@retrivora-ai/rag-engine 0.1.3 → 0.1.4

package/dist/{RAGPipeline-BmkIv1HD.d.ts → Pipeline-Bo6CUCox.d.ts}

@@ -1,120 +1,9 @@
-/**
- * Generic LLM Provider interface.
- * Covers both chat completion and embedding generation so a single
- * provider can handle both responsibilities when appropriate.
- */
-interface ChatMessage {
-    role: 'user' | 'assistant' | 'system';
-    content: string;
-}
-interface ChatOptions {
-    /** Override max tokens for this specific call */
-    maxTokens?: number;
-    /** Override temperature for this specific call */
-    temperature?: number;
-    /** Stop sequences */
-    stop?: string[];
-}
-interface EmbedOptions {
-    /** Override model for this specific embed call */
-    model?: string;
-}
-interface ILLMProvider {
-    /**
-     * Send a chat completion request.
-     * @param messages – the full conversation history
-     * @param context  – retrieved RAG context to inject
-     * @param options  – optional per-call overrides
-     * @returns        – the assistant's reply text
-     */
-    chat(messages: ChatMessage[], context: string, options?: ChatOptions): Promise<string>;
-    /**
-     * Generate an embedding vector for the given text.
-     * @param text    – text to embed
-     * @param options – optional overrides
-     * @returns       – float array (the embedding)
-     */
-    embed(text: string, options?: EmbedOptions): Promise<number[]>;
-    /**
-     * Generate embedding vectors for multiple texts in a single batch.
-     * @param texts   – array of texts to embed
-     * @param options – optional overrides
-     * @returns       – array of float arrays
-     */
-    batchEmbed(texts: string[], options?: EmbedOptions): Promise<number[][]>;
-    /**
-     * Check if the provider endpoint is reachable.
-     */
-    ping(): Promise<boolean>;
-}
-
-/**
- * Generic Vector Database interface.
- * Any adapter (Pinecone, pgVector, REST, Chroma, Qdrant, …) must implement this.
- */
-interface VectorMatch {
-    /** Unique identifier of the stored chunk */
-    id: string;
-    /** Cosine or dot-product similarity score (0–1) */
-    score: number;
-    /** The original text content of the chunk */
-    content: string;
-    /** Arbitrary metadata attached at upsert time */
-    metadata?: Record<string, unknown>;
-}
-interface UpsertDocument {
-    id: string;
-    vector: number[];
-    content: string;
-    metadata?: Record<string, unknown>;
-}
-interface IVectorDB {
-    /**
-     * Initialise the connection (create tables, verify index, etc.)
-     * Called once during RAGPipeline construction.
-     */
-    initialize(): Promise<void>;
-    /**
-     * Insert or update a single vector + content pair.
-     * @param namespace – optional project/tenant namespace
-     */
-    upsert(doc: UpsertDocument, namespace?: string): Promise<void>;
-    /**
-     * Batch upsert for efficient ingestion of many documents.
-     */
-    batchUpsert(docs: UpsertDocument[], namespace?: string): Promise<void>;
-    /**
-     * Find the top-K most similar vectors to the query vector.
-     * @param vector     – embedding of the user query
-     * @param topK       – number of results to return
-     * @param namespace  – optional project/tenant namespace
-     * @param filter     – optional metadata filter (provider-specific shape)
-     */
-    query(vector: number[], topK: number, namespace?: string, filter?: Record<string, unknown>): Promise<VectorMatch[]>;
-    /**
-     * Delete a stored vector by ID.
-     */
-    delete(id: string, namespace?: string): Promise<void>;
-    /**
-     * Delete all vectors in a namespace (full data reset for a project).
-     */
-    deleteNamespace(namespace: string): Promise<void>;
-    /**
-     * Check if the underlying DB is reachable.
-     */
-    ping(): Promise<boolean>;
-    /**
-     * Gracefully close the connection.
-     */
-    disconnect(): Promise<void>;
-}
-
 /**
  * Master configuration interface for Retrivora AI.
  * Each consuming project provides one RagConfig object to drive
  * vector DB selection, LLM selection, embedding, and UI branding.
  */
-type VectorDBProvider = 'pinecone' | 'pgvector' | 'mongodb' | 'rest' | 'universal_rest' | 'custom';
+type VectorDBProvider = 'pinecone' | 'pgvector' | 'postgresql' | 'mongodb' | 'milvus' | 'qdrant' | 'chromadb' | 'redis' | 'weaviate' | 'rest' | 'universal_rest' | 'custom';
 interface VectorDBConfig {
     /** Which vector database to use */
     provider: VectorDBProvider;
@@ -232,67 +121,95 @@ interface RagConfig {
 }
 
