fix: code block segmentation problem of markdown document (#6465)

2025-12-09 10:56:52 +08:00 · 2024-07-25 17:24:37 +08:00
parent 16b4f560cd
commit 5e4ac11df3
1 changed files with 8 additions and 0 deletions
--- a/api/core/rag/extractor/markdown_extractor.py
+++ b/api/core/rag/extractor/markdown_extractor.py
@@ -54,8 +54,16 @@ class MarkdownExtractor(BaseExtractor):

        current_header = None
        current_text = ""
+        code_block_flag = False

        for line in lines:
+            if line.startswith("```"):
+                code_block_flag = not code_block_flag
+                current_text += line + "\n"
+                continue
+            if code_block_flag:
+                current_text += line + "\n"
+                continue
            header_match = re.match(r"^#+\s", line)
            if header_match:
                if current_header is not None: