@neural-tools/vector-db 0.1.3 → 0.1.5

package/dist/index.d.mts

@@ -0,0 +1,110 @@
+interface VectorDBConfig {
+    provider: 'pinecone' | 'qdrant' | 'chromadb' | 'local';
+    apiKey?: string;
+    endpoint?: string;
+    dimension?: number;
+    metric?: 'cosine' | 'euclidean' | 'dotproduct';
+    namespace?: string;
+}
+interface Vector {
+    id: string;
+    values: number[];
+    metadata?: Record<string, any>;
+}
+interface QueryResult {
+    id: string;
+    score: number;
+    metadata?: Record<string, any>;
+    values?: number[];
+}
+interface QueryOptions {
+    topK?: number;
+    filter?: Record<string, any>;
+    includeValues?: boolean;
+    includeMetadata?: boolean;
+}
+interface UpsertOptions {
+    namespace?: string;
+    batch?: boolean;
+}
+declare abstract class VectorStore {
+    protected config: VectorDBConfig;
+    constructor(config: VectorDBConfig);
+    /**
+     * Initialize the vector store connection
+     */
+    abstract connect(): Promise<void>;
+    /**
+     * Insert or update vectors
+     */
+    abstract upsert(vectors: Vector[], options?: UpsertOptions): Promise<void>;
+    /**
+     * Query for similar vectors
+     */
+    abstract query(vector: number[], options?: QueryOptions): Promise<QueryResult[]>;
+    /**
+     * Delete vectors by ID
+     */
+    abstract delete(ids: string[], namespace?: string): Promise<void>;
+    /**
+     * Get vector by ID
+     */
+    abstract fetch(ids: string[], namespace?: string): Promise<Vector[]>;
+    /**
+     * Delete all vectors in a namespace
+     */
+    abstract deleteNamespace(namespace: string): Promise<void>;
+    /**
+     * Close the connection
+     */
+    abstract disconnect(): Promise<void>;
+}
+
+/**
+ * Simple in-memory vector store for local development and testing
+ */
+declare class LocalVectorStore extends VectorStore {
+    private vectors;
+    connect(): Promise<void>;
+    upsert(vectors: Vector[], options?: UpsertOptions): Promise<void>;
+    query(vector: number[], options?: QueryOptions): Promise<QueryResult[]>;
+    delete(ids: string[]): Promise<void>;
+    fetch(ids: string[]): Promise<Vector[]>;
+    deleteNamespace(): Promise<void>;
+    disconnect(): Promise<void>;
+    /**
+     * Calculate cosine similarity between two vectors
+     */
+    private cosineSimilarity;
+    /**
+     * Check if metadata matches filter
+     */
+    private matchesFilter;
+    /**
+     * Get all vectors (for testing/debugging)
+     */
+    getAll(): Vector[];
+    /**
+     * Get count of vectors
+     */
+    count(): number;
+}
+
+/**
+ * Create a vector store instance based on the configuration
+ */
+declare function createVectorStore(config: VectorDBConfig): Promise<VectorStore>;
+/**
+ * Helper function to create embeddings from text
+ * In production, this would call OpenAI, Anthropic, or other embedding APIs
+ */
+declare function createEmbedding(text: string): Promise<number[]>;
+/**
+ * Batch text into chunks for embedding
+ */
+declare function chunkText(text: string, options?: {
+    chunkSize?: number;
+    overlap?: number;
+}): string[];
+
+export { LocalVectorStore, type QueryOptions, type QueryResult, type UpsertOptions, type Vector, type VectorDBConfig, VectorStore, chunkText, createEmbedding, createVectorStore };

package/dist/index.d.ts

@@ -1,19 +1,110 @@
-import { VectorDBConfig, VectorStore } from './types';
-export * from './types';
-export { LocalVectorStore } from './local-store';
+interface VectorDBConfig {
+    provider: 'pinecone' | 'qdrant' | 'chromadb' | 'local';
+    apiKey?: string;
+    endpoint?: string;
+    dimension?: number;
+    metric?: 'cosine' | 'euclidean' | 'dotproduct';
+    namespace?: string;
+}
+interface Vector {
+    id: string;
+    values: number[];
+    metadata?: Record<string, any>;
+}
+interface QueryResult {
+    id: string;
+    score: number;
+    metadata?: Record<string, any>;
+    values?: number[];
+}
+interface QueryOptions {
+    topK?: number;
+    filter?: Record<string, any>;
+    includeValues?: boolean;
+    includeMetadata?: boolean;
+}
+interface UpsertOptions {
+    namespace?: string;
+    batch?: boolean;
+}
+declare abstract class VectorStore {
+    protected config: VectorDBConfig;
+    constructor(config: VectorDBConfig);
+    /**
+     * Initialize the vector store connection
+     */
+    abstract connect(): Promise<void>;
+    /**
+     * Insert or update vectors
+     */
+    abstract upsert(vectors: Vector[], options?: UpsertOptions): Promise<void>;
+    /**
+     * Query for similar vectors
+     */
+    abstract query(vector: number[], options?: QueryOptions): Promise<QueryResult[]>;
+    /**
+     * Delete vectors by ID
+     */
+    abstract delete(ids: string[], namespace?: string): Promise<void>;
+    /**
+     * Get vector by ID
+     */
+    abstract fetch(ids: string[], namespace?: string): Promise<Vector[]>;
+    /**
+     * Delete all vectors in a namespace
+     */
+    abstract deleteNamespace(namespace: string): Promise<void>;
+    /**
+     * Close the connection
+     */
+    abstract disconnect(): Promise<void>;
+}
+
+/**
+ * Simple in-memory vector store for local development and testing
+ */
+declare class LocalVectorStore extends VectorStore {
+    private vectors;
+    connect(): Promise<void>;
+    upsert(vectors: Vector[], options?: UpsertOptions): Promise<void>;
+    query(vector: number[], options?: QueryOptions): Promise<QueryResult[]>;
+    delete(ids: string[]): Promise<void>;
+    fetch(ids: string[]): Promise<Vector[]>;
+    deleteNamespace(): Promise<void>;
+    disconnect(): Promise<void>;
+    /**
+     * Calculate cosine similarity between two vectors
+     */
+    private cosineSimilarity;
+    /**
+     * Check if metadata matches filter
+     */
+    private matchesFilter;
+    /**
+     * Get all vectors (for testing/debugging)
+     */
+    getAll(): Vector[];
+    /**
+     * Get count of vectors
+     */
+    count(): number;
+}
+
 /**
  * Create a vector store instance based on the configuration
  */
-export declare function createVectorStore(config: VectorDBConfig): Promise<VectorStore>;
+declare function createVectorStore(config: VectorDBConfig): Promise<VectorStore>;
 /**
  * Helper function to create embeddings from text
  * In production, this would call OpenAI, Anthropic, or other embedding APIs
  */
-export declare function createEmbedding(text: string): Promise<number[]>;
+declare function createEmbedding(text: string): Promise<number[]>;
 /**
  * Batch text into chunks for embedding
  */
-export declare function chunkText(text: string, options?: {
+declare function chunkText(text: string, options?: {
     chunkSize?: number;
     overlap?: number;
 }): string[];
+
+export { LocalVectorStore, type QueryOptions, type QueryResult, type UpsertOptions, type Vector, type VectorDBConfig, VectorStore, chunkText, createEmbedding, createVectorStore };

