixed(2)}ms total)`);
+                return [];
+            }
+            // Step 3: Retrieve chunks from database using embedding IDs
+            const retrievalStartTime = performance.now();
+            const chunks = await getChunksByEmbeddingIds(this.db, searchResult.embeddingIds);
+            const retrievalTime = performance.now() - retrievalStartTime;
+            // Step 4: Format results as JSON with text, score, and document metadata
+            let results = this.formatSearchResults(chunks, searchResult.distances, searchResult.embeddingIds);
+            // Step 5: Optional reranking with injected rerank function
+            let rerankTime = 0;
+            if (shouldRerank && this.rerankFn && results.length > 1) {
+                try {
+                    const rerankStartTime = performance.now();
+                    results = await this.rerankFn(query, results);
+                    rerankTime = performance.now() - rerankStartTime;
+                }
+                catch (error) {
+                    // Fallback to vector search results and log the error
+                    console.warn(`Reranking failed, using vector search results: ${error instanceof Error ? error.message : 'Unknown error'}`);
+                }
+            }
+            const totalTime = performance.now() - startTime;
+            // Measure latency without premature optimization - just log for monitoring
+            console.log(`Search completed: ${results.length} results in ${totalTime.toFixed(2)}ms ` +
+                `(embed: ${embeddingTime.toFixed(2)}ms, vector: ${vectorSearchTime.toFixed(2)}ms, ` +
+                `retrieval: ${retrievalTime.toFixed(2)}ms${rerankTime > 0 ? `, rerank: ${rerankTime.toFixed(2)}ms` : ''})`);
+            return results;
+        }
+        catch (error) {
+            throw new Error(`Search failed: ${error instanceof Error ? error.message : 'Unknown error'}`);
+        }
+    }
+    /**
+     * Format search results with proper structure
+     * @param chunks - Database chunks with metadata
+     * @param distances - Similarity distances from vector search
+     * @param embeddingIds - Embedding IDs in search result order
+     * @returns Formatted search results
+     */
+    formatSearchResults(chunks, distances, embeddingIds) {
+        const results = [];
+        // Create a map for quick chunk lookup by embedding_id
+        const chunkMap = new Map();
+        chunks.forEach(chunk => {
+            chunkMap.set(chunk.embedding_id, chunk);
+        });
+        // Build results in the order of search results
+        for (let i = 0; i < embeddingIds.length; i++) {
+            const embeddingId = embeddingIds[i];
+            const chunk = chunkMap.get(embeddingId);
+            if (chunk) {
+                // Convert cosine distance to similarity score (1 - distance)
+                // hnswlib-wasm returns cosine distance, we want similarity
+                const score = Math.max(0, 1 - distances[i]);
+                results.push({
+                    content: chunk.content,
+                    score: score,
+                    contentType: chunk.content_type || 'text',
+                    document: {
+                        id: chunk.document_id,
+                        source: chunk.document_source,
+                        title: chunk.document_title,
+                        contentType: chunk.document_content_type || 'text',
+                        contentId: chunk.document_content_id || undefined
+                    }
+                });
+            }
+        }
+        return results;
+    }
+    /**
+     * Get search engine statistics
+     * @returns Object with current search engine stats
+     */
+    async getStats() {
+        const indexStats = await this.indexManager.getStats();
+        return {
+            totalChunks: indexStats.totalVectors,
+            indexSize: indexStats.totalVectors,
+            rerankingEnabled: this.rerankFn !== undefined
+        };
+    }
+    /**
+     * Retrieve content by ID in the specified format
+     * @param contentId - Content ID to retrieve
+     * @param format - Format to return ('file' for CLI clients, 'base64' for MCP clients)
+     * @returns Promise that resolves to content in requested format
+     */
+    async getContent(contentId, format = 'file') {
+        // Lazy initialization of ContentResolver
+        if (!this.contentResolver) {