 /**
- * RAGPipeline — the central orchestrator for the RAG workflow.
- *
- * Responsibilities:
- *  1. INGEST:  chunk → embed → upsert into vector DB
- *  2. ASK:     embed query → retrieve top-K chunks → call LLM with context
- *
- * It is LLM-agnostic and vector-DB-agnostic; concrete implementations are
- * injected via ILLMProvider and IVectorDB.
+ * Generic LLM Provider interface.
+ * Covers both chat completion and embedding generation so a single
+ * provider can handle both responsibilities when appropriate.
  */
+interface ChatMessage {
+    role: 'user' | 'assistant' | 'system';
+    content: string;
+}
+interface ChatOptions {
+    /** Override max tokens for this specific call */
+    maxTokens?: number;
+    /** Override temperature for this specific call */
+    temperature?: number;
+    /** Stop sequences */
+    stop?: string[];
+}
+interface EmbedOptions {
+    /** Override model for this specific embed call */
+    model?: string;
+}
+interface ILLMProvider {
+    /**
+     * Send a chat completion request.
+     * @param messages – the full conversation history
+     * @param context  – retrieved RAG context to inject
+     * @param options  – optional per-call overrides
+     * @returns        – the assistant's reply text
+     */
+    chat(messages: ChatMessage[], context: string, options?: ChatOptions): Promise<string>;
+    /**
+     * Generate an embedding vector for the given text.
+     * @param text    – text to embed
+     * @param options – optional overrides
+     * @returns       – float array (the embedding)
+     */
+    embed(text: string, options?: EmbedOptions): Promise<number[]>;
+    /**
+     * Generate embedding vectors for multiple texts in a single batch.
+     * @param texts   – array of texts to embed
+     * @param options – optional overrides
+     * @returns       – array of float arrays
+     */
+    batchEmbed(texts: string[], options?: EmbedOptions): Promise<number[][]>;
+    /**
+     * Check if the provider endpoint is reachable.
+     */
+    ping(): Promise<boolean>;
+}
+
+interface VectorMatch {
+    id: string;
+    score: number;
+    content: string;
+    metadata?: Record<string, unknown>;
+}
+interface UpsertDocument {
+    id: string;
+    vector: number[];
+    content: string;
+    metadata?: Record<string, unknown>;
+}
 
 interface IngestDocument {
-    /** Unique document ID (used as chunk ID prefix) */
     docId: string;
-    /** Full text content of the document */
     content: string;
-    /** Optional metadata stored alongside each chunk */
     metadata?: Record<string, unknown>;
 }
 interface ChatResponse {
-    /** The LLM's answer */
     reply: string;
-    /** Retrieved source chunks used to generate the answer */
     sources: VectorMatch[];
 }
-declare class RAGPipeline {
-    private readonly vectorDB;
-    private readonly llmProvider;
-    private readonly embeddingProvider;
-    private readonly chunker;
-    private readonly config;
+/**
+ * Pipeline — orchestrates the RAG flow: Embed → Search → Augment → Generate.
+ */
+declare class Pipeline {
+    private vectorDB;
+    private llmProvider;
+    private embeddingProvider;
+    private chunker;
+    private config;
     private initialised;
-    constructor(config: RagConfig, adapters?: {
-        vectorDB?: IVectorDB;
-        llmProvider?: ILLMProvider;
-        embeddingProvider?: ILLMProvider;
-    });
-    /** Lazily initialise the vector DB connection */
-    init(): Promise<void>;
-    /**
-     * Ingest one or more documents into the vector DB.
-     * Automatically chunks, embeds, and upserts each chunk.
-     */
-    ingest(documents: IngestDocument[], namespace?: string): Promise<{
+    constructor(config: RagConfig);
+    initialize(): Promise<void>;
+    ingest(documents: IngestDocument[], namespace?: string): Promise<Array<{
         docId: string;
         chunksIngested: number;
-    }[]>;
-    /**
-     * Run a full RAG query:
-     *  1. Embed the user question
-     *  2. Retrieve top-K relevant chunks from the vector DB
-     *  3. Filter by score threshold
-     *  4. Build context string
-     *  5. Call the LLM with chat history + context
-     */
+    }>>;
     ask(question: string, history?: ChatMessage[], namespace?: string): Promise<ChatResponse>;
-    health(): Promise<{
-        vectorDB: boolean;
-        llm: boolean;
-    }>;
-    deleteDocument(docId: string, namespace?: string): Promise<void>;
-    deleteProject(namespace?: string): Promise<void>;
+    health(): Promise<Record<string, boolean>>;
 }
 
