修复markdown公式显示错误的问题

docutranslate/exporter/md/md2html_exporter.py

@@ -46,7 +46,7 @@ class MD2HTMLExporter(MDExporter):
                     },
                     trust: true,
                     strict: false
-                });
+                })
             });
         </script>"""
 
@@ -84,14 +84,6 @@ class MD2HTMLExporter(MDExporter):
 
         content = document.content.decode()
 
-        # =================================================================
-        # 步骤 2: 预处理 markdown 内容，确保数学公式块周围有正确的空行
-        # 正则表达式 r'(\$\$[\s\S]*?\$\$)' 匹配一个完整的 $$...$$ 块。
-        # [\s\S]*? 匹配包括换行符在内的任何字符，并且是非贪婪模式。
-        # re.sub 将找到的每个匹配项替换为 `\n\n<匹配项>\n\n`，从而强制添加空行。
-        content = re.sub(r'(\$\$[\s\S]*?\$\$)', r'\n\n\1\n\n', content)
-        # =================================================================
-
         html_content = markdown.markdown(
             content,
             extensions=extensions,
@@ -109,12 +101,13 @@ class MD2HTMLExporter(MDExporter):
             mermaid=mermaid,
         )
         return Document.from_bytes(content=render.encode("utf-8"), suffix=".html", stem=document.stem)
+
 if __name__ == '__main__':
     from pathlib import Path
 
     # d = Document.from_path(r"C:\Users\jxgm\Desktop\mcp文件夹\学习笔记\互联网认证授权机制\互联网认证授权机制.md")
-    d = Document.from_path(r"C:\Users\jxgm\Desktop\matrixcalc_translated.md")
+    # d = Document.from_path(r"C:\Users\jxgm\Desktop\matrixcalc_translated.md")
-    # d = Document.from_path(r"C:\Users\jxgm\Downloads\3a8d8999-3e9d-4f32-a32c-5b0830bb4320\full.md")
+    d = Document.from_path(r"C:\Users\jxgm\Desktop\full_translated.md")
     exporter = MD2HTMLExporter()
     d1 = exporter.export(d)
     path = Path(r"C:\Users\jxgm\Desktop\a.html")

docutranslate/translator/ai_translator/md_translator.py

@@ -9,7 +9,6 @@ from docutranslate.ir.markdown_document import MarkdownDocument
 from docutranslate.translator.ai_translator.base import AiTranslatorConfig
 from docutranslate.translator.base import Translator
 from docutranslate.utils.markdown_splitter import split_markdown_text, join_markdown_texts
-from docutranslate.utils.markdown_utils import clean_markdown_math_block
 
 
 @dataclass
@@ -44,7 +43,6 @@ class MDTranslator(Translator):
             # 做一些加强鲁棒性的操作
             content = content.replace(r'\（', r'\(')
             content = content.replace(r'\）', r'\)')
-            content = clean_markdown_math_block(content)
 
             document.content = content.encode()
         self.logger.info("翻译完成")
@@ -62,7 +60,6 @@ class MDTranslator(Translator):
                 # 做一些加强鲁棒性的操作
                 content = content.replace(r'\（', r'\(')
                 content = content.replace(r'\）', r'\)')
-                content = clean_markdown_math_block(content)
                 document.content = content.encode()
 
             await asyncio.to_thread(run)

docutranslate/utils/markdown_splitter.py

@@ -2,6 +2,8 @@ import re
 from typing import List
 
 
+
+
 class MarkdownBlockSplitter:
     def __init__(self, max_block_size: int = 5000):
         """
@@ -202,4 +204,13 @@ def join_markdown_texts(markdown_texts: List[str]) -> str:
 
         joined_text += separator + current_chunk
 
-    return joined_text
+    return joined_text
+
+
+if __name__ == '__main__':
+    from pathlib import Path
+    from docutranslate.utils.markdown_utils import clean_markdown_math_block
+    content=Path(r"C:\Users\jxgm\Desktop\3a8d8999-3e9d-4f32-a32c-5b0830bb4320\full.md").read_text()
+    content=split_markdown_text(content)
+    content=join_markdown_texts(content)
+

docutranslate/utils/markdown_utils.py

@@ -211,16 +211,6 @@ def unembed_base64_images_to_zip(markdown:str,markdown_name:str,image_folder_nam
                     zipf.write(file, file.relative_to(folder_path))
     return zip_buffer.getvalue()
 
-def clean_markdown_math_block(markdown):
-    """清除公式块的多余空格字符"""
-
-    def replace_block(match: re.Match):
-        return f"{match.group(1).strip()}\n{match.group(2).strip()}\n{match.group(3).lstrip()}"
-
-    pattern = re.compile(r"(^\s*\$\$\s*)\n([\s\S]+?)\n(^\s*\$\$\s*$)", re.MULTILINE)
-    cleaned_text = pattern.sub(replace_block, markdown)
-    return cleaned_text
-
 
 if __name__ == '__main__':
     pass