dgx-spark-playbooks/nvidia/txt2kg/assets/frontend/lib/pinecone.ts
2025-10-06 17:05:41 +00:00

672 lines
20 KiB
TypeScript

/**
* Pinecone service for vector embeddings
* Uses direct API calls for Pinecone local server
*/
import { Document } from "@langchain/core/documents";
// Define types for Pinecone requests and responses
interface PineconeRecord {
id: string;
values: number[];
metadata?: Record<string, any>;
}
interface PineconeQueryResponse {
matches: Array<{
id: string;
score: number;
metadata?: Record<string, any>;
}>;
}
// Define interface for document search results
export interface DocumentSearchResult {
id: string;
score: number;
metadata?: Record<string, any>;
}
export class PineconeService {
private dimension: number = 384; // Dimension for MiniLM-L6-v2
private static instance: PineconeService;
private initialized: boolean = false;
private indexName: string = 'entity-embeddings';
private namespace: string = ''; // Empty string is the default namespace for pinecone-index
private hostUrl: string;
private apiKey: string;
private isInitializing = false;
private constructor() {
// Get environment variables with defaults
const host = process.env.PINECONE_HOST || 'localhost';
const port = process.env.PINECONE_PORT || '5081'; // Default to 5081 for pinecone-index
const apiKey = process.env.PINECONE_API_KEY || 'pclocal';
this.hostUrl = `http://${host}:${port}`;
this.apiKey = apiKey;
console.log(`Initializing Pinecone service with host: ${this.hostUrl}`);
}
/**
* Get singleton instance
*/
public static getInstance(): PineconeService {
if (!PineconeService.instance) {
PineconeService.instance = new PineconeService();
}
return PineconeService.instance;
}
/**
* Check if the service is initialized
*/
public isInitialized(): boolean {
return this.initialized;
}
/**
* Make a request to the Pinecone API
*/
private async makeRequest(endpoint: string, method: string = 'GET', body?: any): Promise<any> {
try {
// For the Docker container setup, Pinecone is a separate service
const url = endpoint.startsWith('http') ? endpoint : `${this.hostUrl}${endpoint}`;
console.log(`Making Pinecone request to: ${url}`);
const options: RequestInit = {
method,
headers: {
'Content-Type': 'application/json',
'Api-Key': this.apiKey,
}
};
if (body) {
options.body = JSON.stringify(body);
}
const response = await fetch(url, options);
if (!response.ok) {
const errorText = await response.text();
console.log(`Pinecone API error (${response.status}) for ${url}: ${errorText}`);
return null;
}
// For HEAD requests or empty responses
if (method === 'HEAD' || response.headers.get('content-length') === '0') {
return { status: response.status };
}
return await response.json();
} catch (error) {
console.log(`Error in Pinecone API request to ${endpoint} - request failed`);
return null;
}
}
/**
* Check if the Pinecone server is up and running
*/
private isPineconeRunningCheck = false;
public async isPineconeRunning(): Promise<boolean> {
// Prevent concurrent checks that could cause loops
if (this.isPineconeRunningCheck) {
console.log('Already checking if Pinecone is running, returning true to break cycle');
return true;
}
this.isPineconeRunningCheck = true;
try {
// In Docker Compose setup, Pinecone might be available but need time to initialize
// Try with simplified health checks first
try {
// This is a simplified test to see if the server responds at all
const response = await fetch(this.hostUrl, {
method: 'GET',
headers: {
'Api-Key': this.apiKey
}
});
// Even a 404 is fine - it means the server is responding
if (response.status >= 200 && response.status < 500) {
console.log(`Pinecone server is up (basic connectivity check)`);
this.isPineconeRunningCheck = false;
return true;
}
} catch (e) {
console.log(`Basic connectivity check failed:`, e);
}
// Try multiple health check endpoints
const healthEndpoints = [
'/health',
'/ready',
'/',
'/version',
'/list_indexes',
'/collections',
'/status'
];
for (const endpoint of healthEndpoints) {
try {
// Try with direct fetch without going through makeRequest
const response = await fetch(`${this.hostUrl}${endpoint}`, {
method: 'GET',
headers: {
'Api-Key': this.apiKey
}
});
// Even a 404 might be fine - at least the server is responding
if (response.status >= 200 && response.status < 500) {
console.log(`Pinecone server is up (checked with ${endpoint}, status: ${response.status})`);
this.isPineconeRunningCheck = false;
return true;
}
} catch (e) {
console.log(`Health check failed for ${endpoint}:`, e);
}
}
console.log('All Pinecone health checks failed - server might not be running');
this.isPineconeRunningCheck = false;
return false;
} catch (error) {
console.log('Error checking Pinecone server health - server appears to be down');
this.isPineconeRunningCheck = false;
return false;
}
}
/**
* Initialize Pinecone and create index if needed
*/
public async initialize(forceCreateIndex: boolean = false): Promise<void> {
if ((this.initialized && !forceCreateIndex) || this.isInitializing) {
return;
}
this.isInitializing = true;
try {
console.log('Pinecone service initializing...');
// Check if Pinecone server is running
const isRunning = await this.isPineconeRunning();
if (!isRunning) {
console.log('Pinecone server does not appear to be running. Please ensure it is started in Docker.');