-export { type ChatMessage as C, type EmbedOptions as E, type ILLMProvider as I, type LLMConfig as L, type RAGConfig as R, type UIConfig as U, type VectorMatch as V, type ChatOptions as a, type ChatResponse as b, type EmbeddingConfig as c, type IVectorDB as d, type IngestDocument as e, type RagConfig as f, type UpsertDocument as g, type VectorDBConfig as h, RAGPipeline as i };
+export { type ChatMessage as C, type EmbedOptions as E, type ILLMProvider as I, type LLMConfig as L, Pipeline as P, type RAGConfig as R, type UIConfig as U, type VectorMatch as V, type ChatOptions as a, type ChatResponse as b, type EmbeddingConfig as c, type IngestDocument as d, type RagConfig as e, type UpsertDocument as f, type VectorDBConfig as g };

package/dist/PostgreSQLProvider-ISNMD3BE.mjs

@@ -0,0 +1,8 @@
+import {
+  PostgreSQLProvider
+} from "./chunk-S5DRHETN.mjs";
+import "./chunk-VOIWNO5O.mjs";
+import "./chunk-I4E63NIC.mjs";
+export {
+  PostgreSQLProvider
+};

package/dist/QdrantProvider-LJWOIGES.mjs

@@ -0,0 +1,8 @@
+import {
+  QdrantProvider
+} from "./chunk-6FODXNUF.mjs";
+import "./chunk-VOIWNO5O.mjs";
+import "./chunk-I4E63NIC.mjs";
+export {
+  QdrantProvider
+};

package/dist/RedisProvider-ASONNYBI.mjs

@@ -0,0 +1,8 @@
+import {
+  RedisProvider
+} from "./chunk-VPNRDXIA.mjs";
+import "./chunk-VOIWNO5O.mjs";
+import "./chunk-I4E63NIC.mjs";
+export {
+  RedisProvider
+};

package/dist/WeaviateProvider-PSDCUGC7.mjs

@@ -0,0 +1,8 @@
+import {
+  WeaviateProvider
+} from "./chunk-V75V7BT2.mjs";
+import "./chunk-VOIWNO5O.mjs";
+import "./chunk-I4E63NIC.mjs";
+export {
+  WeaviateProvider
+};

