新增 word 文件预览功能

.claude/settings.local.json

@@ -25,7 +25,9 @@
       "Bash(git -C /d/3.Project/HTY1024/ai-app-database commit -m 'fix\\(ci\\): 将 Secret 名从 GITEA_TOKEN 改为 REGISTRY_TOKEN:*)",
       "Bash(git -C /d/3.Project/HTY1024/ai-app-database log --oneline -3)",
       "Bash(git -C /d/3.Project/HTY1024/ai-app-database add scripts/ .gitattributes)",
-      "Bash(git -C /d/3.Project/HTY1024/ai-app-database log --oneline -4)"
+      "Bash(git -C /d/3.Project/HTY1024/ai-app-database log --oneline -4)",
+      "Bash(xargs ls:*)",
+      "Bash(find /d/3.Project/HTY1024/ai-app-database/app -name \"__init__.py\" -exec head -30 {} \\\\;)"
     ]
   }
 }

app/routes/resources.py

@@ -404,7 +404,7 @@ def progress(resource_id):
 @resources_bp.route('/<int:resource_id>/preview')
 @login_required
 def preview(resource_id):
-    """直接返回文件内容（用于 iframe 文本预览）"""
+    """直接返回文件内容（用于文本/Word 文档预览）"""
     res = db.get_or_404(Resource, resource_id)
     if res.user_id != current_user.id and not current_user.is_admin:
         abort(403)
@@ -413,6 +413,37 @@ def preview(resource_id):
     abs_path = os.path.join(current_app.root_path, 'static', res.file_path)
     if not os.path.exists(abs_path):
         abort(404)
+
+    ext = (res.original_name or res.filename or '').rsplit('.', 1)[-1].lower()
+
+    # Word 文档（.docx）使用 mammoth 转换为 HTML
+    if ext == 'docx':
+        try:
+            import mammoth
+            with open(abs_path, 'rb') as f:
+                result = mammoth.convert_to_html(f)
+            return Response(result.value, mimetype='text/html; charset=utf-8')
+        except ImportError:
+            return Response(
+                '<div style="padding:1rem;color:#dc3545">'
+                '服务器未安装 mammoth 库，无法预览 .docx 文件。请运行 '
+                '<code>pip install mammoth</code> 后重试。</div>',
+                mimetype='text/html; charset=utf-8'
+            )
+        except Exception as e:
+            return Response(
+                f'<div style="padding:1rem;color:#dc3545">无法解析 Word 文档：{e}</div>',
+                mimetype='text/html; charset=utf-8'
+            )
+
+    # 旧版 .doc 二进制格式不支持在线预览
+    if ext == 'doc':
+        return Response(
+            '<div style="padding:1rem;color:#6c757d">'
+            '不支持在线预览旧版 .doc 文件，请下载后查看，或将文件转换为 .docx 格式。</div>',
+            mimetype='text/html; charset=utf-8'
+        )
+
     try:
         with open(abs_path, 'r', encoding='utf-8', errors='replace') as f:
             content = f.read()

app/templates/user/detail.html

@@ -155,6 +155,20 @@
           </div>
 
           {% elif resource.resource_type == 'text' %}
+          {%- set _name = (resource.original_name or resource.filename or '') -%}
+          {%- set _ext = _name.rsplit('.', 1)[-1].lower() if '.' in _name else '' -%}
+          {% if _ext in ('doc', 'docx') %}
+          <div class="text-preview-toolbar p-2 border-bottom d-flex gap-2 align-items-center">
+            <i class="bi bi-file-earmark-word text-primary"></i>
+            <span class="small text-muted">Word 文档预览</span>
+          </div>
+          <div class="word-preview p-4 bg-white" id="textContent"
+               style="max-height:75vh;overflow:auto;line-height:1.7">
+            <div class="text-center py-4 text-muted">
+              <div class="spinner-border spinner-border-sm me-2"></div>加载中…
+            </div>
+          </div>
+          {% else %}
           <div class="text-preview-toolbar p-2 border-bottom d-flex gap-2">
             <button class="btn btn-sm btn-outline-secondary" onclick="changeFont(1)">A+</button>
             <button class="btn btn-sm btn-outline-secondary" onclick="changeFont(-1)">A-</button>
@@ -165,6 +179,7 @@
               <div class="spinner-border spinner-border-sm me-2"></div>加载中…
             </div>
           </div>
+          {% endif %}
 
