feat: 添加索引模式开关 - 支持简单索引模式（不使用LLM）

新增配置:
- USE_LLM_INDEX = False (默认关闭LLM增强索引)

简单索引模式:
- 不调用LLM，速度快
- 使用词频统计提取关键词
- 适合快速建立索引

LLM增强模式(USE_LLM_INDEX=True):
- 使用LLM分析文档，提取摘要、关键词、实体
- 索引质量更高，但速度慢、需要LLM服务

app.py

@@ -177,7 +177,7 @@ def api_upload_document():
 
 @app.route('/api/documents/<int:doc_id>', methods=['GET'])
 def api_get_document(doc_id):
-    """获取文档详情"""
+    """获取文档详情API"""
     doc = Document.query.get_or_404(doc_id)
     
     chunks = DocumentChunk.query.filter_by(document_id=doc_id).all()
@@ -188,6 +188,14 @@ def api_get_document(doc_id):
     })
 
 
+@app.route('/documents/<int:doc_id>')
+def document_detail_page(doc_id):
+    """文档详情页面"""
+    doc = Document.query.get_or_404(doc_id)
+    chunks = DocumentChunk.query.filter_by(document_id=doc_id).order_by(DocumentChunk.chunk_index).all()
+    return render_template('document_detail.html', doc=doc, chunks=chunks)
+
+
 @app.route('/api/documents/<int:doc_id>', methods=['DELETE'])
 def api_delete_document(doc_id):
     """删除文档"""

config.py

@@ -33,6 +33,12 @@ DOC_CONFIG = {
     "max_summary_length": 500,  # 摘要最大长度
 }
 
+# ==================== 索引模式 ====================
+# use_llm_index: 是否使用LLM增强索引
+# True - 使用LLM分析文档，提取关键词、摘要等（需要LLM服务，速度慢）
+# False - 使用简单分词和词频统计（速度快，无需LLM）
+USE_LLM_INDEX = False  # 默认关闭LLM索引
+
 # ==================== 索引配置 ====================
 INDEX_CONFIG = {
     # BM25参数

models.py

@@ -92,6 +92,8 @@ class Document(db.Model):
         return {
             'id': self.id,
             'filename': self.filename,
+            'file_type': self.file_type,
+            'file_size': self.file_size,
             'title': self.title,
             'status': self.status,
             'summary': self.summary,

services.py

@@ -12,7 +12,7 @@ from collections import Counter
 from openai import OpenAI
 from flask import current_app
 
-from config import LLM_CONFIG, DOC_CONFIG, INDEX_CONFIG
+from config import LLM_CONFIG, DOC_CONFIG, INDEX_CONFIG, USE_LLM_INDEX
 from models import db, Document, DocumentChunk, InvertedIndex, IndexStats, QueryLog
 
 
@@ -252,15 +252,25 @@ class DocumentIndexer:
             doc.content = content
             doc.word_count = len(content)
             
-            # 使用LLM分析整个文档
-            print(f"  正在分析文档: {doc.filename}")
-            analysis = self.llm.analyze_document(content, doc.title)
+            # 检查是否使用LLM增强索引
+            use_llm = USE_LLM_INDEX
             
-            doc.summary = analysis.get('summary', '')
-            doc.set_keywords(analysis.get('keywords', []))
-            doc.set_topics(analysis.get('topics', []))
-            doc.category = analysis.get('category', '')
-            doc.set_entities(analysis.get('entities', {}))
+            if use_llm:
+                # 使用LLM分析整个文档
+                print(f"  正在使用LLM分析文档: {doc.filename}")
+                analysis = self.llm.analyze_document(content, doc.title)
+                
+                doc.summary = analysis.get('summary', '')
+                doc.set_keywords(analysis.get('keywords', []))
+                doc.set_topics(analysis.get('topics', []))
+                doc.category = analysis.get('category', '')
+                doc.set_entities(analysis.get('entities', {}))
+            else:
+                # 简单模式：不使用LLM
+                print(f"  正在索引文档(简单模式): {doc.filename}")
+                # 从内容中提取简单关键词
+                simple_keywords = self._extract_simple_keywords(content)
+                doc.set_keywords(simple_keywords[:20])
             