this.isInitializing = false;
return; // Don't throw, just return
}
// Check if we can access the index by getting stats
const stats = await this.getStats();
if (stats.httpHealthy) {
console.log(`Connected to Pinecone index with ${stats.totalVectorCount} vectors`);
this.initialized = true;
} else {
console.log('Failed to access Pinecone index - continuing without initialization');
}
this.isInitializing = false;
console.log('Pinecone service initialization completed');
} catch (error) {
console.log('Error during Pinecone service initialization - continuing without connection');
this.isInitializing = false;
// Don't throw error, just log and continue
}
}
/**
* Store embeddings for entities
*/
public async storeEmbeddings(
entityEmbeddings: Map<string, number[]>,
textContentMap?: Map<string, string>
): Promise<void> {
if (!this.initialized) {
await this.initialize();
}
// If still not initialized after attempt, skip storage
if (!this.initialized) {
console.log('Pinecone not available - skipping embedding storage');
return;
}
try {
const records: PineconeRecord[] = [];
// Convert to Pinecone vector format
for (const [entityName, embedding] of entityEmbeddings.entries()) {
const record: PineconeRecord = {
id: entityName,
values: embedding,
metadata: {
text: textContentMap?.get(entityName) || entityName,
type: 'entity'
}
};
records.push(record);
}
// Use batching for efficient upserts
const batchSize = 100;
for (let i = 0; i < records.length; i += batchSize) {
const batch = records.slice(i, i + batchSize);
const success = await this.upsertVectors(batch);
if (success) {
console.log(`Upserted batch ${Math.floor(i/batchSize) + 1} of ${Math.ceil(records.length/batchSize)}`);
} else {
console.log(`Failed to upsert batch ${Math.floor(i/batchSize) + 1} - continuing`);
}
}
console.log(`Completed embedding storage attempt for ${records.length} embeddings`);
} catch (error) {
console.log('Error storing embeddings - continuing without storage');
}
}
/**
* Upsert vectors to Pinecone
*/
public async upsertVectors(vectors: PineconeRecord[]): Promise<boolean> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - skipping vector upsert');
return false;
}
try {
console.log(`Upserting ${vectors.length} vectors to Pinecone`);
const response = await fetch(`${this.hostUrl}/vectors/upsert`, {
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Api-Key': this.apiKey
},
body: JSON.stringify({
vectors: vectors,
namespace: this.namespace
})
});
if (!response.ok) {
const errorText = await response.text();
console.log(`Pinecone upsert failed: ${response.status} - ${errorText}`);
return false;
}
console.log(`Successfully upserted ${vectors.length} vectors`);
return true;
} catch (error) {
console.log('Error upserting vectors to Pinecone - continuing without storage');
return false;
}
}
/**
* Store embeddings with metadata
*/
public async storeEmbeddingsWithMetadata(
embeddings: Map<string, number[]>,
textContent: Map<string, string>,
metadata: Map<string, any>
): Promise<void> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - skipping embedding storage with metadata');
return;
}
try {
const records: PineconeRecord[] = [];
// Convert to Pinecone vector format
for (const [key, embedding] of embeddings.entries()) {
const record: PineconeRecord = {
id: key,
values: embedding,
metadata: {
text: textContent.get(key) || '',
...metadata.get(key) || {}
}
};
records.push(record);
}
// Use batching for efficient upserts
const batchSize = 100;
for (let i = 0; i < records.length; i += batchSize) {
const batch = records.slice(i, i + batchSize);
const success = await this.upsertVectors(batch);
if (success) {
console.log(`Upserted batch ${Math.floor(i/batchSize) + 1} of ${Math.ceil(records.length/batchSize)}`);
} else {
console.log(`Failed to upsert batch ${Math.floor(i/batchSize) + 1} - continuing`);
}
}
console.log(`Completed embedding storage attempt for ${records.length} embeddings with metadata`);
} catch (error) {
console.log('Error storing embeddings with metadata - continuing without storage');
}
}
/**
* Find similar entities to a query embedding
*/
public async findSimilarEntitiesWithMetadata(
embedding: number[],
limit: number = 10
): Promise<{ entities: string[], metadata: Map<string, any> }> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - returning empty results');
return { entities: [], metadata: new Map() };
}
try {
const queryResponse = await this.queryVectors(embedding, limit, true);
if (!queryResponse) {
return { entities: [], metadata: new Map() };
}
// Extract entities and metadata
const entities = queryResponse.matches.map(match => match.id);
const metadataMap = new Map<string, any>();
queryResponse.matches.forEach(match => {
metadataMap.set(match.id, {
...match.metadata,
score: match.score
});
});
return { entities, metadata: metadataMap };
} catch (error) {
console.log('Error finding similar entities - returning empty results');
return { entities: [], metadata: new Map() };
}
}
/**
* Query vectors in Pinecone
*/
private async queryVectors(
vector: number[],
topK: number = 10,
includeMetadata: boolean = false
): Promise<PineconeQueryResponse | null> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - cannot query vectors');
return null;
}
try {
