Upload 29 files

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+uploads/496449c167344601ba87c24743701b99.pdf filter=lfs diff=lfs merge=lfs -text
+uploads/eb55dde866f64be4a1fdbc81d3d1df0e.pdf filter=lfs diff=lfs merge=lfs -text

modules/knowledge_base/processor.py

@@ -14,9 +14,12 @@ import io
 
 class DocumentLoader:
     """通用文档加载器"""
-    def __init__(self, file_path: str):
+    def __init__(self, file_path: str, original_filename: str = None):
         self.file_path = file_path
-        self.extension = os.path.splitext(file_path)[1].lower()
+        # 使用传入的原始文件名或者从路径提取的文件名
+        self.original_filename = original_filename or os.path.basename(file_path)
+        # 从原始文件名中获取扩展名，确保中文文件名也能正确识别文件类型
+        self.extension = os.path.splitext(self.original_filename)[1].lower()
         self.api_key = os.getenv("API_KEY")
         self.api_base = os.getenv("BASE_URL")
         
@@ -74,15 +77,27 @@ class DocumentLoader:
     
     def load(self):
         try:
+            print(f"正在加载文件: {self.file_path}, 原始文件名: {self.original_filename}, 扩展名: {self.extension}")
+            
             if self.extension == '.md':
-                loader = UnstructuredMarkdownLoader(self.file_path, encoding='utf-8')
-                return loader.load()
+                try:
+                    loader = UnstructuredMarkdownLoader(self.file_path, encoding='utf-8')
+                    return loader.load()
+                except UnicodeDecodeError:
+                    # 如果UTF-8失败，尝试GBK
+                    loader = UnstructuredMarkdownLoader(self.file_path, encoding='gbk')
+                    return loader.load()
             elif self.extension == '.pdf':
                 loader = PyPDFLoader(self.file_path)
                 return loader.load()
             elif self.extension == '.txt':
-                loader = TextLoader(self.file_path, encoding='utf-8')
-                return loader.load()
+                try:
+                    loader = TextLoader(self.file_path, encoding='utf-8')
+                    return loader.load()
+                except UnicodeDecodeError:
+                    # 如果UTF-8失败，尝试GBK
+                    loader = TextLoader(self.file_path, encoding='gbk')
+                    return loader.load()
             elif self.extension in ['.png', '.jpg', '.jpeg', '.gif', '.bmp']:
                 # 处理图片
                 description = self.process_image(self.file_path)
@@ -92,18 +107,29 @@ class DocumentLoader:
                     page_content=description,
                     metadata={
                         'source': self.file_path,
+                        'file_name': self.original_filename,  # 使用原始文件名
                         'img_url': os.path.abspath(self.file_path)  # 存储图片的绝对路径
                     }
                 )
                 return [doc]
             else:
+                print(f"不支持的文件扩展名: {self.extension}")
                 raise ValueError(f"不支持的文件格式: {self.extension}")
                 
         except UnicodeDecodeError:
-            # 如果 utf-8 失败，尝试 gbk
+            # 如果默认编码处理失败，尝试其他编码
+            print(f"文件编码错误，尝试其他编码: {self.file_path}")
             if self.extension in ['.md', '.txt']:
-                loader = TextLoader(self.file_path, encoding='gbk')
-                return loader.load()
+                try:
+                    loader = TextLoader(self.file_path, encoding='gbk')
+                    return loader.load()
+                except Exception as e:
+                    print(f"尝试GBK编码也失败: {str(e)}")
+            raise
+        except Exception as e:
+            print(f"加载文件 {self.file_path} 时出错: {str(e)}")
+            import traceback
+            traceback.print_exc()
             raise
 
 class DocumentProcessor:
@@ -123,12 +149,13 @@ class DocumentProcessor:
             # 如果是文件，使用文件名（不含扩展名）
             return f"rag_{os.path.splitext(os.path.basename(path))[0].lower()}"
         
-    def process(self, path: str, progress_callback: Optional[Callable] = None) -> List[Dict]:
+    def process(self, path: str, progress_callback: Optional[Callable] = None, original_filename: str = None) -> List[Dict]:
         """
         加载并处理文档，支持目录或单个文件
         参数：
             path: 文档路径
             progress_callback: 进度回调函数，用于报告处理进度
+            original_filename: 原始文件名（包括中文）
         返回：处理后的文档列表
         """
         if os.path.isdir(path):
@@ -148,11 +175,12 @@ class DocumentProcessor:
                             processed_files += 1
                             progress_callback(processed_size, f"处理文件 {processed_files}/{total_files}: {file}")
                             
