From e12f34ed5d114175a4a4b251631e568bd3dca9d5 Mon Sep 17 00:00:00 2001
From: Shuming Liu <smshine@qq.com>
Date: Fri, 3 Apr 2026 10:43:11 +0800
Subject: [PATCH] =?UTF-8?q?=E7=BB=A7=E7=BB=AD=E4=BC=98=E5=8C=96?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 server/file_association.py | 68 ++++++++++++++++++++++++++++++++------
 1 file changed, 58 insertions(+), 10 deletions(-)

diff --git a/server/file_association.py b/server/file_association.py
index 89371b8..b838700 100644
--- a/server/file_association.py
+++ b/server/file_association.py
@@ -112,17 +112,47 @@ def read_word_content(doc_path: Path) -> List[str]:
 
 def read_docx_content(doc_path: Path) -> List[str]:
     """读取 .docx 格式 Word 文档内容"""
-    lines = []
     try:
         with zipfile.ZipFile(doc_path, 'r') as z:
             content = z.read('word/document.xml').decode('utf-8')
+            # 提取所有 <w:t>...</w:t> 中的内容，保留原始空白字符
             texts = re.findall(r'<w:t[^>]*>(.*?)</w:t>', content, re.DOTALL)
+
+            # 去除 HTML 标签，不替换内部的空白字符
+            processed = []
             for text in texts:
                 text = re.sub(r'<[^>]+>', '', text).strip()
-                text = re.sub(r'\s+', ' ', text)
                 if text:
-                    lines.append(text)
-            return lines
+                    processed.append(text)
+
+            # 拼接所有文本，保留 \t 和 \n
+            full_text = '\n'.join(processed)
+
+            # 智能检测：判断是使用 \t 分割 6 段，还是每行一个字段
+            lines = full_text.split('\n')
+            lines = [line.strip() for line in lines if line.strip()]
+
+            # 检测是否每行都是单字段（没有\t）
+            # 如果前几行都没有\t，可能是每行一个字段
+            has_tab_in_first_rows = any('\t' in line for line in lines[:6])
+
+            if not has_tab_in_first_rows and len(lines) >= 6:
+                # 可能是每行一个字段，尝试合并为 6 段
+                result = []
+                i = 0
+                while i < len(lines):
+                    if i + 5 <= len(lines):
+                        # 尝试合并接下来的 6 行作为一段
+                        combined = '\t'.join(lines[i:i+6])
+                        result.append(combined)
+                        i += 6
+                    else:
+                        result.append(lines[i])
+                        i += 1
+                return result
+            else:
+                # 每行已经有\t分割，直接返回
+                return lines
     except Exception as e:
         print(f"读取 .docx 失败 ({doc_path.name}): {e}")
         return []
@@ -422,15 +452,33 @@ def main():
                 clear_directory(extract_dir)
 
             try:
-                # 解压
+                # 先提取所有文件
+                with zipfile.ZipFile(zip_file, 'r') as zip_ref:
+                    zip_ref.extractall(data_dir)
+
+                # 然后遍历所有成员，设置时间戳
                 with zipfile.ZipFile(zip_file, 'r') as zip_ref:
                     for member in zip_ref.infolist():
-                        zip_ref.extract(member, data_dir)
-                        member_path = data_dir / zip_file.stem / member.filename
+                        # 构建正确的文件路径 - 先获取解压后的实际路径
+                        base_extract_path = data_dir / zip_file.stem
+                        if member.filename.startswith(zip_file.stem + '/'):
+                            # 文件路径包含 zip 文件名前缀
+                            member_path = base_extract_path / member.filename[len(zip_file.stem) + 1:]
+                        elif member.filename.startswith(zip_file.stem):
+                            # 文件路径直接以 zip 文件名开头
+                            member_path = base_extract_path / member.filename[len(zip_file.stem):].lstrip('/')
+                        else:
+                            # 文件路径不包含前缀
+                            member_path = base_extract_path / member.filename
+
+                        # 如果路径存在，设置时间戳
                         if member_path.exists():
-                            date_time = datetime(*member.date_time[:5])
-                            timestamp = date_time.timestamp()
-                            os.utime(member_path, (timestamp, timestamp))
+                            try:
+                                date_time = datetime(*member.date_time[:5])
+                                timestamp = date_time.timestamp()
+                                os.utime(member_path, (timestamp, timestamp))
+                            except Exception:
+                                pass
 
                 # 检查是否多了一层目录
                 inner_dir = extract_dir / zip_file.stem