+            const { ContentResolver } = await import('./content-resolver.js');
+            this.contentResolver = new ContentResolver(this.db);
+        }
+        return this.contentResolver.getContent(contentId, format);
+    }
+    /**
+     * Retrieve multiple content items efficiently in batch
+     * @param contentIds - Array of content IDs to retrieve
+     * @param format - Format to return ('file' for CLI clients, 'base64' for MCP clients)
+     * @returns Promise that resolves to array of content in requested format
+     */
+    async getContentBatch(contentIds, format = 'file') {
+        // Lazy initialization of ContentResolver
+        if (!this.contentResolver) {
+            const { ContentResolver } = await import('./content-resolver.js');
+            this.contentResolver = new ContentResolver(this.db);
+        }
+        // Convert contentIds array to ContentRequest array
+        const requests = contentIds.map(contentId => ({ contentId, format }));
+        const results = await this.contentResolver.getContentBatch(requests);
+        // Extract content from results, maintaining order and handling errors
+        return results.map(result => {
+            if (!result.success) {
+                throw new Error(`Failed to retrieve content ${result.contentId}: ${result.error}`);
+            }
+            return result.content;
+        });
+    }
+    /**
+     * Retrieve content metadata for result enhancement
+     * @param contentId - Content ID to get metadata for
+     * @returns Promise that resolves to content metadata
+     */
+    async getContentMetadata(contentId) {
+        // Lazy initialization of ContentResolver
+        if (!this.contentResolver) {
+            const { ContentResolver } = await import('./content-resolver.js');
+            this.contentResolver = new ContentResolver(this.db);
+        }
+        return this.contentResolver.getContentMetadata(contentId);
+    }
+    /**
+     * Verify that content exists and is accessible
+     * @param contentId - Content ID to verify
+     * @returns Promise that resolves to true if content exists, false otherwise
+     */
+    async verifyContentExists(contentId) {
+        // Lazy initialization of ContentResolver
+        if (!this.contentResolver) {
+            const { ContentResolver } = await import('./content-resolver.js');
+            this.contentResolver = new ContentResolver(this.db);
+        }
+        return this.contentResolver.verifyContentExists(contentId);
+    }
+    /**
+     * Clean up resources - explicit cleanup method
+     */
+    async cleanup() {
+        try {
+            // Clean up ContentResolver to prevent resource leaks
+            if (this.contentResolver && typeof this.contentResolver.cleanup === 'function') {
+                this.contentResolver.cleanup();
+            }
+            await this.db.close();
+            await this.indexManager.close();
+        }
+        catch (error) {
+            console.error('Error during SearchEngine cleanup:', error instanceof Error ? error.message : String(error));
+        }
+    }
+}
+//# sourceMappingURL=search.js.map

package/dist/core/streaming-operations.d.ts

@@ -0,0 +1,145 @@
+/**
+ * Streaming Operations for Large Content - Task 9.1 Implementation
+ * Provides memory-efficient streaming operations for content ingestion and retrieval
+ * Minimizes memory usage for large files through streaming algorithms
+ */
+/**
+ * Progress callback for long-running operations
+ */
+export interface ProgressCallback {
+    (bytesProcessed: number, totalBytes?: number): void;
+}
+/**
+ * Streaming hash calculation result
+ */
+export interface StreamingHashResult {
+    hash: string;
+    bytesProcessed: number;
+    processingTimeMs: number;
+}
+/**
+ * Streaming file copy result
+ */
+export interface StreamingCopyResult {
+    bytesWritten: number;
+    processingTimeMs: number;
+    hash?: string;
+}
+/**
+ * Configuration for streaming operations
+ */
+export interface StreamingConfig {
+    chunkSize: number;
+    enableProgress: boolean;
+    enableHashing: boolean;
+    timeout: number;
+}
+/**
+ * StreamingOperations class provides memory-efficient operations for large content
+ */
+export declare class StreamingOperations {
+    private config;
+    constructor(config?: Partial<StreamingConfig>);
+    /**
+     * Calculates SHA-256 hash of a file using streaming to minimize memory usage
+     * @param filePath - Path to the file to hash