-                        loader = DocumentLoader(file_path)
+                        # 为目录中的每个文件，传递原始文件名
+                        loader = DocumentLoader(file_path, original_filename=file)
                         docs = loader.load()
                         # 添加文件名到metadata
                         for doc in docs:
-                            doc.metadata['file_name'] = os.path.basename(file_path)
+                            doc.metadata['file_name'] = file  # 使用原始文件名
                         documents.extend(docs)
                     except Exception as e:
                         print(f"警告：加载文件 {file_path} 时出错: {str(e)}")
@@ -161,17 +189,18 @@ class DocumentProcessor:
             try:
                 if progress_callback:
                     file_size = os.path.getsize(path)
-                    progress_callback(file_size * 0.3, f"加载文件: {os.path.basename(path)}")
+                    progress_callback(file_size * 0.3, f"加载文件: {original_filename or os.path.basename(path)}")
                     
-                loader = DocumentLoader(path)
+                # 为单个文件，传递原始文件名
+                loader = DocumentLoader(path, original_filename=original_filename)
                 documents = loader.load()
                 
                 # 更新进度
                 if progress_callback:
                     progress_callback(file_size * 0.6, f"处理文件内容...")
                     
-                # 添加文件名到metadata
-                file_name = os.path.basename(path)
+                # 使用原始文件名而不是存储的文件名
+                file_name = original_filename or os.path.basename(path)
                 for doc in documents:
                     doc.metadata['file_name'] = file_name
             except Exception as e:

modules/knowledge_base/routes.py

@@ -60,12 +60,12 @@ def create_knowledge():
         if not name:
             return jsonify({"success": False, "message": "Knowledge base name cannot be empty"}), 400
             
-        # Check if knowledge base already exists
+        # 检查知识库是否已存在
         indices = retriever.get_all_indices()
         if f"rag_{name}" in indices:
             return jsonify({"success": False, "message": f"Knowledge base '{name}' already exists"}), 400
             
-        # Process uploaded file
+        # 处理上传文件
         if 'file' not in request.files:
             return jsonify({"success": False, "message": "No file uploaded"}), 400
             
@@ -73,53 +73,62 @@ def create_knowledge():
         if file.filename == '':
             return jsonify({"success": False, "message": "No file selected"}), 400
             
-        # Save file
-        filename = secure_filename(file.filename)
-        file_path = os.path.join(UPLOAD_FOLDER, filename)
+        # 保存原始文件名
+        original_filename = file.filename
+        # 从原始文件名中提取扩展名，确保中文文件名也能正确识别文件类型
+        file_ext = os.path.splitext(original_filename)[1].lower()
+        # 使用UUID生成唯一文件名
+        unique_filename = f"{uuid.uuid4().hex}{file_ext}"
+        file_path = os.path.join(UPLOAD_FOLDER, unique_filename)
         file.save(file_path)
         
-        # Create task ID
+        # 创建任务ID
         task_id = f"task_{int(time.time())}_{name}"
         
-        # Initialize task status
+        # 初始化任务状态
         processing_tasks[task_id] = {
             "progress": 0,
             "status": "Starting document processing...",
             "index_name": name,
             "file_path": file_path,
+            "original_filename": original_filename,  # 保存原始文件名
             "error": False,
             "docCount": 0
         }
         
-        # Process documents in a separate thread
+        # 处理文档的线程函数
         def process_in_thread():
             try:
-                # Update task status
+                # 更新任务状态
                 processing_tasks[task_id]["progress"] = 10
                 processing_tasks[task_id]["status"] = "Loading document..."
                 
-                # Process document with progress tracking
+                # 处理文档进度回调
                 def update_progress(progress, status):
                     processing_tasks[task_id]["progress"] = min(95, progress)
                     processing_tasks[task_id]["status"] = status
                 
-                # Process the document
-                processed_docs = doc_processor.process(file_path, progress_callback=update_progress)
+                # 处理文档，传递原始文件名
+                processed_docs = doc_processor.process(
+                    file_path, 
+                    progress_callback=update_progress,
+                    original_filename=original_filename  # 传递原始文件名
+                )
                 
-                # Update task status
+                # 更新任务状态
                 processing_tasks[task_id]["progress"] = 95
                 processing_tasks[task_id]["status"] = "Creating vector store..."
                 processing_tasks[task_id]["docCount"] = len(processed_docs)
                 
-                # Store vectors
+                # 存储向量
                 vector_store.store(processed_docs, f"rag_{name}")
                 
-                # Complete task
+                # 完成任务
                 processing_tasks[task_id]["progress"] = 100
                 processing_tasks[task_id]["status"] = "Processing complete"
                 
             except Exception as e:
-                # Record error
+                # 记录错误
                 processing_tasks[task_id]["error"] = True
                 processing_tasks[task_id]["status"] = f"Processing failed: {str(e)}"
                 import traceback
@@ -158,12 +167,12 @@ def get_progress(task_id):
 def add_documents(index_id):
     """Add documents to a knowledge base"""
     try:
-        # Check if knowledge base exists
+        # 检查知识库是否存在
         indices = retriever.get_all_indices()
         if index_id not in indices:
             return jsonify({"success": False, "message": "Knowledge base does not exist"}), 404
             