             # 分块处理
             chunks = self._split_content(content)
@@ -279,10 +289,14 @@ class DocumentIndexer:
                     end_char=len(chunk_content)
                 )
                 
-                # LLM分析分块
-                chunk_analysis = self.llm.analyze_chunk(chunk_content)
-                chunk.summary = chunk_analysis.get('summary', '')
-                chunk.set_keywords(chunk_analysis.get('keywords', []))
+                if use_llm:
+                    # LLM分析分块
+                    chunk_analysis = self.llm.analyze_chunk(chunk_content)
+                    chunk.summary = chunk_analysis.get('summary', '')
+                    chunk.set_keywords(chunk_analysis.get('keywords', []))
+                else:
+                    # 简单模式：从分块提取关键词
+                    chunk.set_keywords(self._extract_simple_keywords(chunk_content)[:10])
                 
                 # 计算词频
                 term_freq = self._compute_term_freq(chunk_content)
@@ -313,23 +327,70 @@ class DocumentIndexer:
             print(f"  ✗ 索引失败: {e}")
             return False
     
+    def _extract_simple_keywords(self, content):
+        """简单提取关键词（不使用LLM）"""
+        import re
+        from collections import Counter
+        
+        # 提取中文词组（2-4字）
+        chinese_words = re.findall(r'[\u4e00-\u9fff]{2,4}', content)
+        
+        # 过滤常见无意义词
+        stopwords = {'的', '是', '在', '了', '和', '与', '或', '等', '这', '那', 
+                     '有', '为', '对', '也', '被', '把', '给', '向', '从', '到',
+                     '一个', '一些', '这种', '那种', '什么', '怎么', '如何', '为什么'}
+        
+        # 统计词频
+        word_freq = Counter(w for w in chinese_words if w not in stopwords)
+        
+        # 返回高频词
+        return [w for w, _ in word_freq.most_common(30)]
+    
     def _read_document(self, filepath):
         """读取文档内容"""
         ext = os.path.splitext(filepath)[1].lower()
         
-        # 尝试读取文本文件（包括没有扩展名的）
-        try:
-            with open(filepath, 'r', encoding='utf-8', errors='ignore') as f:
-                content = f.read()
-                if content.strip():  # 如果能读取到内容
-                    return content
-        except:
-            pass
+        # 尝试多种编码读取文本文件
+        encodings = ['utf-8', 'gbk', 'gb2312', 'gb18030', 'big5', 'utf-16', 'latin-1']
         
-        # 按扩展名处理特定格式
-        if ext in ['.txt', '.md', '.json', '.html']:
-            with open(filepath, 'r', encoding='utf-8', errors='ignore') as f:
-                return f.read()
+        def try_read_with_encoding(encodings_list):
+            for enc in encodings_list:
+                try:
+                    with open(filepath, 'r', encoding=enc) as f:
+                        content = f.read()
+                        # 检查是否有有效的中文字符
+                        if content.strip() and len(content) > 0:
+                            # 简单验证：检查是否有乱码
+                            # 如果内容看起来合理，返回它
+                            return content, enc
+                except (UnicodeDecodeError, UnicodeError):
+                    continue
+            return None, None
+        
+        # 按扩展名处理
+        if ext in ['.txt', '.md', '.json', '.html', '']:
+            # 先尝试常见编码
+            content, used_enc = try_read_with_encoding(encodings)
+            if content:
+                print(f"  使用编码 {used_enc} 读取文件")
+                return content
+            
+            # 如果都失败，尝试二进制读取后解码
+            try:
+                with open(filepath, 'rb') as f:
+                    raw = f.read()
+                # 尝试chardet检测编码
+                try:
+                    import chardet
+                    detected = chardet.detect(raw)
+                    if detected['encoding']:
+                        return raw.decode(detected['encoding'])
+                except:
+                    pass
+                # 最后尝试
+                return raw.decode('utf-8', errors='replace')
+            except:
+                return ""
         
         elif ext == '.pdf':
             try:

