@hazeljs/rag 0.2.0-beta.1

package/README.md

@@ -0,0 +1,379 @@
+# @hazeljs/rag
+
+**Retrieval-Augmented Generation (RAG) and Vector Search for HazelJS**
+
+Build powerful AI applications with semantic search, document retrieval, and LLM-augmented responses.
+
+## Features
+
+- 🔍 **Vector Search** - Semantic similarity search using embeddings
+- 📚 **Document Management** - Load, split, and index documents
+- 🤖 **RAG Pipeline** - Complete retrieval-augmented generation workflow
+- 🎯 **Multiple Strategies** - Similarity, MMR (Maximal Marginal Relevance), Hybrid search
+- 🔌 **Pluggable Backends** - Support for Pinecone, Weaviate, Qdrant, ChromaDB, and in-memory
+- 🌐 **Multiple Embedding Providers** - OpenAI, Cohere, HuggingFace
+- ✂️ **Smart Text Splitting** - Recursive text splitter with overlap
+- 📊 **Metadata Filtering** - Filter results by custom metadata
+
+## Installation
+
+```bash
+npm install @hazeljs/rag
+```
+
+### Optional Peer Dependencies
+
+Install the vector store and embedding provider you want to use:
+
+```bash
+# OpenAI Embeddings
+npm install openai
+
+# Vector Stores (choose one or more)
+npm install @pinecone-database/pinecone  # Pinecone
+npm install weaviate-ts-client            # Weaviate
+npm install @qdrant/js-client-rest        # Qdrant
+npm install chromadb                      # ChromaDB
+
+# Additional Embedding Providers
+npm install cohere-ai                     # Cohere
+npm install @huggingface/inference        # HuggingFace
+```
+
+## Quick Start
+
+### Basic RAG Pipeline
+
+```typescript
+import {
+  RAGPipeline,
+  MemoryVectorStore,
+  OpenAIEmbeddings,
+  RecursiveTextSplitter,
+} from '@hazeljs/rag';
+
+// 1. Setup embedding provider
+const embeddings = new OpenAIEmbeddings({
+  apiKey: process.env.OPENAI_API_KEY!,
+  model: 'text-embedding-3-small',
+});
+
+// 2. Setup vector store
+const vectorStore = new MemoryVectorStore(embeddings);
+
+// 3. Setup text splitter
+const textSplitter = new RecursiveTextSplitter({
+  chunkSize: 1000,
+  chunkOverlap: 200,
+});
+
+// 4. Create RAG pipeline
+const rag = new RAGPipeline({
+  vectorStore,
+  embeddingProvider: embeddings,
+  textSplitter,
+  topK: 5,
+});
+
+// 5. Initialize
+await rag.initialize();
+
+// 6. Add documents
+await rag.addDocuments([
+  {
+    content: 'HazelJS is a modern TypeScript framework for building scalable applications.',
+    metadata: { source: 'docs', category: 'intro' },
+  },
+  {
+    content: 'The framework includes built-in support for microservices, caching, and AI.',
+    metadata: { source: 'docs', category: 'features' },
+  },
+]);
+
+// 7. Query
+const result = await rag.query('What is HazelJS?', {
+  topK: 3,
+  filter: { source: 'docs' },
+});
+
+console.log(result.answer);
+console.log(result.sources);
+```
+
+### With LLM Integration
+
+```typescript
+import OpenAI from 'openai';
+
+const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY! });
+
+// Create LLM function
+const llmFunction = async (prompt: string) => {
+  const response = await openai.chat.completions.create({
+    model: 'gpt-4',
+    messages: [{ role: 'user', content: prompt }],
+  });
+  return response.choices[0].message.content || '';
+};
+
+// Create RAG pipeline with LLM
+const rag = new RAGPipeline(config, llmFunction);
+
+// Query with custom prompt
+const result = await rag.query('What is HazelJS?', {
+  llmPrompt: `Based on the following context, answer the question.