package/dist/index.js

@@ -1,87 +1 @@
-"use strict";
-var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    var desc = Object.getOwnPropertyDescriptor(m, k);
-    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
-      desc = { enumerable: true, get: function() { return m[k]; } };
-    }
-    Object.defineProperty(o, k2, desc);
-}) : (function(o, m, k, k2) {
-    if (k2 === undefined) k2 = k;
-    o[k2] = m[k];
-}));
-var __exportStar = (this && this.__exportStar) || function(m, exports) {
-    for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
-};
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.LocalVectorStore = void 0;
-exports.createVectorStore = createVectorStore;
-exports.createEmbedding = createEmbedding;
-exports.chunkText = chunkText;
-const local_store_1 = require("./local-store");
-const core_1 = require("@neural-tools/core");
-__exportStar(require("./types"), exports);
-var local_store_2 = require("./local-store");
-Object.defineProperty(exports, "LocalVectorStore", { enumerable: true, get: function () { return local_store_2.LocalVectorStore; } });
-/**
- * Create a vector store instance based on the configuration
- */
-async function createVectorStore(config) {
-    switch (config.provider) {
-        case 'local':
-            return new local_store_1.LocalVectorStore(config);
-        case 'pinecone':
-            await (0, core_1.requireFeature)('vector-db', 'Vector Database Integration');
-            throw new Error('Pinecone integration coming soon. Use local provider for now.');
-        case 'qdrant':
-            await (0, core_1.requireFeature)('vector-db', 'Vector Database Integration');
-            throw new Error('Qdrant integration coming soon. Use local provider for now.');
-        case 'chromadb':
-            await (0, core_1.requireFeature)('vector-db', 'Vector Database Integration');
-            throw new Error('ChromaDB integration coming soon. Use local provider for now.');
-        default:
-            throw new Error(`Unknown vector store provider: ${config.provider}`);
-    }
-}
-/**
- * Helper function to create embeddings from text
- * In production, this would call OpenAI, Anthropic, or other embedding APIs
- */
-async function createEmbedding(text) {
-    // Placeholder: Simple hash-based embedding for testing
-    // In production, use proper embedding models
-    const hash = simpleHash(text);
-    const dimension = 384; // Common embedding dimension
-    const embedding = new Array(dimension).fill(0);
-    // Create a deterministic but distributed embedding from hash
-    for (let i = 0; i < dimension; i++) {
-        embedding[i] = Math.sin(hash + i) * Math.cos(hash / (i + 1));
-    }
-    // Normalize
-    const norm = Math.sqrt(embedding.reduce((sum, val) => sum + val * val, 0));
-    return embedding.map(val => val / norm);
-}
-function simpleHash(str) {
-    let hash = 0;
-    for (let i = 0; i < str.length; i++) {
-        const char = str.charCodeAt(i);
-        hash = ((hash << 5) - hash) + char;
-        hash = hash & hash; // Convert to 32-bit integer
-    }
-    return Math.abs(hash);
-}
-/**
- * Batch text into chunks for embedding
- */
-function chunkText(text, options = {}) {