package/dist/chunk-6FODXNUF.mjs

@@ -0,0 +1,91 @@
+import {
+  BaseVectorProvider
+} from "./chunk-VOIWNO5O.mjs";
+import {
+  __spreadValues
+} from "./chunk-I4E63NIC.mjs";
+
+// src/providers/vectordb/QdrantProvider.ts
+import axios from "axios";
+var QdrantProvider = class extends BaseVectorProvider {
+  constructor(config) {
+    super(config);
+    const opts = config.options;
+    const baseUrl = opts.baseUrl;
+    if (!baseUrl) throw new Error("[QdrantProvider] baseUrl is required");
+    this.http = axios.create({
+      baseURL: baseUrl,
+      headers: __spreadValues({
+        "Content-Type": "application/json"
+      }, opts.apiKey ? { "api-key": opts.apiKey } : {})
+    });
+  }
+  async initialize() {
+    await this.ping();
+  }
+  async upsert(doc, namespace) {
+    await this.batchUpsert([doc], namespace);
+  }
+  async batchUpsert(docs, namespace) {
+    const payload = {
+      points: docs.map((doc) => ({
+        id: doc.id,
+        vector: doc.vector,
+        payload: __spreadValues({
+          content: doc.content,
+          metadata: doc.metadata || {}
+        }, namespace ? { namespace } : {})
+      }))
+    };
+    await this.http.put(`/collections/${this.indexName}/points`, payload);
+  }
+  // eslint-disable-next-line @typescript-eslint/no-unused-vars
+  async query(vector, topK, namespace, _filter) {
+    const payload = {
+      vector,
+      limit: topK,
+      with_payload: true,
+      filter: namespace ? {
+        must: [{ key: "namespace", match: { value: namespace } }]
+      } : void 0
+    };
+    const { data } = await this.http.post(`/collections/${this.indexName}/points/search`, payload);
+    return (data.result || []).map((res) => {
+      var _a, _b;
+      return {
+        id: res["id"],
+        score: res["score"],
+        content: ((_a = res["payload"]) == null ? void 0 : _a.content) || "",
+        metadata: ((_b = res["payload"]) == null ? void 0 : _b.metadata) || {}
+      };
+    });
+  }
+  // eslint-disable-next-line @typescript-eslint/no-unused-vars
+  async delete(id, _namespace) {
+    await this.http.post(`/collections/${this.indexName}/points/delete`, {
+      points: [id]
+    });
+  }
+  // eslint-disable-next-line @typescript-eslint/no-unused-vars
+  async deleteNamespace(_namespace) {
+    await this.http.post(`/collections/${this.indexName}/points/delete`, {
+      filter: {
+        must: [{ key: "namespace", match: { value: _namespace } }]
+      }
+    });
+  }
+  async ping() {
+    try {
+      await this.http.get("/healthz");
+      return true;
+    } catch (e) {
+      return false;
+    }
+  }
+  async disconnect() {
+  }
+};
+
+export {
+  QdrantProvider
+};

package/dist/chunk-7NXI6ZWX.mjs

@@ -0,0 +1,89 @@
+import {
+  BaseVectorProvider
+} from "./chunk-VOIWNO5O.mjs";
+import {
+  __spreadValues
+} from "./chunk-I4E63NIC.mjs";
+
+// src/providers/vectordb/MilvusProvider.ts
+import axios from "axios";
+var MilvusProvider = class extends BaseVectorProvider {
+  constructor(config) {
+    super(config);
+    const opts = config.options;
+    const baseUrl = opts.baseUrl || opts.uri;
+    if (!baseUrl) throw new Error("[MilvusProvider] baseUrl/uri is required");
+    this.http = axios.create({
+      baseURL: baseUrl,
+      headers: __spreadValues(__spreadValues({
+        "Content-Type": "application/json"
+      }, opts.headers || {}), opts.apiKey ? { Authorization: `Bearer ${opts.apiKey}` } : {})
+    });
+  }
+  async initialize() {
+    await this.ping();
+  }
+  async upsert(doc, namespace) {
+    const payload = {
+      collectionName: this.indexName,
+      data: [__spreadValues({
+        vector: doc.vector,
+        content: doc.content,
+        metadata: doc.metadata || {}
+      }, namespace ? { namespace } : {})]
+    };
+    await this.http.post("/v1/vector/upsert", payload);
+  }
+  async batchUpsert(docs, namespace) {
+    const payload = {
+      collectionName: this.indexName,
+      data: docs.map((doc) => __spreadValues({
+        vector: doc.vector,
+        content: doc.content,
+        metadata: doc.metadata || {}
+      }, namespace ? { namespace } : {}))
+    };
+    await this.http.post("/v1/vector/upsert", payload);
+  }
+  // eslint-disable-next-line @typescript-eslint/no-unused-vars
+  async query(vector, topK, namespace, _filter) {
+    const payload = {
+      collectionName: this.indexName,
+      vector,
+      limit: topK,
+      filter: namespace ? `namespace == "${namespace}"` : void 0,
+      outputFields: ["content", "metadata"]
+    };
+    const { data } = await this.http.post("/v1/vector/search", payload);
+    return (data.data || []).map((res) => ({
+      id: String(res["id"]),
+      score: res["distance"],
+      content: res["content"],
+      metadata: res["metadata"]
+    }));
+  }
+  // eslint-disable-next-line @typescript-eslint/no-unused-vars
+  async delete(id, _namespace) {
+    await this.http.post("/v1/vector/delete", {
+      collectionName: this.indexName,
+      id
+    });
+  }
+  // eslint-disable-next-line @typescript-eslint/no-unused-vars
+  async deleteNamespace(_namespace) {
+  }
+  async ping() {
+    try {
+      await this.http.get("/v1/health");
+      return true;
+    } catch (e) {
+      return false;
+    }
+  }
+  async disconnect() {
+  }
+};
+
+export {
+  MilvusProvider
+};

