@matperez/coderag 0.1.24

package/dist/db/schema.js

@@ -0,0 +1,85 @@
+/**
+ * Database schema for persistent codebase index
+ */
+import { index, integer, real, sqliteTable, text } from 'drizzle-orm/sqlite-core';
+/**
+ * Indexed files table
+ * Stores file metadata and content
+ */
+export const files = sqliteTable('files', {
+    id: integer('id').primaryKey({ autoIncrement: true }),
+    path: text('path').notNull().unique(),
+    content: text('content').notNull(),
+    hash: text('hash').notNull(),
+    size: integer('size').notNull(),
+    mtime: integer('mtime').notNull(), // Unix timestamp in milliseconds
+    language: text('language'),
+    indexedAt: integer('indexed_at').notNull(), // Unix timestamp
+    magnitude: real('magnitude').default(0), // Pre-computed TF-IDF vector magnitude for cosine similarity
+    tokenCount: integer('token_count').default(0), // Total tokens in document (for BM25 length normalization)
+}, (table) => ({
+    pathIdx: index('files_path_idx').on(table.path),
+    hashIdx: index('files_hash_idx').on(table.hash),
+}));
+/**
+ * Chunks table
+ * Stores semantic chunks extracted from files (functions, classes, sections, etc.)
+ */
+export const chunks = sqliteTable('chunks', {
+    id: integer('id').primaryKey({ autoIncrement: true }),
+    fileId: integer('file_id')
+        .notNull()
+        .references(() => files.id, { onDelete: 'cascade' }),
+    content: text('content').notNull(), // Chunk content
+    type: text('type').notNull(), // Chunk type (FunctionDeclaration, heading, etc.)
+    startLine: integer('start_line').notNull(),
+    endLine: integer('end_line').notNull(),
+    metadata: text('metadata'), // JSON string for additional metadata
+    tokenCount: integer('token_count').default(0), // Tokens in chunk (for BM25)
+    magnitude: real('magnitude').default(0), // TF-IDF magnitude for chunk
+}, (table) => ({
+    fileIdIdx: index('chunks_file_id_idx').on(table.fileId),
+    typeIdx: index('chunks_type_idx').on(table.type),
+}));
+/**
+ * TF-IDF vectors table
+ * Stores term frequencies and TF-IDF scores per CHUNK (not file)
+ */
+export const documentVectors = sqliteTable('document_vectors', {
+    id: integer('id').primaryKey({ autoIncrement: true }),
+    chunkId: integer('chunk_id')
+        .notNull()
+        .references(() => chunks.id, { onDelete: 'cascade' }),
+    term: text('term').notNull(),
+    tf: real('tf').notNull(), // Term frequency
+    tfidf: real('tfidf').notNull(), // TF-IDF score
+    rawFreq: integer('raw_freq').notNull(), // Raw term count
+}, (table) => ({
+    chunkIdIdx: index('vectors_chunk_id_idx').on(table.chunkId),
+    termIdx: index('vectors_term_idx').on(table.term),
+    tfidfIdx: index('vectors_tfidf_idx').on(table.tfidf),
+    termChunkIdx: index('vectors_term_chunk_idx').on(table.term, table.chunkId), // Composite index for search
+}));
+/**
+ * IDF (Inverse Document Frequency) table
+ * Stores global IDF scores for terms
+ */
+export const idfScores = sqliteTable('idf_scores', {
+    id: integer('id').primaryKey({ autoIncrement: true }),
+    term: text('term').notNull().unique(),
+    idf: real('idf').notNull(),
+    documentFrequency: integer('document_frequency').notNull(), // How many docs contain this term
+}, (table) => ({
+    termIdx: index('idf_term_idx').on(table.term),
+}));
+/**
+ * Index metadata table
+ * Stores global index information
+ */
+export const indexMetadata = sqliteTable('index_metadata', {
+    id: integer('id').primaryKey({ autoIncrement: true }),
+    key: text('key').notNull().unique(),
+    value: text('value').notNull(),
+    updatedAt: integer('updated_at').notNull(),
+});
+//# sourceMappingURL=schema.js.map