-    const chunkSize = options.chunkSize || 500;
-    const overlap = options.overlap || 50;
-    const chunks = [];
-    let start = 0;
-    while (start < text.length) {
-        const end = Math.min(start + chunkSize, text.length);
-        chunks.push(text.slice(start, end));
-        start += chunkSize - overlap;
-    }
-    return chunks;
-}
+"use strict";var d=Object.defineProperty;var m=Object.getOwnPropertyDescriptor;var p=Object.getOwnPropertyNames;var f=Object.prototype.hasOwnProperty;var h=(t,e)=>{for(var r in e)d(t,r,{get:e[r],enumerable:!0})},g=(t,e,r,o)=>{if(e&&typeof e=="object"||typeof e=="function")for(let n of p(e))!f.call(t,n)&&n!==r&&d(t,n,{get:()=>e[n],enumerable:!(o=m(e,n))||o.enumerable});return t};var v=t=>g(d({},"__esModule",{value:!0}),t);var P={};h(P,{LocalVectorStore:()=>c,VectorStore:()=>a,chunkText:()=>w,createEmbedding:()=>y,createVectorStore:()=>b});module.exports=v(P);var a=class{config;constructor(e){this.config=e}};var c=class extends a{vectors=new Map;async connect(){}async upsert(e,r){for(let o of e)this.vectors.set(o.id,o)}async query(e,r){let o=r?.topK||10,n=[];for(let[s,i]of this.vectors){if(r?.filter&&!this.matchesFilter(i.metadata,r.filter))continue;let l=this.cosineSimilarity(e,i.values);n.push({id:s,score:l,metadata:r?.includeMetadata!==!1?i.metadata:void 0,values:r?.includeValues?i.values:void 0})}return n.sort((s,i)=>i.score-s.score),n.slice(0,o)}async delete(e){for(let r of e)this.vectors.delete(r)}async fetch(e){let r=[];for(let o of e){let n=this.vectors.get(o);n&&r.push(n)}return r}async deleteNamespace(){this.vectors.clear()}async disconnect(){}cosineSimilarity(e,r){if(e.length!==r.length)throw new Error("Vectors must have same length");let o=0,n=0,s=0;for(let i=0;i<e.length;i++)o+=e[i]*r[i],n+=e[i]*e[i],s+=r[i]*r[i];return o/(Math.sqrt(n)*Math.sqrt(s))}matchesFilter(e,r){if(!e)return!1;for(let[o,n]of Object.entries(r))if(e[o]!==n)return!1;return!0}getAll(){return Array.from(this.vectors.values())}count(){return this.vectors.size}};var u=require("@neural-tools/core");async function b(t){switch(t.provider){case"local":return new c(t);case"pinecone":throw await(0,u.requireFeature)("vector-db","Vector Database Integration"),new Error("Pinecone integration coming soon. Use local provider for now.");case"qdrant":throw await(0,u.requireFeature)("vector-db","Vector Database Integration"),new Error("Qdrant integration coming soon. Use local provider for now.");case"chromadb":throw await(0,u.requireFeature)("vector-db","Vector Database Integration"),new Error("ChromaDB integration coming soon. Use local provider for now.");default:throw new Error(`Unknown vector store provider: ${t.provider}`)}}async function y(t){let e=V(t),r=384,o=new Array(r).fill(0);for(let s=0;s<r;s++)o[s]=Math.sin(e+s)*Math.cos(e/(s+1));let n=Math.sqrt(o.reduce((s,i)=>s+i*i,0));return o.map(s=>s/n)}function V(t){let e=0;for(let r=0;r<t.length;r++){let o=t.charCodeAt(r);e=(e<<5)-e+o,e=e&e}return Math.abs(e)}function w(t,e={}){let r=e.chunkSize||500,o=e.overlap||50,n=[],s=0;for(;s<t.length;){let i=Math.min(s+r,t.length);n.push(t.slice(s,i)),s+=r-o}return n}0&&(module.exports={LocalVectorStore,VectorStore,chunkText,createEmbedding,createVectorStore});

