From 5f780435ac9a3d3c3a2ff7ed57122846c60924a8 Mon Sep 17 00:00:00 2001
From: Han Xiao <han.xiao@jina.ai>
Date: Wed, 12 Mar 2025 14:21:52 +0800
Subject: [PATCH] feat: late chunking

---
 src/tools/jina-latechunk.ts | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/tools/jina-latechunk.ts b/src/tools/jina-latechunk.ts
index 42e7585..14554bb 100644
--- a/src/tools/jina-latechunk.ts
+++ b/src/tools/jina-latechunk.ts
@@ -6,15 +6,15 @@ import {Schemas} from "../utils/schemas";
 export async function cherryPick(question: string, longContext: string, options: any = {}, trackers: TrackerContext, schemaGen: Schemas) {
 
   const {
-    snippetLength = 2000,
-    numSnippets = 2,
+    snippetLength = 3000,
+    numSnippets = Math.max(2, Math.min(5, Math.floor(longContext.length / snippetLength))),
     chunkSize = 200,
     maxTokensPerRequest = 8192, // Maximum tokens per embedding request
     // Rough estimate of tokens per character (can be adjusted based on your text)
     tokensPerCharacter = 0.5
   } = options;
 
-  if (longContext.length < snippetLength * numSnippets) {
+  if (longContext.length < snippetLength * 2) {
     // If the context is shorter than the snippet length, return the whole context
     return longContext;
   }