templates/document_detail.html

@@ -0,0 +1,212 @@
+<!DOCTYPE html>
+<html lang="zh-CN">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>{{ doc.title or doc.filename }} - 文档详情</title>
+    <link href="https://cdn.jsdelivr.net/npm/bootstrap@5.3.0/dist/css/bootstrap.min.css" rel="stylesheet">
+    <link href="https://cdn.jsdelivr.net/npm/bootstrap-icons@1.10.0/font/bootstrap-icons.css" rel="stylesheet">
+    <style>
+        body { background-color: #f8f9fa; }
+        .chunk-card { border-left: 4px solid #667eea; }
+    </style>
+</head>
+<body>
+    <nav class="navbar navbar-expand-lg navbar-dark bg-dark">
+        <div class="container">
+            <a class="navbar-brand" href="/"><i class="bi bi-search"></i> LLM Index RAG</a>
+            <div class="navbar-nav ms-auto">
+                <a class="nav-link" href="/">首页</a>
+                <a class="nav-link" href="/documents">文档管理</a>
+                <a class="nav-link" href="/search">知识检索</a>
+                <a class="nav-link" href="/settings">系统设置</a>
+            </div>
+        </div>
+    </nav>
+
+    <div class="container py-4">
+        <nav aria-label="breadcrumb" class="mb-3">
+            <ol class="breadcrumb">
+                <li class="breadcrumb-item"><a href="/documents">文档管理</a></li>
+                <li class="breadcrumb-item active">{{ doc.title or doc.filename }}</li>
+            </ol>
+        </nav>
+
+        <!-- 文档信息 -->
+        <div class="card mb-4">
+            <div class="card-header d-flex justify-content-between align-items-center">
+                <h5 class="mb-0"><i class="bi bi-file-earmark-text"></i> 文档信息</h5>
+                <div>
+                    {% if doc.status == 'indexed' %}
+                    <span class="badge bg-success">已索引</span>
+                    {% elif doc.status == 'processing' %}
+                    <span class="badge bg-warning">处理中</span>
+                    {% elif doc.status == 'failed' %}
+                    <span class="badge bg-danger">失败</span>
+                    {% else %}
+                    <span class="badge bg-secondary">待索引</span>
+                    {% endif %}
+                </div>
+            </div>
+            <div class="card-body">
+                <div class="row">
+                    <div class="col-md-6">
+                        <table class="table table-sm">
+                            <tr>
+                                <th width="120">文件名</th>
+                                <td>{{ doc.filename }}</td>
+                            </tr>
+                            <tr>
+                                <th>标题</th>
+                                <td>{{ doc.title or '-' }}</td>
+                            </tr>
+                            <tr>
+                                <th>文件类型</th>
+                                <td>{{ doc.file_type }}</td>
+                            </tr>
+                            <tr>
+                                <th>文件大小</th>
+                                <td>{{ (doc.file_size / 1024)|round(1) }} KB</td>
+                            </tr>
+                        </table>
+                    </div>
+                    <div class="col-md-6">
+                        <table class="table table-sm">
+                            <tr>
+                                <th width="120">分块数</th>
+                                <td>{{ doc.chunk_count }}</td>
+                            </tr>
+                            <tr>
+                                <th>字数</th>
+                                <td>{{ doc.word_count|default(0) }}</td>
+                            </tr>
+                            <tr>
+                                <th>上传时间</th>
+                                <td>{{ doc.created_at.strftime('%Y-%m-%d %H:%M') if doc.created_at else '-' }}</td>
+                            </tr>
+                            <tr>
+                                <th>索引时间</th>
+                                <td>{{ doc.indexed_at.strftime('%Y-%m-%d %H:%M') if doc.indexed_at else '-' }}</td>
+                            </tr>
+                        </table>
+                    </div>
+                </div>
+
+                {% if doc.summary %}
+                <div class="mt-3">
+                    <h6><i class="bi bi-card-text"></i> 文档摘要</h6>
+                    <p class="text-muted">{{ doc.summary }}</p>
+                </div>
+                {% endif %}
+
+                {% if doc.keywords %}
+                <div class="mt-3">
+                    <h6><i class="bi bi-tags"></i> 关键词</h6>