package/dist/index.mjs

@@ -0,0 +1 @@
+var a=class{config;constructor(e){this.config=e}};var c=class extends a{vectors=new Map;async connect(){}async upsert(e,r){for(let n of e)this.vectors.set(n.id,n)}async query(e,r){let n=r?.topK||10,s=[];for(let[t,o]of this.vectors){if(r?.filter&&!this.matchesFilter(o.metadata,r.filter))continue;let d=this.cosineSimilarity(e,o.values);s.push({id:t,score:d,metadata:r?.includeMetadata!==!1?o.metadata:void 0,values:r?.includeValues?o.values:void 0})}return s.sort((t,o)=>o.score-t.score),s.slice(0,n)}async delete(e){for(let r of e)this.vectors.delete(r)}async fetch(e){let r=[];for(let n of e){let s=this.vectors.get(n);s&&r.push(s)}return r}async deleteNamespace(){this.vectors.clear()}async disconnect(){}cosineSimilarity(e,r){if(e.length!==r.length)throw new Error("Vectors must have same length");let n=0,s=0,t=0;for(let o=0;o<e.length;o++)n+=e[o]*r[o],s+=e[o]*e[o],t+=r[o]*r[o];return n/(Math.sqrt(s)*Math.sqrt(t))}matchesFilter(e,r){if(!e)return!1;for(let[n,s]of Object.entries(r))if(e[n]!==s)return!1;return!0}getAll(){return Array.from(this.vectors.values())}count(){return this.vectors.size}};import{requireFeature as u}from"@neural-tools/core";async function w(i){switch(i.provider){case"local":return new c(i);case"pinecone":throw await u("vector-db","Vector Database Integration"),new Error("Pinecone integration coming soon. Use local provider for now.");case"qdrant":throw await u("vector-db","Vector Database Integration"),new Error("Qdrant integration coming soon. Use local provider for now.");case"chromadb":throw await u("vector-db","Vector Database Integration"),new Error("ChromaDB integration coming soon. Use local provider for now.");default:throw new Error(`Unknown vector store provider: ${i.provider}`)}}async function P(i){let e=l(i),r=384,n=new Array(r).fill(0);for(let t=0;t<r;t++)n[t]=Math.sin(e+t)*Math.cos(e/(t+1));let s=Math.sqrt(n.reduce((t,o)=>t+o*o,0));return n.map(t=>t/s)}function l(i){let e=0;for(let r=0;r<i.length;r++){let n=i.charCodeAt(r);e=(e<<5)-e+n,e=e&e}return Math.abs(e)}function x(i,e={}){let r=e.chunkSize||500,n=e.overlap||50,s=[],t=0;for(;t<i.length;){let o=Math.min(t+r,i.length);s.push(i.slice(t,o)),t+=r-n}return s}export{c as LocalVectorStore,a as VectorStore,x as chunkText,P as createEmbedding,w as createVectorStore};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@neural-tools/vector-db",
-  "version": "0.1.3",
+  "version": "0.1.5",
   "description": "Vector database abstraction layer for Neural Tools",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -10,11 +10,24 @@
   },
   "repository": {
     "type": "git",
-    "url": "https://github.com/MacLeanLuke/ai-toolkit.git",
+    "url": "https://github.com/MacLeanLuke/neural-tools.git",
     "directory": "packages/vector-db"
   },
+  "homepage": "https://neural-tools.com/docs/vector-db.html",
+  "bugs": {
+    "url": "https://github.com/MacLeanLuke/neural-tools/issues"
+  },
+  "keywords": [
+    "vector-database",
+    "pinecone",
+    "chroma",
+    "qdrant",
+    "embeddings",
+    "neural-tools",
+    "ai"
+  ],
   "dependencies": {
-    "@neural-tools/core": "0.1.3"
+    "@neural-tools/core": "0.1.5"
   },
   "devDependencies": {
     "@types/node": "^20.11.5",
@@ -40,8 +53,8 @@
     "dist"
   ],
   "scripts": {
-    "build": "tsc",
-    "dev": "tsc --watch",
+    "build": "tsup",
+    "dev": "tsup --watch",
     "clean": "rm -rf dist",
     "test": "echo 'Tests coming soon'"
   }

package/dist/local-store.d.ts

@@ -1,30 +0,0 @@
-import { VectorStore, Vector, QueryResult, QueryOptions, UpsertOptions } from './types';
-/**
- * Simple in-memory vector store for local development and testing
- */
-export declare class LocalVectorStore extends VectorStore {
-    private vectors;
-    connect(): Promise<void>;
-    upsert(vectors: Vector[], options?: UpsertOptions): Promise<void>;
-    query(vector: number[], options?: QueryOptions): Promise<QueryResult[]>;
-    delete(ids: string[]): Promise<void>;
-    fetch(ids: string[]): Promise<Vector[]>;
-    deleteNamespace(): Promise<void>;
-    disconnect(): Promise<void>;
-    /**
-     * Calculate cosine similarity between two vectors
-     */
-    private cosineSimilarity;
-    /**
-     * Check if metadata matches filter
-     */
-    private matchesFilter;
-    /**
-     * Get all vectors (for testing/debugging)
-     */
-    getAll(): Vector[];
-    /**
-     * Get count of vectors
-     */
-    count(): number;
-}