package/dist/db/schema.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"schema.js","sourceRoot":"","sources":["../../src/db/schema.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,MAAM,yBAAyB,CAAA;AAEjF;;;GAGG;AACH,MAAM,CAAC,MAAM,KAAK,GAAG,WAAW,CAC/B,OAAO,EACP;IACC,EAAE,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;IACrD,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,CAAC,MAAM,EAAE;IACrC,OAAO,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC,OAAO,EAAE;IAClC,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE;IAC5B,IAAI,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE;IAC/B,KAAK,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,iCAAiC;IACpE,QAAQ,EAAE,IAAI,CAAC,UAAU,CAAC;IAC1B,SAAS,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC,OAAO,EAAE,EAAE,iBAAiB;IAC7D,SAAS,EAAE,IAAI,CAAC,WAAW,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,6DAA6D;IACtG,UAAU,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,2DAA2D;CAC1G,EACD,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACX,OAAO,EAAE,KAAK,CAAC,gBAAgB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC;IAC/C,OAAO,EAAE,KAAK,CAAC,gBAAgB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC;CAC/C,CAAC,CACF,CAAA;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,MAAM,GAAG,WAAW,CAChC,QAAQ,EACR;IACC,EAAE,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;IACrD,MAAM,EAAE,OAAO,CAAC,SAAS,CAAC;SACxB,OAAO,EAAE;SACT,UAAU,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC;IACrD,OAAO,EAAE,IAAI,CAAC,SAAS,CAAC,CAAC,OAAO,EAAE,EAAE,gBAAgB;IACpD,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,kDAAkD;IAChF,SAAS,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC,OAAO,EAAE;IAC1C,OAAO,EAAE,OAAO,CAAC,UAAU,CAAC,CAAC,OAAO,EAAE;IACtC,QAAQ,EAAE,IAAI,CAAC,UAAU,CAAC,EAAE,sCAAsC;IAClE,UAAU,EAAE,OAAO,CAAC,aAAa,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,6BAA6B;IAC5E,SAAS,EAAE,IAAI,CAAC,WAAW,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,6BAA6B;CACtE,EACD,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACX,SAAS,EAAE,KAAK,CAAC,oBAAoB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC;IACvD,OAAO,EAAE,KAAK,CAAC,iBAAiB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC;CAChD,CAAC,CACF,CAAA;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,WAAW,CACzC,kBAAkB,EAClB;IACC,EAAE,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;IACrD,OAAO,EAAE,OAAO,CAAC,UAAU,CAAC;SAC1B,OAAO,EAAE;SACT,UAAU,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC;IACtD,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE;IAC5B,EAAE,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,iBAAiB;IAC3C,KAAK,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,eAAe;IAC/C,OAAO,EAAE,OAAO,CAAC,UAAU,CAAC,CAAC,OAAO,EAAE,EAAE,iBAAiB;CACzD,EACD,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACX,UAAU,EAAE,KAAK,CAAC,sBAAsB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC;IAC3D,OAAO,EAAE,KAAK,CAAC,kBAAkB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC;IACjD,QAAQ,EAAE,KAAK,CAAC,mBAAmB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC;IACpD,YAAY,EAAE,KAAK,CAAC,wBAAwB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,OAAO,CAAC,EAAE,6BAA6B;CAC1G,CAAC,CACF,CAAA;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,SAAS,GAAG,WAAW,CACnC,YAAY,EACZ;IACC,EAAE,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;IACrD,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,CAAC,MAAM,EAAE;IACrC,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,OAAO,EAAE;IAC1B,iBAAiB,EAAE,OAAO,CAAC,oBAAoB,CAAC,CAAC,OAAO,EAAE,EAAE,kCAAkC;CAC9F,EACD,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;IACX,OAAO,EAAE,KAAK,CAAC,cAAc,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC;CAC7C,CAAC,CACF,CAAA;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,WAAW,CAAC,gBAAgB,EAAE;IAC1D,EAAE,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,UAAU,CAAC,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;IACrD,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,OAAO,EAAE,CAAC,MAAM,EAAE;IACnC,KAAK,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE;IAC9B,SAAS,EAAE,OAAO,CAAC,YAAY,CAAC,CAAC,OAAO,EAAE;CAC1C,CAAC,CAAA"}

package/dist/embeddings.d.ts

@@ -0,0 +1,92 @@
+/**
+ * Embedding generation utilities - Pure Functional