+
+Context:
+{context}
+
+Question: {query}
+
+Answer:`,
+});
+```
+
+## Vector Stores
+
+### Memory Vector Store (Development)
+
+```typescript
+import { MemoryVectorStore, OpenAIEmbeddings } from '@hazeljs/rag';
+
+const embeddings = new OpenAIEmbeddings({ apiKey: process.env.OPENAI_API_KEY! });
+const vectorStore = new MemoryVectorStore(embeddings);
+```
+
+### Pinecone (Production)
+
+```typescript
+import { Pinecone } from '@pinecone-database/pinecone';
+import { PineconeVectorStore } from '@hazeljs/rag';
+
+const pinecone = new Pinecone({ apiKey: process.env.PINECONE_API_KEY! });
+const index = pinecone.index('my-index');
+
+const vectorStore = new PineconeVectorStore(index, embeddings);
+```
+
+### Qdrant
+
+```typescript
+import { QdrantClient } from '@qdrant/js-client-rest';
+import { QdrantVectorStore } from '@hazeljs/rag';
+
+const client = new QdrantClient({ url: 'http://localhost:6333' });
+const vectorStore = new QdrantVectorStore(client, embeddings, {
+  collectionName: 'my-collection',
+});
+```
+
+## Embedding Providers
+
+### OpenAI
+
+```typescript
+import { OpenAIEmbeddings } from '@hazeljs/rag';
+
+const embeddings = new OpenAIEmbeddings({
+  apiKey: process.env.OPENAI_API_KEY!,
+  model: 'text-embedding-3-small', // or 'text-embedding-3-large'
+  dimensions: 1536,
+});
+```
+
+### Cohere
+
+```typescript
+import { CohereEmbeddings } from '@hazeljs/rag';
+
+const embeddings = new CohereEmbeddings({
+  apiKey: process.env.COHERE_API_KEY!,
+  model: 'embed-english-v3.0',
+});
+```
+
+### HuggingFace
+
+```typescript
+import { HuggingFaceEmbeddings } from '@hazeljs/rag';
+
+const embeddings = new HuggingFaceEmbeddings({
+  apiKey: process.env.HUGGINGFACE_API_KEY!,
+  model: 'sentence-transformers/all-MiniLM-L6-v2',
+});
+```
+
+## Text Splitting
+
+### Recursive Text Splitter
+
+```typescript
+import { RecursiveTextSplitter } from '@hazeljs/rag';
+
+const splitter = new RecursiveTextSplitter({
+  chunkSize: 1000,
+  chunkOverlap: 200,
+  separators: ['\n\n', '\n', '. ', ' ', ''],
+});
+
+const chunks = splitter.split(longText);
+```
+
+## Retrieval Strategies
+
+### Similarity Search (Default)
+
+```typescript
+const results = await rag.retrieve('query', {
+  topK: 5,
+  strategy: RetrievalStrategy.SIMILARITY,
+});
+```
+
+### MMR (Maximal Marginal Relevance)
+
+Balances relevance and diversity to avoid redundant results:
+
+```typescript
+const results = await rag.retrieve('query', {
+  topK: 5,
+  strategy: RetrievalStrategy.MMR,
+});
+```
+
+### Hybrid Search
+
+Combines keyword and semantic search:
+
+```typescript
+const results = await rag.retrieve('query', {
+  topK: 5,
+  strategy: RetrievalStrategy.HYBRID,
+});
+```
+
+## Metadata Filtering
+
+```typescript
+await rag.addDocuments([
+  {
+    content: 'Document 1',
+    metadata: { category: 'tech', year: 2024 },
+  },
+  {
+    content: 'Document 2',
+    metadata: { category: 'science', year: 2023 },
+  },
+]);
+
+// Filter by metadata
+const results = await rag.query('query', {
+  filter: { category: 'tech', year: 2024 },
+});
+```
+
+## Advanced Usage
+
+### Custom Document Loaders
+
+```typescript
+import { DocumentLoader, Document } from '@hazeljs/rag';
+
+class PDFLoader implements DocumentLoader {
+  constructor(private filePath: string) {}
+