package/dist/local-store.js

@@ -1,103 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.LocalVectorStore = void 0;
-const types_1 = require("./types");
-/**
- * Simple in-memory vector store for local development and testing
- */
-class LocalVectorStore extends types_1.VectorStore {
-    vectors = new Map();
-    async connect() {
-        // No connection needed for local store
-    }
-    async upsert(vectors, options) {
-        for (const vector of vectors) {
-            this.vectors.set(vector.id, vector);
-        }
-    }
-    async query(vector, options) {
-        const topK = options?.topK || 10;
-        const results = [];
-        // Calculate similarity for all vectors
-        for (const [id, storedVector] of this.vectors) {
-            // Skip if filter doesn't match
-            if (options?.filter && !this.matchesFilter(storedVector.metadata, options.filter)) {
-                continue;
-            }
-            const score = this.cosineSimilarity(vector, storedVector.values);
-            results.push({
-                id,
-                score,
-                metadata: options?.includeMetadata !== false ? storedVector.metadata : undefined,
-                values: options?.includeValues ? storedVector.values : undefined
-            });
-        }
-        // Sort by score descending and take topK
-        results.sort((a, b) => b.score - a.score);
-        return results.slice(0, topK);
-    }
-    async delete(ids) {
-        for (const id of ids) {
-            this.vectors.delete(id);
-        }
-    }
-    async fetch(ids) {
-        const results = [];
-        for (const id of ids) {
-            const vector = this.vectors.get(id);
-            if (vector) {
-                results.push(vector);
-            }
-        }
-        return results;
-    }
-    async deleteNamespace() {
-        this.vectors.clear();
-    }
-    async disconnect() {
-        // No cleanup needed for local store
-    }
-    /**
-     * Calculate cosine similarity between two vectors
-     */
-    cosineSimilarity(a, b) {
-        if (a.length !== b.length) {
-            throw new Error('Vectors must have same length');
-        }
-        let dotProduct = 0;
-        let normA = 0;
-        let normB = 0;
-        for (let i = 0; i < a.length; i++) {
-            dotProduct += a[i] * b[i];
-            normA += a[i] * a[i];
-            normB += b[i] * b[i];
-        }
-        return dotProduct / (Math.sqrt(normA) * Math.sqrt(normB));
-    }
-    /**
-     * Check if metadata matches filter
-     */
-    matchesFilter(metadata, filter) {
-        if (!metadata)
-            return false;
-        for (const [key, value] of Object.entries(filter)) {
-            if (metadata[key] !== value) {
-                return false;
-            }
-        }
-        return true;
-    }
-    /**
-     * Get all vectors (for testing/debugging)
-     */
-    getAll() {
-        return Array.from(this.vectors.values());
-    }
-    /**
-     * Get count of vectors
-     */
-    count() {
-        return this.vectors.size;
-    }
-}
-exports.LocalVectorStore = LocalVectorStore;

package/dist/types.d.ts

@@ -1,61 +0,0 @@
-export interface VectorDBConfig {
-    provider: 'pinecone' | 'qdrant' | 'chromadb' | 'local';
-    apiKey?: string;
-    endpoint?: string;
-    dimension?: number;
-    metric?: 'cosine' | 'euclidean' | 'dotproduct';
-    namespace?: string;
-}
-export interface Vector {
-    id: string;
-    values: number[];
-    metadata?: Record<string, any>;
-}
-export interface QueryResult {
-    id: string;
-    score: number;
-    metadata?: Record<string, any>;
-    values?: number[];
-}
-export interface QueryOptions {
-    topK?: number;
-    filter?: Record<string, any>;
-    includeValues?: boolean;
-    includeMetadata?: boolean;
-}
-export interface UpsertOptions {
-    namespace?: string;
-    batch?: boolean;
-}
-export declare abstract class VectorStore {
-    protected config: VectorDBConfig;
-    constructor(config: VectorDBConfig);
-    /**
-     * Initialize the vector store connection
-     */
-    abstract connect(): Promise<void>;
-    /**
-     * Insert or update vectors
-     */
-    abstract upsert(vectors: Vector[], options?: UpsertOptions): Promise<void>;
-    /**
-     * Query for similar vectors
-     */
-    abstract query(vector: number[], options?: QueryOptions): Promise<QueryResult[]>;
-    /**
-     * Delete vectors by ID
-     */
-    abstract delete(ids: string[], namespace?: string): Promise<void>;
-    /**
-     * Get vector by ID
-     */
-    abstract fetch(ids: string[], namespace?: string): Promise<Vector[]>;
-    /**
-     * Delete all vectors in a namespace
-     */
-    abstract deleteNamespace(namespace: string): Promise<void>;
-    /**
-     * Close the connection
-     */
-    abstract disconnect(): Promise<void>;
-}

package/dist/types.js

@@ -1,10 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.VectorStore = void 0;
-class VectorStore {
-    config;
-    constructor(config) {
-        this.config = config;
-    }
-}
-exports.VectorStore = VectorStore;