+ * Uses Vercel AI SDK with OpenAI Provider
+ *
+ * Design:
+ * - Pure functions (no classes)
+ * - Immutable data structures
+ * - Function composition
+ * - Config from environment variables
+ */
+/**
+ * Embedding Provider Config
+ */
+export interface EmbeddingConfig {
+    readonly provider: 'openai' | 'openai-compatible' | 'mock';
+    readonly model: string;
+    readonly dimensions: number;
+    readonly apiKey?: string;
+    readonly baseURL?: string;
+    readonly batchSize?: number;
+}
+/**
+ * Embedding Provider Interface
+ */
+export interface EmbeddingProvider {
+    readonly name: string;
+    readonly model: string;
+    readonly dimensions: number;
+    readonly generateEmbedding: (text: string) => Promise<number[]>;
+    readonly generateEmbeddings: (texts: string[]) => Promise<number[][]>;
+}
+/**
+ * Generate mock embedding using deterministic hash
+ * Used as fallback when OpenAI is not available
+ */
+export declare const generateMockEmbedding: (text: string, dimensions: number) => number[];
+/**
+ * Create default config from environment
+ */
+export declare const createDefaultConfig: () => EmbeddingConfig;
+/**
+ * Create OpenAI Embedding Provider (pure function)
+ */
+export declare const createOpenAIProvider: (config: EmbeddingConfig) => EmbeddingProvider;
+/**
+ * Create Mock Embedding Provider (pure function)
+ */
+export declare const createMockProvider: (dimensions?: number) => EmbeddingProvider;
+/**
+ * Provider Factory Type
+ */
+type ProviderFactory = (config: EmbeddingConfig) => EmbeddingProvider;
+/**
+ * Register a provider factory
+ */
+export declare const registerProvider: (name: string, factory: ProviderFactory) => void;
+/**
+ * Get registered provider names
+ */
+export declare const getRegisteredProviders: () => string[];
+/**
+ * Create Embedding Provider from config (pure function)
+ * Uses provider registry for modularity
+ */
+export declare const createEmbeddingProvider: (config: EmbeddingConfig) => EmbeddingProvider;
+/**
+ * Get default embedding provider (pure function with I/O)
+ */
+export declare const getDefaultEmbeddingProvider: () => Promise<EmbeddingProvider>;
+/**
+ * Chunk text into smaller pieces
+ * Pure function - no side effects
+ */
+export declare const chunkText: (text: string, options?: {
+    readonly maxChunkSize?: number;
+    readonly overlap?: number;
+}) => readonly string[];
+/**
+ * Calculate cosine similarity between two vectors
+ * Returns value between -1 and 1 (1 = identical, 0 = orthogonal, -1 = opposite)
+ */
+export declare const cosineSimilarity: (vecA: number[], vecB: number[]) => number;
+/**
+ * Normalize vector to unit length
+ */
+export declare const normalizeVector: (vector: number[]) => number[];
+/**
+ * Compose embedding providers (primary + fallback)
+ */
+export declare const composeProviders: (primary: EmbeddingProvider, fallback: EmbeddingProvider) => EmbeddingProvider;
+export {};
+//# sourceMappingURL=embeddings.d.ts.map

package/dist/embeddings.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"embeddings.d.ts","sourceRoot":"","sources":["../src/embeddings.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAKH;;GAEG;AACH,MAAM,WAAW,eAAe;IAC/B,QAAQ,CAAC,QAAQ,EAAE,QAAQ,GAAG,mBAAmB,GAAG,MAAM,CAAA;IAC1D,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAA;IACtB,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAA;IAC3B,QAAQ,CAAC,MAAM,CAAC,EAAE,MAAM,CAAA;IACxB,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAA;IACzB,QAAQ,CAAC,SAAS,CAAC,EAAE,MAAM,CAAA;CAC3B;AAED;;GAEG;AACH,MAAM,WAAW,iBAAiB;IACjC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;IACrB,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAA;IACtB,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAA;IAC3B,QAAQ,CAAC,iBAAiB,EAAE,CAAC,IAAI,EAAE,MAAM,KAAK,OAAO,CAAC,MAAM,EAAE,CAAC,CAAA;IAC/D,QAAQ,CAAC,kBAAkB,EAAE,CAAC,KAAK,EAAE,MAAM,EAAE,KAAK,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAA;CACrE;AAED;;;GAGG;AACH,eAAO,MAAM,qBAAqB,GAAI,MAAM,MAAM,EAAE,YAAY,MAAM,KAAG,MAAM,EAuB9E,CAAA;AAkBD;;GAEG;AACH,eAAO,MAAM,mBAAmB,QAAO,eAyBtC,CAAA;AAiDD;;GAEG;AACH,eAAO,MAAM,oBAAoB,GAAI,QAAQ,eAAe,KAAG,iBAW9D,CAAA;AAED;;GAEG;AACH,eAAO,MAAM,kBAAkB,GAAI,mBAAiB,KAAG,iBAOrD,CAAA;AAEF;;GAEG;AACH,KAAK,eAAe,GAAG,CAAC,MAAM,EAAE,eAAe,KAAK,iBAAiB,CAAA;AAOrE;;GAEG;AACH,eAAO,MAAM,gBAAgB,GAAI,MAAM,MAAM,EAAE,SAAS,eAAe,KAAG,IAEzE,CAAA;AAED;;GAEG;AACH,eAAO,MAAM,sBAAsB,QAAO,MAAM,EAE/C,CAAA;AAqBD;;;GAGG;AACH,eAAO,MAAM,uBAAuB,GAAI,QAAQ,eAAe,KAAG,iBAUjE,CAAA;AAED;;GAEG;AACH,eAAO,MAAM,2BAA2B,QAAa,OAAO,CAAC,iBAAiB,CAG7E,CAAA;AAED;;;GAGG;AACH,eAAO,MAAM,SAAS,GACrB,MAAM,MAAM,EACZ,UAAS;IACR,QAAQ,CAAC,YAAY,CAAC,EAAE,MAAM,CAAA;IAC9B,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAA;CACpB,KACJ,SAAS,MAAM,EAgCjB,CAAA;AAED;;;GAGG;AACH,eAAO,MAAM,gBAAgB,GAAI,MAAM,MAAM,EAAE,EAAE,MAAM,MAAM,EAAE,KAAG,MAsBjE,CAAA;AAED;;GAEG;AACH,eAAO,MAAM,eAAe,GAAI,QAAQ,MAAM,EAAE,KAAG,MAAM,EAMxD,CAAA;AAED;;GAEG;AACH,eAAO,MAAM,gBAAgB,GAC5B,SAAS,iBAAiB,EAC1B,UAAU,iBAAiB,KACzB,iBAkBD,CAAA"}

package/dist/embeddings.js

@@ -0,0 +1,275 @@
+/**
+ * Embedding generation utilities - Pure Functional
+ * Uses Vercel AI SDK with OpenAI Provider
+ *
+ * Design:
+ * - Pure functions (no classes)
+ * - Immutable data structures
+ * - Function composition
+ * - Config from environment variables
+ */
+import { createOpenAI } from '@ai-sdk/openai';
+import { embed, embedMany } from 'ai';
+/**
+ * Generate mock embedding using deterministic hash
+ * Used as fallback when OpenAI is not available
+ */
+export const generateMockEmbedding = (text, dimensions) => {
+    // Simple hash function for consistency
+    const hash = (str) => {
+        let h = 0;
+        for (let i = 0; i < str.length; i++) {
+            h = (Math.imul(31, h) + str.charCodeAt(i)) | 0;
+        }
+        return h;
+    };
+    const baseHash = hash(text);
+    const vector = [];
+    for (let i = 0; i < dimensions; i++) {
+        // Generate deterministic pseudo-random values
+        const seed = baseHash + i;
+        const value = (Math.sin(seed) + Math.cos(seed * 0.5)) / 2;
+        vector.push(value);
+    }
+    // Normalize to unit vector
+    const magnitude = Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
+    return vector.map((val) => val / magnitude);
+};
+/**
+ * Get dimensions for OpenAI model
+ * Returns undefined for unknown models (must be specified in config)
+ */
+const getModelDimensions = (model) => {
+    switch (model) {
+        case 'text-embedding-3-large':
+            return 3072;
+        case 'text-embedding-3-small':
+        case 'text-embedding-ada-002':
+            return 1536;
+        default:
+            return undefined; // For custom models, dimensions must be specified
+    }
+};
+/**
+ * Create default config from environment
+ */
+export const createDefaultConfig = () => {
+    const apiKey = process.env.OPENAI_API_KEY;
+    const baseURL = process.env.OPENAI_BASE_URL; // Support custom base URL
+    const model = process.env.EMBEDDING_MODEL || 'text-embedding-3-small';
+    const customDimensions = process.env.EMBEDDING_DIMENSIONS
+        ? parseInt(process.env.EMBEDDING_DIMENSIONS, 10)
+        : undefined;
+    // Determine provider type
+    let provider = 'mock';
+    if (apiKey) {
+        provider = baseURL ? 'openai-compatible' : 'openai';
+    }
+    // Get dimensions (from env, model default, or fallback)
+    const dimensions = customDimensions || getModelDimensions(model) || 1536;
+    return {
+        provider,
+        model,
+        dimensions,
+        apiKey,
+        baseURL,
+        batchSize: 10,
+    };
+};
+/**
+ * Create OpenAI embedding model instance
+ * Supports both official OpenAI and OpenAI-compatible endpoints
+ */
+const createEmbeddingModel = (config) => {
+    const provider = createOpenAI({
+        apiKey: config.apiKey || process.env.OPENAI_API_KEY,
+        baseURL: config.baseURL, // Support custom endpoints (OpenRouter, Together AI, etc.)
