github/node-DeepResearch
1
0
Fork 0
mirror of https://github.com/jina-ai/node-DeepResearch.git synced 2025-12-26 06:28:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

update embedding dimensions

Browse Source
This commit is contained in:
Sha Zhou 2025-04-30 16:56:02 +08:00
parent 03a2657fb6
commit dc9b530fb5
2 changed files with 7 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
src/tools/embeddings.ts
View File

@ -93,10 +93,10 @@ async function getBatchEmbeddingsWithRetry(
task: options.task || "text-matching",
input: textsToProcess,
truncate: true,
dimensions: options.dimensions || 128,
};
// Add optional parameters if provided
if (options.dimensions) request.dimensions = options.dimensions;
if (options.late_chunking) request.late_chunking = options.late_chunking;
if (options.embedding_type) request.embedding_type = options.embedding_type;
@ -116,7 +116,7 @@ async function getBatchEmbeddingsWithRetry(
console.error('No data returned from Jina API');
if (retryCount === MAX_RETRIES - 1) {
// On last retry, create placeholder embeddings
const dimensionSize = options.dimensions || 128;
const dimensionSize = options.dimensions || 1024;
const placeholderEmbeddings = textsToProcess.map(text => {
console.error(`Failed to get embedding after all retries: [${text.substring(0, 50)}...]`);
return new Array(dimensionSize).fill(0);
@ -136,7 +136,7 @@ async function getBatchEmbeddingsWithRetry(
}
const receivedIndices = new Set(response.data.data.map(item => item.index));
const dimensionSize = response.data.data[0]?.embedding?.length || options.dimensions || 128;
const dimensionSize = response.data.data[0]?.embedding?.length || options.dimensions || 1024;
// Process successful embeddings
const successfulEmbeddings: number[][] = [];
@ -187,7 +187,7 @@ async function getBatchEmbeddingsWithRetry(
// On last retry, create placeholder embeddings
if (retryCount === MAX_RETRIES - 1) {
const dimensionSize = options.dimensions || 128;
const dimensionSize = options.dimensions || 1024;
for (let idx = 0; idx < textsToProcess.length; idx++) {
const originalIndex = indexMap.get(idx)!;
console.error(`Failed to get embedding after all retries for index ${originalIndex}: [${textsToProcess[idx].substring(0, 50)}...]`);
@ -213,7 +213,7 @@ async function getBatchEmbeddingsWithRetry(
// Handle any remaining missing embeddings after max retries
if (textsToProcess.length > 0) {
console.error(`[embeddings] Failed to get embeddings for ${textsToProcess.length} texts after ${MAX_RETRIES} retries`);
const dimensionSize = options.dimensions || 128;
const dimensionSize = options.dimensions || 1024;
for (let idx = 0; idx < textsToProcess.length; idx++) {
const originalIndex = indexMap.get(idx)!;

4
src/tools/jina-latechunk.ts
View File

@ -38,7 +38,7 @@ export async function cherryPick(question: string, longContext: string, options:
trackers.tokenTracker,
{
task: "retrieval.passage",
dimensions: 128,
dimensions: 1024,
late_chunking: true,
embedding_type: "float"
}
@ -52,7 +52,7 @@ export async function cherryPick(question: string, longContext: string, options:
trackers.tokenTracker,
{
task: "retrieval.query",
dimensions: 128,
dimensions: 1024,
embedding_type: "float"
}
);