+  async load(): Promise<Document[]> {
+    // Load and parse PDF
+    const text = await this.parsePDF(this.filePath);
+    return [{ content: text, metadata: { source: this.filePath } }];
+  }
+
+  private async parsePDF(path: string): Promise<string> {
+    // PDF parsing logic
+    return '';
+  }
+}
+
+const loader = new PDFLoader('./document.pdf');
+const documents = await loader.load();
+await rag.addDocuments(documents);
+```
+
+### Batch Operations
+
+```typescript
+// Add multiple documents efficiently
+const ids = await rag.addDocuments(documents);
+
+// Delete multiple documents
+await rag.deleteDocuments(ids);
+
+// Clear all documents
+await rag.clear();
+```
+
+## API Reference
+
+### RAGPipeline
+
+```typescript
+class RAGPipeline {
+  constructor(config: RAGConfig, llmFunction?: LLMFunction);
+  initialize(): Promise<void>;
+  addDocuments(documents: Document[]): Promise<string[]>;
+  query(query: string, options?: RAGQueryOptions): Promise<RAGResponse>;
+  retrieve(query: string, options?: QueryOptions, strategy?: RetrievalStrategy): Promise<SearchResult[]>;
+  deleteDocuments(ids: string[]): Promise<void>;
+  clear(): Promise<void>;
+}
+```
+
+### Types
+
+```typescript
+interface Document {
+  id?: string;
+  content: string;
+  metadata?: Record<string, any>;
+  embedding?: number[];
+}
+
+interface SearchResult {
+  id: string;
+  content: string;
+  metadata?: Record<string, any>;
+  score: number;
+  embedding?: number[];
+}
+
+interface RAGResponse {
+  answer: string;
+  sources: SearchResult[];
+  context: string;
+}
+```
+
+## Use Cases
+
+- 📖 **Documentation Search** - Semantic search across documentation
+- 💬 **Chatbots** - Context-aware conversational AI
+- 🔍 **Knowledge Base** - Internal knowledge management
+- 📝 **Content Recommendations** - Similar content discovery
+- 🎓 **Educational Tools** - Q&A systems with source citations
+- 🏢 **Enterprise Search** - Semantic search across company data
+
+## Performance Tips
+
+1. **Batch Operations** - Add documents in batches for better performance
+2. **Chunk Size** - Balance between context and precision (500-1500 tokens)
+3. **Overlap** - Use 10-20% overlap for better context continuity
+4. **Caching** - Cache embeddings for frequently accessed documents
+5. **Filtering** - Use metadata filters to reduce search space
+
+## License
+
+MIT
+
+## Contributing
+
+Contributions are welcome! Please see [CONTRIBUTING.md](../../CONTRIBUTING.md) for details.

package/dist/agentic/agentic-rag.service.d.ts

@@ -0,0 +1,49 @@
+/**
+ * Agentic RAG Service
+ * Main service class demonstrating all agentic RAG capabilities
+ */
+import { VectorStore, SearchResult, QueryOptions } from '../types';
+import { AgenticLLMProvider } from './types';
+export interface AgenticRAGServiceConfig {
+    vectorStore: VectorStore;
+    llmProvider?: AgenticLLMProvider;
+    enableAllFeatures?: boolean;
+}
+export declare class AgenticRAGService {
+    private vectorStore;
+    private llmProvider?;
+    constructor(config: AgenticRAGServiceConfig);
+    /**
+     * Basic retrieval with all agentic features
+     */
+    retrieve(query: string, options?: QueryOptions): Promise<SearchResult[]>;
+    /**
+     * Advanced retrieval with HyDE
+     */
+    hydeRetrieve(query: string, options?: QueryOptions): Promise<SearchResult[]>;