+    });
+    return provider.embedding(config.model);
+};
+/**
+ * Generate single embedding (OpenAI)
+ */
+const generateOpenAIEmbedding = async (model, text, dimensions) => {
+    try {
+        const { embedding } = await embed({ model, value: text });
+        return embedding;
+    }
+    catch (error) {
+        console.error('[WARN] OpenAI embedding failed, falling back to mock:', error);
+        return generateMockEmbedding(text, dimensions);
+    }
+};
+/**
+ * Generate multiple embeddings (OpenAI)
+ */
+const generateOpenAIEmbeddings = async (model, texts, dimensions) => {
+    try {
+        const { embeddings } = await embedMany({ model, values: texts });
+        return embeddings;
+    }
+    catch (error) {
+        console.error('[WARN] OpenAI embeddings failed, falling back to mock:', error);
+        return texts.map((text) => generateMockEmbedding(text, dimensions));
+    }
+};
+/**
+ * Create OpenAI Embedding Provider (pure function)
+ */
+export const createOpenAIProvider = (config) => {
+    const model = createEmbeddingModel(config);
+    return {
+        name: 'openai',
+        model: config.model,
+        dimensions: config.dimensions,
+        generateEmbedding: (text) => generateOpenAIEmbedding(model, text, config.dimensions),
+        generateEmbeddings: (texts) => generateOpenAIEmbeddings(model, texts, config.dimensions),
+    };
+};
+/**
+ * Create Mock Embedding Provider (pure function)
+ */
+export const createMockProvider = (dimensions = 1536) => ({
+    name: 'mock',
+    model: 'mock',
+    dimensions,
+    generateEmbedding: (text) => Promise.resolve(generateMockEmbedding(text, dimensions)),
+    generateEmbeddings: (texts) => Promise.resolve(texts.map((text) => generateMockEmbedding(text, dimensions))),
+});
+/**
+ * Provider Registry (modular design)
+ */
+const providerRegistry = new Map();
+/**
+ * Register a provider factory
+ */
+export const registerProvider = (name, factory) => {
+    providerRegistry.set(name, factory);
+};
+/**
+ * Get registered provider names
+ */
+export const getRegisteredProviders = () => {
+    return Array.from(providerRegistry.keys());
+};
+// Register built-in providers
+registerProvider('openai', (config) => {
+    console.error(`[INFO] Creating OpenAI provider: ${config.model} (${config.dimensions} dims)`);
+    return createOpenAIProvider(config);
+});
+registerProvider('openai-compatible', (config) => {
+    console.error(`[INFO] Creating OpenAI-compatible provider: ${config.model} (${config.dimensions} dims)`, config.baseURL ? `at ${config.baseURL}` : '');
+    return createOpenAIProvider(config);
+});
+registerProvider('mock', (config) => {
+    console.error('[INFO] Creating Mock embedding provider');
+    return createMockProvider(config.dimensions);
+});
+/**
+ * Create Embedding Provider from config (pure function)
+ * Uses provider registry for modularity
+ */
+export const createEmbeddingProvider = (config) => {
+    const factory = providerRegistry.get(config.provider);
+    if (!factory) {
+        console.error(`[WARN] Unknown provider '${config.provider}', using mock`);
+        const mockFactory = providerRegistry.get('mock');
+        return mockFactory?.(config);
+    }
+    return factory(config);
+};
+/**
+ * Get default embedding provider (pure function with I/O)
+ */
+export const getDefaultEmbeddingProvider = async () => {
+    const config = createDefaultConfig();
+    return createEmbeddingProvider(config);
+};
+/**
+ * Chunk text into smaller pieces
+ * Pure function - no side effects
+ */
+export const chunkText = (text, options = {}) => {
+    const maxChunkSize = options.maxChunkSize ?? 1000;
+    const overlap = options.overlap ?? 100;
+    if (text.length === 0) {
+        return [];
+    }
+    const chunks = [];
+    let start = 0;
+    while (start < text.length) {