package/dist/chunk-AALIF3AL.mjs

@@ -0,0 +1,91 @@
+import {
+  BaseVectorProvider
+} from "./chunk-VOIWNO5O.mjs";
+import {
+  __spreadValues
+} from "./chunk-I4E63NIC.mjs";
+
+// src/providers/vectordb/PineconeProvider.ts
+import { Pinecone } from "@pinecone-database/pinecone";
+var PineconeProvider = class extends BaseVectorProvider {
+  constructor(config) {
+    super(config);
+    const opts = config.options;
+    if (!opts.apiKey) throw new Error("[PineconeProvider] options.apiKey is required");
+    this.apiKey = opts.apiKey;
+  }
+  async initialize() {
+    var _a, _b;
+    this.client = new Pinecone({ apiKey: this.apiKey });
+    const indexes = await this.client.listIndexes();
+    const names = (_b = (_a = indexes.indexes) == null ? void 0 : _a.map((i) => i.name)) != null ? _b : [];
+    if (!names.includes(this.indexName)) {
+      throw new Error(`[PineconeProvider] Index "${this.indexName}" not found.`);
+    }
+  }
+  index(namespace) {
+    const idx = this.client.index(this.indexName);
+    return namespace ? idx.namespace(namespace) : idx.namespace("");
+  }
+  async upsert(doc, namespace) {
+    var _a;
+    await this.index(namespace).upsert({
+      records: [{
+        id: doc.id,
+        values: doc.vector,
+        metadata: __spreadValues({ content: doc.content }, (_a = doc.metadata) != null ? _a : {})
+      }]
+    });
+  }
+  async batchUpsert(docs, namespace) {
+    const BATCH = 100;
+    for (let i = 0; i < docs.length; i += BATCH) {
+      const records = docs.slice(i, i + BATCH).map((d) => {
+        var _a;
+        return {
+          id: d.id,
+          values: d.vector,
+          metadata: __spreadValues({ content: d.content }, (_a = d.metadata) != null ? _a : {})
+        };
+      });
+      await this.index(namespace).upsert({ records });
+    }
+  }
+  async query(vector, topK, namespace, filter) {
+    var _a;
+    const result = await this.index(namespace).query(__spreadValues({
+      vector,
+      topK,
+      includeMetadata: true
+    }, filter ? { filter } : {}));
+    return ((_a = result.matches) != null ? _a : []).map((m) => {
+      var _a2, _b;
+      return {
+        id: m.id,
+        score: (_a2 = m.score) != null ? _a2 : 0,
+        content: ((_b = m.metadata) == null ? void 0 : _b.content) || "",
+        metadata: m.metadata
+      };
+    });
+  }
+  async delete(id, namespace) {
+    await this.index(namespace).deleteOne({ id });
+  }
+  async deleteNamespace(namespace) {
+    await this.client.index(this.indexName).namespace(namespace).deleteAll();
+  }
+  async ping() {
+    try {
+      await this.client.listIndexes();
+      return true;
+    } catch (e) {
+      return false;
+    }
+  }
+  async disconnect() {
+  }
+};
+
+export {
+  PineconeProvider
+};