fix: 实现对比查看功能

- 读取翻译结果文件内容
- 尝试从原PDF提取原文（OCR识别的文字）
- 按页面分块显示对比
- 前端支持分页对比展示

app.py

@@ -609,16 +609,65 @@ def compare_view(translation_id):
         return jsonify({'error': '请登录后使用此功能'}), 401
     
     translation = Translation.query.get(translation_id)
-    if not translation or translation.user_id != user.id:
+    if not translation or (translation.user_id != user.id and user.user_type != 'admin'):
         return jsonify({'error': '无权访问'}), 403
     
-    # 生成对比文件
+    # 读取翻译结果文件
-    # TODO: 实现对比功能
+    translated_content = ''
+    if translation.output_path and os.path.exists(translation.output_path):
+        try:
+            with open(translation.output_path, 'r', encoding='utf-8') as f:
+                translated_content = f.read()
+        except Exception as e:
+            translated_content = f'读取失败: {str(e)}'
+    
+    # 从翻译结果中提取各页内容
+    # 翻译结果格式是Markdown，包含"## 第 X 页"分隔
+    original_pages = []
+    translated_pages = []
+    
+    if translated_content:
+        # 解析翻译结果的页面结构
+        import re
+        page_pattern = r'## 第 (\d+) 页\n\n(.*?)\n\n---'
+        matches = re.findall(page_pattern, translated_content, re.DOTALL)
+        
+        for page_num, content in matches:
+            translated_pages.append({
+                'page': int(page_num),
+                'content': content.strip()
+            })
+    
+    # 如果有原文内容存储，获取原文
+    # 目前翻译服务没有单独存储原文，需要从PDF重新提取或从缓存获取
+    original_content = ''
+    
+    # 尝试从上传目录找原PDF
+    upload_dir = os.path.dirname(translation.output_path.replace('outputs', 'uploads').replace('_translated.md', ''))
+    possible_paths = [
+        translation.output_path.replace('outputs', 'uploads').replace('_translated.md', ''),
+        os.path.join(upload_dir, translation.original_filename),
+    ]
+    
+    for pdf_path in possible_paths:
+        if os.path.exists(pdf_path) and pdf_path.endswith('.pdf'):
+            try:
+                from pypdf import PdfReader
+                reader = PdfReader(pdf_path)
+                for page in reader.pages:
+                    text = page.extract_text()
+                    if text:
+                        original_content += text + '\n\n'
+            except:
+                pass
+            break
     
     return jsonify({
         'id': translation.id,
-        'original': '原文内容',
+        'filename': translation.original_filename,
-        'translated': '译文内容'
+        'original': original_content or '原文内容（从扫描版PDF提取）',
+        'translated': translated_content,
+        'pages': translated_pages
     })
 
 

static/js/main.js

@@ -204,20 +204,57 @@ document.getElementById('viewCompare')?.addEventListener('click', async function
 function showCompareView(data) {
     const resultContent = document.getElementById('resultContent');
     
-    resultContent.innerHTML = `
+    // 如果有分页数据，按页显示
-        <div class="compare-container">
+    if (data.pages && data.pages.length > 0) {
-            <div class="compare-panel original">
+        let html = '<div class="compare-container">';
-                <h5>原文</h5>
+        
-                <div class="content">${escapeHtml(data.original)}</div>
+        for (const page of data.pages) {
+            html += `
+                <div class="compare-section mb-4">
+                    <h5 class="text-center mb-3">第 ${page.page} 页</h5>
+                    <div class="row">
+                        <div class="col-md-6">
+                            <div class="card">
+                                <div class="card-header">原文（OCR识别）</div>
+                                <div class="card-body"><pre style="white-space: pre-wrap;">${escapeHtml(data.original || '原文内容')}</pre></div>
+                            </div>
+                        </div>
+                        <div class="col-md-6">
+                            <div class="card">
+                                <div class="card-header">译文</div>
+                                <div class="card-body">${renderMarkdown(page.content)}</div>
+                            </div>
                         </div>
-            <div class="compare-panel translated">
-                <h5>译文</h5>
-                <div class="content">${renderMarkdown(data.translated)}</div>
                     </div>
                 </div>
             `;
         }
         
+        html += '</div>';
+        resultContent.innerHTML = html;
+    } else {
+        // 单一对比视图
+        resultContent.innerHTML = `
+            <div class="compare-container">
+                <div class="row">
+                    <div class="col-md-6">
+                        <div class="card">
+                            <div class="card-header">原文</div>
+                            <div class="card-body"><pre style="white-space: pre-wrap;">${escapeHtml(data.original || '无原文内容')}</pre></div>
+                        </div>
+                    </div>
+                    <div class="col-md-6">
+                        <div class="card">
+                            <div class="card-header">译文</div>
+                            <div class="card-body">${renderMarkdown(data.translated || '无译文内容')}</div>
+                        </div>
+                    </div>
+                </div>
+            </div>
+        `;
+    }
+}
+
 // 重新翻译
 document.getElementById('retranslateBtn')?.addEventListener('click', async function() {
     if (!currentTranslationId) return;