+        const end = Math.min(start + maxChunkSize, text.length);
+        const chunk = text.slice(start, end);
+        chunks.push(chunk);
+        // If we've reached the end, break
+        if (end >= text.length) {
+            break;
+        }
+        // Move start forward, ensuring it advances
+        const nextStart = end - overlap;
+        if (nextStart <= start) {
+            // Avoid infinite loop: ensure we always move forward
+            start = end;
+        }
+        else {
+            start = nextStart;
+        }
+    }
+    return chunks;
+};
+/**
+ * Calculate cosine similarity between two vectors
+ * Returns value between -1 and 1 (1 = identical, 0 = orthogonal, -1 = opposite)
+ */
+export const cosineSimilarity = (vecA, vecB) => {
+    if (vecA.length !== vecB.length) {
+        throw new Error(`Vector dimension mismatch: ${vecA.length} vs ${vecB.length}`);
+    }
+    const { dotProduct, normA, normB } = vecA.reduce((acc, aVal, i) => {
+        const bVal = vecB[i];
+        return {
+            dotProduct: acc.dotProduct + aVal * bVal,
+            normA: acc.normA + aVal * aVal,
+            normB: acc.normB + bVal * bVal,
+        };
+    }, { dotProduct: 0, normA: 0, normB: 0 });
+    if (normA === 0 || normB === 0) {
+        return 0;
+    }
+    return dotProduct / (Math.sqrt(normA) * Math.sqrt(normB));
+};
+/**
+ * Normalize vector to unit length
+ */
+export const normalizeVector = (vector) => {
+    const magnitude = Math.sqrt(vector.reduce((sum, val) => sum + val * val, 0));
+    if (magnitude === 0) {
+        return vector;
+    }
+    return vector.map((val) => val / magnitude);
+};
+/**
+ * Compose embedding providers (primary + fallback)
+ */
+export const composeProviders = (primary, fallback) => ({
+    name: `${primary.name}+${fallback.name}`,
+    model: primary.model,
+    dimensions: primary.dimensions,
+    generateEmbedding: async (text) => {
+        try {
+            return await primary.generateEmbedding(text);
+        }
+        catch {
+            return await fallback.generateEmbedding(text);
+        }
+    },
+    generateEmbeddings: async (texts) => {
+        try {
+            return await primary.generateEmbeddings(texts);
+        }
+        catch {
+            return await fallback.generateEmbeddings(texts);
+        }
+    },
+});
+//# sourceMappingURL=embeddings.js.map

package/dist/embeddings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"embeddings.js","sourceRoot":"","sources":["../src/embeddings.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,YAAY,EAAE,MAAM,gBAAgB,CAAA;AAC7C,OAAO,EAAE,KAAK,EAAE,SAAS,EAAE,MAAM,IAAI,CAAA;AAyBrC;;;GAGG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAE,UAAkB,EAAY,EAAE;IACnF,uCAAuC;IACvC,MAAM,IAAI,GAAG,CAAC,GAAW,EAAU,EAAE;QACpC,IAAI,CAAC,GAAG,CAAC,CAAA;QACT,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACrC,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;QAC/C,CAAC;QACD,OAAO,CAAC,CAAA;IACT,CAAC,CAAA;IAED,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,CAAA;IAC3B,MAAM,MAAM,GAAa,EAAE,CAAA;IAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,8CAA8C;QAC9C,MAAM,IAAI,GAAG,QAAQ,GAAG,CAAC,CAAA;QACzB,MAAM,KAAK,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;QACzD,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;IACnB,CAAC;IAED,2BAA2B;IAC3B,MAAM,SAAS,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,GAAG,GAAG,GAAG,EAAE,CAAC,CAAC,CAAC,CAAA;IAC5E,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,SAAS,CAAC,CAAA;AAC5C,CAAC,CAAA;AAED;;;GAGG;AACH,MAAM,kBAAkB,GAAG,CAAC,KAAa,EAAsB,EAAE;IAChE,QAAQ,KAAK,EAAE,CAAC;QACf,KAAK,wBAAwB;YAC5B,OAAO,IAAI,CAAA;QACZ,KAAK,wBAAwB,CAAC;QAC9B,KAAK,wBAAwB;YAC5B,OAAO,IAAI,CAAA;QACZ;YACC,OAAO,SAAS,CAAA,CAAC,kDAAkD;IACrE,CAAC;AACF,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,mBAAmB,GAAG,GAAoB,EAAE;IACxD,MAAM,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,cAAc,CAAA;IACzC,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,eAAe,CAAA,CAAC,0BAA0B;IACtE,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,CAAC,eAAe,IAAI,wBAAwB,CAAA;IACrE,MAAM,gBAAgB,GAAG,OAAO,CAAC,GAAG,CAAC,oBAAoB;QACxD,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,oBAAoB,EAAE,EAAE,CAAC;QAChD,CAAC,CAAC,SAAS,CAAA;IAEZ,0BAA0B;IAC1B,IAAI,QAAQ,GAAgC,MAAM,CAAA;IAClD,IAAI,MAAM,EAAE,CAAC;QACZ,QAAQ,GAAG,OAAO,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,QAAQ,CAAA;IACpD,CAAC;IAED,wDAAwD;IACxD,MAAM,UAAU,GAAG,gBAAgB,IAAI,kBAAkB,CAAC,KAAK,CAAC,IAAI,IAAI,CAAA;IAExE,OAAO;QACN,QAAQ;QACR,KAAK;QACL,UAAU;QACV,MAAM;QACN,OAAO;QACP,SAAS,EAAE,EAAE;KACb,CAAA;AACF,CAAC,CAAA;AAED;;;GAGG;AACH,MAAM,oBAAoB,GAAG,CAAC,MAAuB,EAAE,EAAE;IACxD,MAAM,QAAQ,GAAG,YAAY,CAAC;QAC7B,MAAM,EAAE,MAAM,CAAC,MAAM,IAAI,OAAO,CAAC,GAAG,CAAC,cAAc;QACnD,OAAO,EAAE,MAAM,CAAC,OAAO,EAAE,2DAA2D;KACpF,CAAC,CAAA;IAEF,OAAO,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,CAAC,CAAA;AACxC,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,uBAAuB,GAAG,KAAK,EACpC,KAA8C,EAC9C,IAAY,EACZ,UAAkB,EACE,EAAE;IACtB,IAAI,CAAC;QACJ,MAAM,EAAE,SAAS,EAAE,GAAG,MAAM,KAAK,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAA;QACzD,OAAO,SAAS,CAAA;IACjB,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QAChB,OAAO,CAAC,KAAK,CAAC,uDAAuD,EAAE,KAAK,CAAC,CAAA;QAC7E,OAAO,qBAAqB,CAAC,IAAI,EAAE,UAAU,CAAC,CAAA;IAC/C,CAAC;AACF,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,wBAAwB,GAAG,KAAK,EACrC,KAA8C,EAC9C,KAAe,EACf,UAAkB,EACI,EAAE;IACxB,IAAI,CAAC;QACJ,MAAM,EAAE,UAAU,EAAE,GAAG,MAAM,SAAS,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAA;QAChE,OAAO,UAAU,CAAA;IAClB,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QAChB,OAAO,CAAC,KAAK,CAAC,wDAAwD,EAAE,KAAK,CAAC,CAAA;QAC9E,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,qBAAqB,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC,CAAA;IACpE,CAAC;AACF,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,MAAuB,EAAqB,EAAE;IAClF,MAAM,KAAK,GAAG,oBAAoB,CAAC,MAAM,CAAC,CAAA;IAE1C,OAAO;QACN,IAAI,EAAE,QAAQ;QACd,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,UAAU,EAAE,MAAM,CAAC,UAAU;QAC7B,iBAAiB,EAAE,CAAC,IAAY,EAAE,EAAE,CAAC,uBAAuB,CAAC,KAAK,EAAE,IAAI,EAAE,MAAM,CAAC,UAAU,CAAC;QAC5F,kBAAkB,EAAE,CAAC,KAAe,EAAE,EAAE,CACvC,wBAAwB,CAAC,KAAK,EAAE,KAAK,EAAE,MAAM,CAAC,UAAU,CAAC;KAC1D,CAAA;AACF,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,UAAU,GAAG,IAAI,EAAqB,EAAE,CAAC,CAAC;IAC5E,IAAI,EAAE,MAAM;IACZ,KAAK,EAAE,MAAM;IACb,UAAU;IACV,iBAAiB,EAAE,CAAC,IAAY,EAAE,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC,qBAAqB,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;IAC7F,kBAAkB,EAAE,CAAC,KAAe,EAAE,EAAE,CACvC,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,qBAAqB,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC,CAAC;CAC9E,CAAC,CAAA;AAOF;;GAEG;AACH,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAA2B,CAAA;AAE3D;;GAEG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,IAAY,EAAE,OAAwB,EAAQ,EAAE;IAChF,gBAAgB,CAAC,GAAG,CAAC,IAAI,EAAE,OAAO,CAAC,CAAA;AACpC,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,sBAAsB,GAAG,GAAa,EAAE;IACpD,OAAO,KAAK,CAAC,IAAI,CAAC,gBAAgB,CAAC,IAAI,EAAE,CAAC,CAAA;AAC3C,CAAC,CAAA;AAED,8BAA8B;AAC9B,gBAAgB,CAAC,QAAQ,EAAE,CAAC,MAAM,EAAE,EAAE;IACrC,OAAO,CAAC,KAAK,CAAC,oCAAoC,MAAM,CAAC,KAAK,KAAK,MAAM,CAAC,UAAU,QAAQ,CAAC,CAAA;IAC7F,OAAO,oBAAoB,CAAC,MAAM,CAAC,CAAA;AACpC,CAAC,CAAC,CAAA;AAEF,gBAAgB,CAAC,mBAAmB,EAAE,CAAC,MAAM,EAAE,EAAE;IAChD,OAAO,CAAC,KAAK,CACZ,+CAA+C,MAAM,CAAC,KAAK,KAAK,MAAM,CAAC,UAAU,QAAQ,EACzF,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,MAAM,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAC5C,CAAA;IACD,OAAO,oBAAoB,CAAC,MAAM,CAAC,CAAA;AACpC,CAAC,CAAC,CAAA;AAEF,gBAAgB,CAAC,MAAM,EAAE,CAAC,MAAM,EAAE,EAAE;IACnC,OAAO,CAAC,KAAK,CAAC,yCAAyC,CAAC,CAAA;IACxD,OAAO,kBAAkB,CAAC,MAAM,CAAC,UAAU,CAAC,CAAA;AAC7C,CAAC,CAAC,CAAA;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,uBAAuB,GAAG,CAAC,MAAuB,EAAqB,EAAE;IACrF,MAAM,OAAO,GAAG,gBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAA;IAErD,IAAI,CAAC,OAAO,EAAE,CAAC;QACd,OAAO,CAAC,KAAK,CAAC,4BAA4B,MAAM,CAAC,QAAQ,eAAe,CAAC,CAAA;QACzE,MAAM,WAAW,GAAG,gBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;QAChD,OAAO,WAAW,EAAE,CAAC,MAAM,CAAC,CAAA;IAC7B,CAAC;IAED,OAAO,OAAO,CAAC,MAAM,CAAC,CAAA;AACvB,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,2BAA2B,GAAG,KAAK,IAAgC,EAAE;IACjF,MAAM,MAAM,GAAG,mBAAmB,EAAE,CAAA;IACpC,OAAO,uBAAuB,CAAC,MAAM,CAAC,CAAA;AACvC,CAAC,CAAA;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,SAAS,GAAG,CACxB,IAAY,EACZ,UAGI,EAAE,EACc,EAAE;IACtB,MAAM,YAAY,GAAG,OAAO,CAAC,YAAY,IAAI,IAAI,CAAA;IACjD,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,GAAG,CAAA;IAEtC,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,EAAE,CAAA;IACV,CAAC;IAED,MAAM,MAAM,GAAa,EAAE,CAAA;IAC3B,IAAI,KAAK,GAAG,CAAC,CAAA;IAEb,OAAO,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QAC5B,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,YAAY,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;QACvD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,CAAA;QACpC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QAElB,kCAAkC;QAClC,IAAI,GAAG,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC;YACxB,MAAK;QACN,CAAC;QAED,2CAA2C;QAC3C,MAAM,SAAS,GAAG,GAAG,GAAG,OAAO,CAAA;QAC/B,IAAI,SAAS,IAAI,KAAK,EAAE,CAAC;YACxB,qDAAqD;YACrD,KAAK,GAAG,GAAG,CAAA;QACZ,CAAC;aAAM,CAAC;YACP,KAAK,GAAG,SAAS,CAAA;QAClB,CAAC;IACF,CAAC;IAED,OAAO,MAAM,CAAA;AACd,CAAC,CAAA;AAED;;;GAGG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,IAAc,EAAE,IAAc,EAAU,EAAE;IAC1E,IAAI,IAAI,CAAC,MAAM,KAAK,IAAI,CAAC,MAAM,EAAE,CAAC;QACjC,MAAM,IAAI,KAAK,CAAC,8BAA8B,IAAI,CAAC,MAAM,OAAO,IAAI,CAAC,MAAM,EAAE,CAAC,CAAA;IAC/E,CAAC;IAED,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,KAAK,EAAE,GAAG,IAAI,CAAC,MAAM,CAC/C,CAAC,GAAG,EAAE,IAAI,EAAE,CAAC,EAAE,EAAE;QAChB,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAA;QACpB,OAAO;YACN,UAAU,EAAE,GAAG,CAAC,UAAU,GAAG,IAAI,GAAG,IAAI;YACxC,KAAK,EAAE,GAAG,CAAC,KAAK,GAAG,IAAI,GAAG,IAAI;YAC9B,KAAK,EAAE,GAAG,CAAC,KAAK,GAAG,IAAI,GAAG,IAAI;SAC9B,CAAA;IACF,CAAC,EACD,EAAE,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CACrC,CAAA;IAED,IAAI,KAAK,KAAK,CAAC,IAAI,KAAK,KAAK,CAAC,EAAE,CAAC;QAChC,OAAO,CAAC,CAAA;IACT,CAAC;IAED,OAAO,UAAU,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAA;AAC1D,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,MAAgB,EAAY,EAAE;IAC7D,MAAM,SAAS,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,GAAG,GAAG,GAAG,EAAE,CAAC,CAAC,CAAC,CAAA;IAC5E,IAAI,SAAS,KAAK,CAAC,EAAE,CAAC;QACrB,OAAO,MAAM,CAAA;IACd,CAAC;IACD,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,SAAS,CAAC,CAAA;AAC5C,CAAC,CAAA;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAC/B,OAA0B,EAC1B,QAA2B,EACP,EAAE,CAAC,CAAC;IACxB,IAAI,EAAE,GAAG,OAAO,CAAC,IAAI,IAAI,QAAQ,CAAC,IAAI,EAAE;IACxC,KAAK,EAAE,OAAO,CAAC,KAAK;IACpB,UAAU,EAAE,OAAO,CAAC,UAAU;IAC9B,iBAAiB,EAAE,KAAK,EAAE,IAAY,EAAE,EAAE;QACzC,IAAI,CAAC;YACJ,OAAO,MAAM,OAAO,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAA;QAC7C,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,MAAM,QAAQ,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAA;QAC9C,CAAC;IACF,CAAC;IACD,kBAAkB,EAAE,KAAK,EAAE,KAAe,EAAE,EAAE;QAC7C,IAAI,CAAC;YACJ,OAAO,MAAM,OAAO,CAAC,kBAAkB,CAAC,KAAK,CAAC,CAAA;QAC/C,CAAC;QAAC,MAAM,CAAC;YACR,OAAO,MAAM,QAAQ,CAAC,kBAAkB,CAAC,KAAK,CAAC,CAAA;QAChD,CAAC;IACF,CAAC;CACD,CAAC,CAAA"}

package/dist/embeddings.test.d.ts

@@ -0,0 +1,5 @@
+/**
+ * Tests for Embedding utilities
+ */
+export {};
+//# sourceMappingURL=embeddings.test.d.ts.map

package/dist/embeddings.test.d.ts.map

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