           {% else %}
           <div class="d-flex flex-column align-items-center justify-content-center p-5 text-muted">
@@ -195,6 +210,30 @@
 <script>
 // ── 文本内容加载 ────────────────────────────────────────────────────────────
 {% if resource.resource_type == 'text' and resource.file_path and resource.download_status in ('na','done') %}
+{%- set _name = (resource.original_name or resource.filename or '') -%}
+{%- set _ext = _name.rsplit('.', 1)[-1].lower() if '.' in _name else '' -%}
+{% if _ext in ('doc', 'docx') %}
+// Word 文档：服务端已转换为 HTML，直接渲染
+fetch("{{ url_for('resources.preview', resource_id=resource.id) }}")
+  .then(r => r.text())
+  .then(html => {
+    const container = document.getElementById('textContent');
+    container.innerHTML = html;
+    // 让文档内的图片自适应宽度
+    container.querySelectorAll('img').forEach(img => {
+      img.style.maxWidth = '100%';
+      img.style.height = 'auto';
+    });
+    // 表格基础样式
+    container.querySelectorAll('table').forEach(t => {
+      t.classList.add('table', 'table-bordered');
+    });
+  })
+  .catch(() => {
+    document.getElementById('textContent').innerHTML =
+      '<div class="text-danger p-3">文档加载失败</div>';
+  });
+{% else %}
 fetch("{{ url_for('resources.preview', resource_id=resource.id) }}")
   .then(r => r.text())
   .then(text => {
@@ -222,6 +261,7 @@ function toggleWrap() {
   pre.style.whiteSpace = pre.style.whiteSpace === 'pre' ? 'pre-wrap' : 'pre';
 }
 {% endif %}
+{% endif %}
 
 // ── 下载进度轮询 ────────────────────────────────────────────────────────────
 {% if resource.download_status in ('pending','downloading') %}

app/utils/file_handler.py

@@ -15,6 +15,8 @@ MIME_TYPE_MAP = {
     'text/xml': 'text',
     'application/json': 'text',
     'application/xml': 'text',
+    'application/msword': 'text',
+    'application/vnd.openxmlformats-officedocument.wordprocessingml.document': 'text',
     # 图片
     'image/jpeg': 'image',
     'image/png': 'image',
@@ -45,6 +47,7 @@ MIME_TYPE_MAP = {
 EXT_TYPE_MAP = {
     'txt': 'text', 'md': 'text', 'csv': 'text', 'json': 'text',
     'xml': 'text', 'log': 'text', 'html': 'text', 'htm': 'text',
+    'doc': 'text', 'docx': 'text',
     'jpg': 'image', 'jpeg': 'image', 'png': 'image', 'gif': 'image',
     'webp': 'image', 'bmp': 'image', 'svg': 'image', 'ico': 'image',
     'mp3': 'audio', 'wav': 'audio', 'ogg': 'audio', 'flac': 'audio',

config.py

@@ -29,7 +29,7 @@ class Config:
     WTF_CSRF_TIME_LIMIT = 3600
 
     # 允许的文件类型
-    ALLOWED_TEXT_EXT = {'txt', 'md', 'csv', 'json', 'xml', 'log', 'html', 'htm'}
+    ALLOWED_TEXT_EXT = {'txt', 'md', 'csv', 'json', 'xml', 'log', 'html', 'htm', 'doc', 'docx'}
     ALLOWED_IMAGE_EXT = {'jpg', 'jpeg', 'png', 'gif', 'webp', 'bmp', 'svg', 'ico'}
     ALLOWED_AUDIO_EXT = {'mp3', 'wav', 'ogg', 'flac', 'm4a', 'aac', 'wma'}
     ALLOWED_VIDEO_EXT = {'mp4', 'webm', 'avi', 'mkv', 'mov', 'wmv', 'flv', 'm4v'}

requirements.txt

@@ -13,5 +13,6 @@ requests==2.32.3
 python-dotenv==1.0.1
 python-magic-bin==0.4.14; sys_platform == 'win32'
 python-magic==0.4.27; sys_platform != 'win32'
+mammoth==1.8.0
 # WSGI 服务器（容器生产环境）
 gunicorn==23.0.0