+                    <div>
+                        {% for kw in doc.get_keywords() %}
+                        <span class="badge bg-primary me-1">{{ kw }}</span>
+                        {% endfor %}
+                    </div>
+                </div>
+                {% endif %}
+
+                {% if doc.category %}
+                <div class="mt-3">
+                    <h6><i class="bi bi-folder"></i> 分类</h6>
+                    <span class="badge bg-info">{{ doc.category }}</span>
+                </div>
+                {% endif %}
+            </div>
+        </div>
+
+        <!-- 文档分块 -->
+        <div class="card">
+            <div class="card-header">
+                <h5 class="mb-0"><i class="bi bi-puzzle"></i> 文档分块 ({{ chunks|length }} 个)</h5>
+            </div>
+            <div class="card-body">
+                {% if chunks %}
+                <div class="accordion" id="chunksAccordion">
+                    {% for chunk in chunks %}
+                    <div class="accordion-item">
+                        <h2 class="accordion-header">
+                            <button class="accordion-button collapsed" type="button" data-bs-toggle="collapse" data-bs-target="#chunk{{ loop.index }}">
+                                <strong>分块 {{ loop.index }}</strong>
+                                <span class="ms-3 text-muted small">
+                                    {{ chunk.content[:50] }}...
+                                </span>
+                            </button>
+                        </h2>
+                        <div id="chunk{{ loop.index }}" class="accordion-collapse collapse" data-bs-parent="#chunksAccordion">
+                            <div class="accordion-body">
+                                {% if chunk.summary %}
+                                <div class="alert alert-info small">
+                                    <strong>摘要:</strong> {{ chunk.summary }}
+                                </div>
+                                {% endif %}
+
+                                {% if chunk.keywords %}
+                                <div class="mb-2">
+                                    {% for kw in chunk.get_keywords() %}
+                                    <span class="badge bg-secondary me-1">{{ kw }}</span>
+                                    {% endfor %}
+                                </div>
+                                {% endif %}
+
+                                <pre class="bg-light p-3 rounded" style="white-space: pre-wrap; word-wrap: break-word;">{{ chunk.content }}</pre>
+                            </div>
+                        </div>
+                    </div>
+                    {% endfor %}
+                </div>
+                {% else %}
+                <p class="text-muted text-center py-4">暂无分块数据，请先索引文档</p>
+                {% endif %}
+            </div>
+        </div>
+
+        <div class="mt-4">
+            <a href="/documents" class="btn btn-secondary">
+                <i class="bi bi-arrow-left"></i> 返回列表
+            </a>
+            {% if doc.status == 'pending' %}
+            <button class="btn btn-primary" onclick="indexDoc({{ doc.id }})">
+                <i class="bi bi-arrow-repeat"></i> 索引文档
+            </button>
+            {% endif %}
+            <button class="btn btn-danger" onclick="deleteDoc({{ doc.id }})">
+                <i class="bi bi-trash"></i> 删除文档
+            </button>
+        </div>
+    </div>
+
+    <script src="https://cdn.jsdelivr.net/npm/bootstrap@5.3.0/dist/js/bootstrap.bundle.min.js"></script>
+    <script>
+        function indexDoc(id) {
+            if (!confirm('确定要索引此文档吗？')) return;
+            
+            fetch(`/api/index/${id}`, { method: 'POST' })
+                .then(r => r.json())
+                .then(data => {
+                    alert(data.success ? '索引完成' : '索引失败: ' + (data.error || '未知错误'));
+                    if (data.success) location.reload();
+                });
+        }
+
+        function deleteDoc(id) {
+            if (!confirm('确定要删除此文档吗？此操作不可恢复！')) return;
+            
+            fetch(`/api/documents/${id}`, { method: 'DELETE' })
+                .then(r => r.json())
+                .then(data => {
+                    if (data.success) {
+                        alert('删除成功');
+                        window.location.href = '/documents';
+                    } else {
+                        alert('删除失败');
+                    }
+                });
+        }
+    </script>
+</body>
+</html>