+    /**
+     * Multi-hop reasoning retrieval
+     */
+    deepRetrieve(query: string, options?: QueryOptions): Promise<SearchResult[]>;
+    /**
+     * Context-aware conversational retrieval
+     */
+    conversationalRetrieve(query: string, sessionId: string, options?: QueryOptions): Promise<SearchResult[]>;
+    /**
+     * Verified retrieval with source checking
+     */
+    verifiedRetrieve(query: string, options?: QueryOptions): Promise<SearchResult[]>;
+    /**
+     * Learning-enabled retrieval
+     */
+    learningRetrieve(query: string, options?: QueryOptions): Promise<SearchResult[]>;
+    /**
+     * Provide feedback for active learning
+     */
+    provideFeedback(_resultId: string, _rating: number, _relevant: boolean): Promise<void>;
+    /**
+     * Combined: All features enabled
+     */
+    superRetrieve(query: string, sessionId: string, options?: QueryOptions): Promise<SearchResult[]>;
+}
+//# sourceMappingURL=agentic-rag.service.d.ts.map

package/dist/agentic/agentic-rag.service.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"agentic-rag.service.d.ts","sourceRoot":"","sources":["../../src/agentic/agentic-rag.service.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,WAAW,EAAE,YAAY,EAAE,YAAY,EAAE,MAAM,UAAU,CAAC;AACnE,OAAO,EAAE,kBAAkB,EAAE,MAAM,SAAS,CAAC;AAgB7C,MAAM,WAAW,uBAAuB;IACtC,WAAW,EAAE,WAAW,CAAC;IACzB,WAAW,CAAC,EAAE,kBAAkB,CAAC;IACjC,iBAAiB,CAAC,EAAE,OAAO,CAAC;CAC7B;AAED,qBAAa,iBAAiB;IAC5B,OAAO,CAAC,WAAW,CAAc;IACjC,OAAO,CAAC,WAAW,CAAC,CAAqB;gBAE7B,MAAM,EAAE,uBAAuB;IAK3C;;OAEG;IAKG,QAAQ,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,YAAY,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;IAI9E;;OAEG;IAIG,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,YAAY,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;IAIlF;;OAEG;IAEG,YAAY,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,YAAY,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;IAIlF;;OAEG;IAIG,sBAAsB,CAC1B,KAAK,EAAE,MAAM,EACb,SAAS,EAAE,MAAM,EACjB,OAAO,CAAC,EAAE,YAAY,GACrB,OAAO,CAAC,YAAY,EAAE,CAAC;IAO1B;;OAEG;IAOG,gBAAgB,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,YAAY,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;IAItF;;OAEG;IAGG,gBAAgB,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,EAAE,YAAY,GAAG,OAAO,CAAC,YAAY,EAAE,CAAC;IAItF;;OAEG;IAEG,eAAe,CAAC,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,GAAG,OAAO,CAAC,IAAI,CAAC;IAI5F;;OAEG;IAWG,aAAa,CACjB,KAAK,EAAE,MAAM,EACb,SAAS,EAAE,MAAM,EACjB,OAAO,CAAC,EAAE,YAAY,GACrB,OAAO,CAAC,YAAY,EAAE,CAAC;CAM3B"}

package/dist/agentic/agentic-rag.service.js

@@ -0,0 +1,149 @@
+"use strict";
+/**
+ * Agentic RAG Service
+ * Main service class demonstrating all agentic RAG capabilities
+ */
+var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
+    var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+    if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+    else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+    return c > 3 && r && Object.defineProperty(target, key, r), r;
+};
+var __metadata = (this && this.__metadata) || function (k, v) {
+    if (typeof Reflect === "object" && typeof Reflect.metadata === "function") return Reflect.metadata(k, v);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.AgenticRAGService = void 0;