+     * @param progressCallback - Optional callback for progress reporting
+     * @returns Promise that resolves to hash result
+     */
+    calculateFileHashStreaming(filePath: string, progressCallback?: ProgressCallback): Promise<StreamingHashResult>;
+    /**
+     * Calculates SHA-256 hash of a buffer using streaming to minimize memory usage
+     * @param content - Buffer to hash
+     * @param progressCallback - Optional callback for progress reporting
+     * @returns Promise that resolves to hash result
+     */
+    calculateBufferHashStreaming(content: Buffer, progressCallback?: ProgressCallback): Promise<StreamingHashResult>;
+    /**
+     * Copies a file using streaming operations with optional hashing
+     * @param sourcePath - Source file path
+     * @param destinationPath - Destination file path
+     * @param progressCallback - Optional callback for progress reporting
+     * @returns Promise that resolves to copy result
+     */
+    copyFileStreaming(sourcePath: string, destinationPath: string, progressCallback?: ProgressCallback): Promise<StreamingCopyResult>;
+    /**
+     * Writes buffer content to file using streaming operations
+     * @param content - Buffer to write
+     * @param destinationPath - Destination file path
+     * @param progressCallback - Optional callback for progress reporting
+     * @returns Promise that resolves to write result
+     */
+    writeBufferStreaming(content: Buffer, destinationPath: string, progressCallback?: ProgressCallback): Promise<StreamingCopyResult>;
+    /**
+     * Reads file content and converts to base64 using streaming to minimize memory usage
+     * @param filePath - Path to the file to read
+     * @param progressCallback - Optional callback for progress reporting
+     * @returns Promise that resolves to base64 string
+     */
+    readFileAsBase64Streaming(filePath: string, progressCallback?: ProgressCallback): Promise<string>;
+    /**
+     * Validates file integrity by comparing streaming hash with expected hash
+     * @param filePath - Path to the file to validate
+     * @param expectedHash - Expected SHA-256 hash
+     * @param progressCallback - Optional callback for progress reporting
+     * @returns Promise that resolves to validation result
+     */
+    validateFileIntegrityStreaming(filePath: string, expectedHash: string, progressCallback?: ProgressCallback): Promise<{
+        isValid: boolean;
+        actualHash: string;
+        bytesProcessed: number;
+    }>;
+    /**
+     * Gets file information without loading content into memory
+     * @param filePath - Path to the file
+     * @returns Promise that resolves to file information
+     */
+    getFileInfo(filePath: string): Promise<{
+        size: number;
+        isFile: boolean;
+        isDirectory: boolean;
+        lastModified: Date;
+        canRead: boolean;
+        canWrite: boolean;
+    }>;
+    /**
+     * Converts buffer to chunks for streaming
+     * @param buffer - Buffer to chunk
+     * @returns Generator that yields buffer chunks
+     */
+    private bufferToChunks;
+    /**
+     * Wraps a promise with timeout functionality
+     * @param promise - Promise to wrap
+     * @param timeoutMs - Timeout in milliseconds
+     * @param errorMessage - Error message for timeout
+     * @returns Promise that rejects if timeout is reached
+     */
+    private withTimeout;
+}
+/**
+ * Creates a StreamingOperations instance with default configuration
+ * @param config - Optional configuration overrides
+ * @returns StreamingOperations instance
+ */
+export declare function createStreamingOperations(config?: Partial<StreamingConfig>): StreamingOperations;
+/**
+ * Utility function to format bytes for progress reporting
+ * @param bytes - Number of bytes
+ * @returns Formatted string (e.g., "1.5 MB")
+ */
+export declare function formatBytes(bytes: number): string;
+/**
+ * Utility function to format processing time
+ * @param milliseconds - Processing time in milliseconds
+ * @returns Formatted string (e.g., "1.5s" or "150ms")
+ */
+export declare function formatProcessingTime(milliseconds: number): string;
+/**
+ * Utility function to calculate processing speed
+ * @param bytes - Number of bytes processed
+ * @param milliseconds - Processing time in milliseconds
+ * @returns Speed in MB/s
+ */
+export declare function calculateProcessingSpeed(bytes: number, milliseconds: number): number;
+//# sourceMappingURL=streaming-operations.d.ts.map