'commit'

dsLightRag/Util/DocxUtil.py

@@ -29,9 +29,11 @@ def get_docx_content_by_pandoc(docx_file):
     # 最后拼接的内容
     content = ""
     # output_file 设置为临时目录下的uuid.md
-    temp_markdown = os.path.join(os.environ['TEMP'], uuid.uuid4().hex + '.md')
+    file_name=uuid.uuid4().hex
+    temp_markdown = os.path.join(os.environ['TEMP'], file_name + '.md')
     # 调用pandoc将docx文件转换成markdown
-    subprocess.run(['pandoc', docx_file, '-f', 'docx', '-t', 'markdown', '-o', temp_markdown])
+    os.mkdir("../static/Images/"+file_name)
+    subprocess.run(['pandoc', docx_file, '-f', 'docx', '-t', 'markdown', '-o', temp_markdown,'--extract-media=../static/Images/'+file_name])
     # 打印 temp_markdown 文件
     # with open(temp_markdown, 'r', encoding='utf-8') as f:
     #     print(f.read())