+const decorators_1 = require("./decorators");
+class AgenticRAGService {
+    constructor(config) {
+        this.vectorStore = config.vectorStore;
+        this.llmProvider = config.llmProvider;
+    }
+    /**
+     * Basic retrieval with all agentic features
+     */
+    async retrieve(query, options) {
+        return this.vectorStore.search(query, options);
+    }
+    /**
+     * Advanced retrieval with HyDE
+     */
+    async hydeRetrieve(query, options) {
+        return this.vectorStore.search(query, options);
+    }
+    /**
+     * Multi-hop reasoning retrieval
+     */
+    async deepRetrieve(query, options) {
+        return this.vectorStore.search(query, options);
+    }
+    /**
+     * Context-aware conversational retrieval
+     */
+    async conversationalRetrieve(query, sessionId, options) {
+        return this.vectorStore.search(query, {
+            ...options,
+            sessionId: sessionId,
+        });
+    }
+    /**
+     * Verified retrieval with source checking
+     */
+    async verifiedRetrieve(query, options) {
+        return this.vectorStore.search(query, options);
+    }
+    /**
+     * Learning-enabled retrieval
+     */
+    async learningRetrieve(query, options) {
+        return this.vectorStore.search(query, options);
+    }
+    /**
+     * Provide feedback for active learning
+     */
+    async provideFeedback(_resultId, _rating, _relevant) {
+        // Feedback is stored by decorator
+    }
+    /**
+     * Combined: All features enabled
+     */
+    async superRetrieve(query, sessionId, options) {
+        return this.vectorStore.search(query, {
+            ...options,
+            sessionId: sessionId,
+        });
+    }
+}
+exports.AgenticRAGService = AgenticRAGService;
+__decorate([
+    (0, decorators_1.QueryPlanner)({ decompose: true, maxSubQueries: 5, parallel: true }),
+    (0, decorators_1.SelfReflective)({ maxIterations: 3, qualityThreshold: 0.8 }),
+    (0, decorators_1.AdaptiveRetrieval)({ autoSelect: true, contextAware: true }),
+    (0, decorators_1.Cached)({ ttl: 3600 }),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "retrieve", null);
+__decorate([
+    (0, decorators_1.HyDE)({ generateHypothesis: true, numHypotheses: 3 }),
+    (0, decorators_1.CorrectiveRAG)({ relevanceThreshold: 0.7, fallbackToWeb: true }),
+    (0, decorators_1.Cached)({ ttl: 1800 }),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "hydeRetrieve", null);
+__decorate([
+    (0, decorators_1.MultiHop)({ maxHops: 3, strategy: 'breadth-first' }),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "deepRetrieve", null);
+__decorate([
+    (0, decorators_1.ContextAware)({ windowSize: 5, entityTracking: true, topicModeling: true }),
+    (0, decorators_1.QueryRewriter)({ techniques: ['expansion', 'synonym'], llmBased: true }),
+    (0, decorators_1.Cached)({ ttl: 600 }),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, String, Object]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "conversationalRetrieve", null);
+__decorate([
+    (0, decorators_1.SourceVerification)({
+        checkFreshness: true,
+        verifyAuthority: true,
+        requireCitations: true,
+    }),
+    (0, decorators_1.SelfReflective)({ maxIterations: 2, qualityThreshold: 0.85 }),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "verifiedRetrieve", null);
+__decorate([
+    (0, decorators_1.ActiveLearning)({ feedbackEnabled: true, retrainThreshold: 100 }),
+    (0, decorators_1.AdaptiveRetrieval)({ autoSelect: true }),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "learningRetrieve", null);
+__decorate([
+    (0, decorators_1.Feedback)(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Number, Boolean]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "provideFeedback", null);
+__decorate([
+    (0, decorators_1.QueryPlanner)({ decompose: true, maxSubQueries: 5 }),
+    (0, decorators_1.SelfReflective)({ maxIterations: 3, qualityThreshold: 0.8 }),
+    (0, decorators_1.AdaptiveRetrieval)({ autoSelect: true, contextAware: true }),
+    (0, decorators_1.HyDE)({ generateHypothesis: true, numHypotheses: 2 }),
+    (0, decorators_1.CorrectiveRAG)({ relevanceThreshold: 0.7 }),
+    (0, decorators_1.ContextAware)({ windowSize: 5, entityTracking: true }),
+    (0, decorators_1.QueryRewriter)({ techniques: ['expansion', 'synonym'] }),
+    (0, decorators_1.SourceVerification)({ checkFreshness: true, verifyAuthority: true }),
+    (0, decorators_1.ActiveLearning)({ feedbackEnabled: true }),
+    (0, decorators_1.Cached)({ ttl: 3600 }),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, String, Object]),
+    __metadata("design:returntype", Promise)
+], AgenticRAGService.prototype, "superRetrieve", null);
+//# sourceMappingURL=agentic-rag.service.js.map

package/dist/agentic/agentic-rag.service.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"agentic-rag.service.js","sourceRoot":"","sources":["../../src/agentic/agentic-rag.service.ts"],"names":[],"mappings":";AAAA;;;GAGG;;;;;;;;;;;;AAIH,6CAasB;AAQtB,MAAa,iBAAiB;IAI5B,YAAY,MAA+B;QACzC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;QACtC,IAAI,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;IACxC,CAAC;IAED;;OAEG;IAKG,AAAN,KAAK,CAAC,QAAQ,CAAC,KAAa,EAAE,OAAsB;QAClD,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACjD,CAAC;IAED;;OAEG;IAIG,AAAN,KAAK,CAAC,YAAY,CAAC,KAAa,EAAE,OAAsB;QACtD,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACjD,CAAC;IAED;;OAEG;IAEG,AAAN,KAAK,CAAC,YAAY,CAAC,KAAa,EAAE,OAAsB;QACtD,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACjD,CAAC;IAED;;OAEG;IAIG,AAAN,KAAK,CAAC,sBAAsB,CAC1B,KAAa,EACb,SAAiB,EACjB,OAAsB;QAEtB,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE;YACpC,GAAG,OAAO;YACV,SAAS,EAAE,SAAoB;SAChB,CAAC,CAAC;IACrB,CAAC;IAED;;OAEG;IAOG,AAAN,KAAK,CAAC,gBAAgB,CAAC,KAAa,EAAE,OAAsB;QAC1D,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACjD,CAAC;IAED;;OAEG;IAGG,AAAN,KAAK,CAAC,gBAAgB,CAAC,KAAa,EAAE,OAAsB;QAC1D,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACjD,CAAC;IAED;;OAEG;IAEG,AAAN,KAAK,CAAC,eAAe,CAAC,SAAiB,EAAE,OAAe,EAAE,SAAkB;QAC1E,kCAAkC;IACpC,CAAC;IAED;;OAEG;IAWG,AAAN,KAAK,CAAC,aAAa,CACjB,KAAa,EACb,SAAiB,EACjB,OAAsB;QAEtB,OAAO,IAAI,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE;YACpC,GAAG,OAAO;YACV,SAAS,EAAE,SAAoB;SAChB,CAAC,CAAC;IACrB,CAAC;CACF;AA5GD,8CA4GC;AA5FO;IAJL,IAAA,yBAAY,EAAC,EAAE,SAAS,EAAE,IAAI,EAAE,aAAa,EAAE,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;IACnE,IAAA,2BAAc,EAAC,EAAE,aAAa,EAAE,CAAC,EAAE,gBAAgB,EAAE,GAAG,EAAE,CAAC;IAC3D,IAAA,8BAAiB,EAAC,EAAE,UAAU,EAAE,IAAI,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC;IAC3D,IAAA,mBAAM,EAAC,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC;;;;iDAGrB;AAQK;IAHL,IAAA,iBAAI,EAAC,EAAE,kBAAkB,EAAE,IAAI,EAAE,aAAa,EAAE,CAAC,EAAE,CAAC;IACpD,IAAA,0BAAa,EAAC,EAAE,kBAAkB,EAAE,GAAG,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;IAC/D,IAAA,mBAAM,EAAC,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC;;;;qDAGrB;AAMK;IADL,IAAA,qBAAQ,EAAC,EAAE,OAAO,EAAE,CAAC,EAAE,QAAQ,EAAE,eAAe,EAAE,CAAC;;;;qDAGnD;AAQK;IAHL,IAAA,yBAAY,EAAC,EAAE,UAAU,EAAE,CAAC,EAAE,cAAc,EAAE,IAAI,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;IAC1E,IAAA,0BAAa,EAAC,EAAE,UAAU,EAAE,CAAC,WAAW,EAAE,SAAS,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;IACvE,IAAA,mBAAM,EAAC,EAAE,GAAG,EAAE,GAAG,EAAE,CAAC;;;;+DAUpB;AAWK;IANL,IAAA,+BAAkB,EAAC;QAClB,cAAc,EAAE,IAAI;QACpB,eAAe,EAAE,IAAI;QACrB,gBAAgB,EAAE,IAAI;KACvB,CAAC;IACD,IAAA,2BAAc,EAAC,EAAE,aAAa,EAAE,CAAC,EAAE,gBAAgB,EAAE,IAAI,EAAE,CAAC;;;;yDAG5D;AAOK;IAFL,IAAA,2BAAc,EAAC,EAAE,eAAe,EAAE,IAAI,EAAE,gBAAgB,EAAE,GAAG,EAAE,CAAC;IAChE,IAAA,8BAAiB,EAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC;;;;yDAGvC;AAMK;IADL,IAAA,qBAAQ,GAAE;;;;wDAGV;AAeK;IAVL,IAAA,yBAAY,EAAC,EAAE,SAAS,EAAE,IAAI,EAAE,aAAa,EAAE,CAAC,EAAE,CAAC;IACnD,IAAA,2BAAc,EAAC,EAAE,aAAa,EAAE,CAAC,EAAE,gBAAgB,EAAE,GAAG,EAAE,CAAC;IAC3D,IAAA,8BAAiB,EAAC,EAAE,UAAU,EAAE,IAAI,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC;IAC3D,IAAA,iBAAI,EAAC,EAAE,kBAAkB,EAAE,IAAI,EAAE,aAAa,EAAE,CAAC,EAAE,CAAC;IACpD,IAAA,0BAAa,EAAC,EAAE,kBAAkB,EAAE,GAAG,EAAE,CAAC;IAC1C,IAAA,yBAAY,EAAC,EAAE,UAAU,EAAE,CAAC,EAAE,cAAc,EAAE,IAAI,EAAE,CAAC;IACrD,IAAA,0BAAa,EAAC,EAAE,UAAU,EAAE,CAAC,WAAW,EAAE,SAAS,CAAC,EAAE,CAAC;IACvD,IAAA,+BAAkB,EAAC,EAAE,cAAc,EAAE,IAAI,EAAE,eAAe,EAAE,IAAI,EAAE,CAAC;IACnE,IAAA,2BAAc,EAAC,EAAE,eAAe,EAAE,IAAI,EAAE,CAAC;IACzC,IAAA,mBAAM,EAAC,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC;;;;sDAUrB"}

package/dist/agentic/decorators/active-learning.decorator.d.ts

@@ -0,0 +1,19 @@
+/**
+ * Active Learning Decorator
+ * Learns from user feedback to improve retrieval
+ */
+import 'reflect-metadata';
+export interface ActiveLearningConfig {
+    feedbackEnabled?: boolean;
+    retrainThreshold?: number;
+    storageKey?: string;
+}
+export declare function ActiveLearning(config?: ActiveLearningConfig): MethodDecorator;
+export declare function Feedback(): MethodDecorator;
+export declare function getFeedbackStats(storageKey?: string): {
+    totalFeedback: number;
+    avgRating: number;
+    relevantCount: number;
+};
+export declare function clearFeedback(storageKey?: string): void;
+//# sourceMappingURL=active-learning.decorator.d.ts.map

package/dist/agentic/decorators/active-learning.decorator.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"active-learning.decorator.d.ts","sourceRoot":"","sources":["../../../src/agentic/decorators/active-learning.decorator.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,kBAAkB,CAAC;AAG1B,MAAM,WAAW,oBAAoB;IACnC,eAAe,CAAC,EAAE,OAAO,CAAC;IAC1B,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAID,wBAAgB,cAAc,CAAC,MAAM,GAAE,oBAAyB,GAAG,eAAe,CAmBjF;AAED,wBAAgB,QAAQ,IAAI,eAAe,CAkB1C;AAqDD,wBAAgB,gBAAgB,CAAC,UAAU,GAAE,MAAkB,GAAG;IAChE,aAAa,EAAE,MAAM,CAAC;IACtB,SAAS,EAAE,MAAM,CAAC;IAClB,aAAa,EAAE,MAAM,CAAC;CACvB,CAeA;AAED,wBAAgB,aAAa,CAAC,UAAU,GAAE,MAAkB,GAAG,IAAI,CAElE"}