From db4371c74598894e2cf72e5d102fe92a731d0e2d Mon Sep 17 00:00:00 2001
From: Stephen Hu <stephenhu@seismic.com>
Date: Fri, 23 May 2025 14:30:19 +0800
Subject: [PATCH] Fix: Improve First Chunk Size (#7806)

### What problem does this PR solve?

https://github.com/infiniflow/ragflow/issues/7790

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
---
 rag/nlp/__init__.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/rag/nlp/__init__.py b/rag/nlp/__init__.py
index 71ca2aca1..0fe0653fc 100644
--- a/rag/nlp/__init__.py
+++ b/rag/nlp/__init__.py
@@ -524,7 +524,7 @@ def naive_merge(sections, chunk_token_num=128, delimiter="\n。；！？"):
         if tnum < 8:
             pos = ""
         # Ensure that the length of the merged chunk does not exceed chunk_token_num  
-        if tk_nums[-1] > chunk_token_num:
+        if cks[-1] == "" or tk_nums[-1] > chunk_token_num:
 
             if t.find(pos) < 0:
                 t += pos
@@ -560,7 +560,7 @@ def naive_merge_with_images(texts, images, chunk_token_num=128, delimiter="\n。
         if tnum < 8:
             pos = ""
         # Ensure that the length of the merged chunk does not exceed chunk_token_num
-        if tk_nums[-1] > chunk_token_num:
+        if cks[-1] == "" or tk_nums[-1] > chunk_token_num:
             if t.find(pos) < 0:
                 t += pos
             cks.append(t)
@@ -627,7 +627,7 @@ def naive_merge_docx(sections, chunk_token_num=128, delimiter="\n。；！？"):
         tnum = num_tokens_from_string(t)
         if tnum < 8:
             pos = ""
-        if tk_nums[-1] > chunk_token_num:
+        if cks[-1] == "" or tk_nums[-1] > chunk_token_num:
             if t.find(pos) < 0:
                 t += pos
             cks.append(t)