const response = await fetch(`${this.hostUrl}/query`, {
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Api-Key': this.apiKey
},
body: JSON.stringify({
vector: vector,
topK: topK,
includeMetadata: includeMetadata,
namespace: this.namespace
})
});
if (!response.ok) {
const errorText = await response.text();
console.log(`Pinecone query failed: ${response.status} - ${errorText}`);
return null;
}
return await response.json();
} catch (error) {
console.log('Error querying vectors from Pinecone - returning null');
return null;
}
}
/**
* Find similar entities to a query embedding
*/
public async findSimilarEntities(queryEmbedding: number[], topK: number = 10): Promise<string[]> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - returning empty entity list');
return [];
}
try {
const queryResponse = await this.queryVectors(queryEmbedding, topK, false);
if (!queryResponse) {
return [];
}
return queryResponse.matches.map(match => match.id);
} catch (error) {
console.log('Error finding similar entities - returning empty list');
return [];
}
}
/**
* Get all entities in the index (up to limit)
*/
public async getAllEntities(limit: number = 1000): Promise<string[]> {
if (!this.initialized) {
await this.initialize();
}
try {
// Create a dummy query that will match all vectors
const dummyVector = Array(this.dimension).fill(0);
const queryResponse = await this.queryVectors(dummyVector, limit, false);
return queryResponse.matches.map(match => match.id);
} catch (error) {
console.error('Error getting all entities:', error);
return [];
}
}
/**
* Delete entities from the index
*/
public async deleteEntities(entityIds: string[]): Promise<boolean> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - cannot delete entities');
return false;
}
try {
console.log(`Deleting ${entityIds.length} entities from Pinecone`);
const response = await fetch(`${this.hostUrl}/vectors/delete`, {
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Api-Key': this.apiKey
},
body: JSON.stringify({
ids: entityIds,
namespace: this.namespace
})
});
if (!response.ok) {
const errorText = await response.text();
console.log(`Pinecone delete failed: ${response.status} - ${errorText}`);
return false;
}
console.log(`Successfully deleted ${entityIds.length} entities`);
return true;
} catch (error) {
console.log('Error deleting entities from Pinecone - operation failed');
return false;
}
}
/**
* Get index statistics from Pinecone
*/
public async getStats(): Promise<any> {
try {
// Try direct HTTP requests to the describe_index_stats endpoint
try {
console.log('Getting stats from Pinecone...');
const response = await fetch(`${this.hostUrl}/describe_index_stats`, {
method: 'GET',
headers: {
'Content-Type': 'application/json',
'Api-Key': this.apiKey
}
});
if (response.ok) {
const statsData = await response.json();
console.log('Successfully retrieved stats from Pinecone');
return {
totalVectorCount: statsData.totalVectorCount || 0,
namespaces: statsData.namespaces || {},
source: 'direct-http',
httpHealthy: true
};
} else {
console.log(`Pinecone stats request failed with status: ${response.status}`);
return {
totalVectorCount: 0,
source: 'error',
httpHealthy: false,
error: `Failed to get stats: ${response.status}`
};
}
} catch (error) {
console.log('Pinecone connection failed - server may not be running');
return {
totalVectorCount: 0,
source: 'error',
httpHealthy: false,
error: error instanceof Error ? error.message : String(error)
};
}
} catch (error) {
console.log('Error accessing Pinecone service');
return {
totalVectorCount: 0,
source: 'error',
httpHealthy: false,
error: error instanceof Error ? error.message : String(error)
};
}
}
/**
* Delete all entities in the index
*/
public async deleteAllEntities(): Promise<boolean> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - cannot delete all entities');
return false;
}
try {
console.log('Deleting all entities from Pinecone');
const response = await fetch(`${this.hostUrl}/vectors/delete`, {
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Api-Key': this.apiKey
},
body: JSON.stringify({
deleteAll: true,
namespace: this.namespace
})
});
if (!response.ok) {
const errorText = await response.text();
console.log(`Pinecone delete all failed: ${response.status} - ${errorText}`);
return false;
}
console.log('Successfully deleted all entities from Pinecone');
return true;
} catch (error) {
console.log('Error deleting all entities from Pinecone - operation failed');
return false;
}
}
/**
* Find similar documents to a query embedding
* @param queryEmbedding Query embedding vector
* @param topK Number of results to return
* @returns Promise resolving to array of document search results
*/
public async findSimilarDocuments(queryEmbedding: number[], topK: number = 10): Promise<DocumentSearchResult[]> {
if (!this.initialized) {
await this.initialize();
}
if (!this.initialized) {
console.log('Pinecone not available - returning empty document results');
return [];
}
try {
const queryResponse = await this.queryVectors(queryEmbedding, topK, true);
if (!queryResponse) {
return [];
}
return queryResponse.matches.map(match => ({
id: match.id,
score: match.score,
metadata: match.metadata
}));
} catch (error) {
console.log('Error finding similar documents - returning empty results');
return [];
}
}
}