currently works

1 files changed, 23 insertions, 24 deletions

diff --git a/src/client/views/nodes/ChatBox/vectorstore/Vectorstore.ts b/src/client/views/nodes/ChatBox/vectorstore/Vectorstore.ts
index 25aec751f..8e7be6eec 100644
--- a/src/client/views/nodes/ChatBox/vectorstore/Vectorstore.ts
+++ b/src/client/views/nodes/ChatBox/vectorstore/Vectorstore.ts
@@ -4,7 +4,7 @@ import { EmbedResponse } from 'cohere-ai/api';
 import dotenv from 'dotenv';
 import axios from 'axios';
 
-import { Chunk, AI_Document, CHUNK_TYPE } from '../types';
+import { RAGChunk, AI_Document, CHUNK_TYPE } from '../types';
 import { Doc } from '../../../../../fields/Doc';
 import { DocData } from '../../../../../fields/DocSymbols';
 import { CsvCast, PDFCast, StrCast } from '../../../../../fields/Types';
@@ -44,7 +44,7 @@ export class Vectorstore {
         if (!indexList.indexes?.some(index => index.name === this.indexName)) {
             await this.pinecone.createIndex({
                 name: this.indexName,
-                dimension: 768,
+                dimension: 1024,
                 metric: 'cosine',
                 spec: {
                     serverless: {
@@ -98,7 +98,7 @@ export class Vectorstore {
                     image_chunks: [],
                 };
 
-                document_json.chunks.forEach((chunk: Chunk) => {
+                document_json.chunks.forEach((chunk: RAGChunk) => {
                     let chunk_to_add: { chunk_id: string; start_page: number; end_page: number }[] | { chunk_id: string; location: string; page: number }[];
                     switch (chunk.metadata.type) {
                         case CHUNK_TYPE.TEXT:
@@ -125,7 +125,7 @@ export class Vectorstore {
 
     private async indexDocument(document: any) {
         console.log('Uploading vectors to content namespace...');
-        const pineconeRecords: PineconeRecord<RecordMetadata>[] = (document.chunks as Chunk[]).map(
+        const pineconeRecords: PineconeRecord<RecordMetadata>[] = (document.chunks as RAGChunk[]).map(
             chunk =>
                 ({
                     id: chunk.id,
@@ -136,29 +136,28 @@ export class Vectorstore {
         await this.index.upsert(pineconeRecords);
     }
 
-    async retrieve(query: string, topK: number = 10): Promise<Chunk[]> {
+    async retrieve(query: string, topK: number = 10): Promise<RAGChunk[]> {
         console.log(`Retrieving chunks for query: ${query}`);
         try {
-            const url = 'https://api.jina.ai/v1/embeddings';
-            const headers = {
-                'Content-Type': 'application/json',
-                Authorization: `Bearer ${process.env.JINA_API_KEY}`,
-            };
-            const data = {
-                model: 'jina-clip-v1',
-                normalized: true,
-                embedding_type: 'float',
-                input: [{ text: query }],
-            };
-
-            const response = await axios.post(url, data, { headers });
-            const embeddings = response.data?.data?.[0]?.embedding;
-
-            if (!embeddings || !Array.isArray(embeddings)) {
-                throw new Error('Invalid embedding response format from Jina API');
+            const queryEmbeddingResponse: EmbedResponse = await this.cohere.embed({
+                texts: [query],
+                model: 'embed-english-v3.0',
+                inputType: 'search_query',
+            });
+
+            let queryEmbedding: number[];
+
+            if (Array.isArray(queryEmbeddingResponse.embeddings)) {
+                queryEmbedding = queryEmbeddingResponse.embeddings[0];
+            } else if (queryEmbeddingResponse.embeddings && 'embeddings' in queryEmbeddingResponse.embeddings) {
+                queryEmbedding = (queryEmbeddingResponse.embeddings as { embeddings: number[][] }).embeddings[0];
+            } else {
+                throw new Error('Invalid embedding response format');
             }
 
-            const queryEmbedding = embeddings;
+            if (!Array.isArray(queryEmbedding)) {
+                throw new Error('Query embedding is not an array');
+            }
 
             const queryResponse: QueryResponse<RecordMetadata> = await this.index.query({
                 vector: queryEmbedding,
@@ -176,7 +175,7 @@ export class Vectorstore {
                         id: match.id,
                         values: match.values as number[],
                         metadata: match.metadata as { text: string; type: string; original_document: string; file_path: string; doc_id: string; location: string; start_page: number; end_page: number },
-                    }) as Chunk
+                    }) as RAGChunk
             );
         } catch (error) {
             console.error(`Error retrieving chunks: ${error}`);