-        # Process uploaded file
+        # 处理上传文件
         if 'file' not in request.files:
             return jsonify({"success": False, "message": "No file uploaded"}), 400
             
@@ -171,56 +180,63 @@ def add_documents(index_id):
         if file.filename == '':
             return jsonify({"success": False, "message": "No file selected"}), 400
             
-        # Save file
-        filename = secure_filename(file.filename)
-        file_path = os.path.join(UPLOAD_FOLDER, filename)
+        # 保存原始文件名并使用UUID生成唯一文件名
+        original_filename = file.filename
+        file_ext = os.path.splitext(original_filename)[1].lower()
+        unique_filename = f"{uuid.uuid4().hex}{file_ext}"
+        file_path = os.path.join(UPLOAD_FOLDER, unique_filename)
         file.save(file_path)
         
-        # Extract knowledge base name from index ID
+        # 提取知识库名称
         kb_name = index_id[4:] if index_id.startswith('rag_') else index_id
         
-        # Create task ID
-        task_id = f"task_{int(time.time())}_{kb_name}_{filename}"
+        # 创建任务ID
+        task_id = f"task_{int(time.time())}_{kb_name}_{uuid.uuid4().hex[:8]}"
         
-        # Initialize task status
+        # 初始化任务状态
         processing_tasks[task_id] = {
             "progress": 0,
             "status": "Starting document processing...",
             "index_name": kb_name,
             "file_path": file_path,
+            "original_filename": original_filename,  # 保存原始文件名
             "error": False,
             "docCount": 0
         }
         
-        # Process documents in a separate thread
+        # 处理文档的线程函数
         def process_in_thread():
             try:
-                # Update task status
+                # 更新任务状态
                 processing_tasks[task_id]["progress"] = 10
                 processing_tasks[task_id]["status"] = "Loading document..."
                 
-                # Process document with progress tracking
+                # 处理文档进度回调
                 def update_progress(progress, status):
                     processing_tasks[task_id]["progress"] = min(95, progress)
                     processing_tasks[task_id]["status"] = status
                 
-                # Process the document
-                processed_docs = doc_processor.process(file_path, progress_callback=update_progress)
+                # 处理文档，传递原始文件名
+                processed_docs = doc_processor.process(
+                    file_path, 
+                    progress_callback=update_progress,
+                    original_filename=original_filename  # 传递原始文件名
+                )
                 
-                # Update task status
+                # 更新任务状态
                 processing_tasks[task_id]["progress"] = 95
                 processing_tasks[task_id]["status"] = "Creating vector store..."
                 processing_tasks[task_id]["docCount"] = len(processed_docs)
                 
-                # Store vectors
+                # 存储向量
                 vector_store.store(processed_docs, index_id)
                 
-                # Complete task
+                # 完成任务
                 processing_tasks[task_id]["progress"] = 100
                 processing_tasks[task_id]["status"] = "Processing complete"
                 
             except Exception as e:
-                # Record error
+                # 记录错误
                 processing_tasks[task_id]["error"] = True
                 processing_tasks[task_id]["status"] = f"Processing failed: {str(e)}"
                 import traceback

templates/index.html

@@ -4,7 +4,7 @@
 <head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
-   <meta http-equiv="Content-Security-Policy" content="upgrade-insecure-requests">
+
    <title>教育AI助手开发平台</title>
    <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/dist/css/bootstrap.min.css">
    <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/[email protected]/font/bootstrap-icons.css">
@@ -1653,249 +1653,327 @@
     }
 
     // 加载知识库列表
-    async function loadKnowledgeBases() {
-        try {
-            const response = await fetch('/api/knowledge');
-            const result = await response.json();
+async function loadKnowledgeBases() {
+    try {
+        const response = await fetch('/api/knowledge');
+        const result = await response.json();
+        
+        if (result.success) {
+            allKnowledgeBases = result.data || [];
             
-            if (result.success) {
-                allKnowledgeBases = result.data || [];
-                
-                // 更新仪表盘计数
-                document.getElementById('knowledge-count').textContent = allKnowledgeBases.length;
-                
-                // 更新Agent创建页面的知识库列表
-                const knowledgeListElement = document.getElementById('knowledge-list');
-                if (knowledgeListElement) {
-                    if (allKnowledgeBases.length === 0) {
-                        knowledgeListElement.innerHTML = `
-                            <div class="alert alert-info">
-                                <i class="bi bi-info-circle me-2"></i>
-                                暂无知识库，请先创建知识库
+            // 更新仪表盘计数
+            document.getElementById('knowledge-count').textContent = allKnowledgeBases.length;
+            
+            // 更新Agent创建页面的知识库列表
+            const knowledgeListElement = document.getElementById('knowledge-list');
+            if (knowledgeListElement) {
+                if (allKnowledgeBases.length === 0) {
+                    knowledgeListElement.innerHTML = `
+                        <div class="alert alert-info">
+                            <i class="bi bi-info-circle me-2"></i>
+                            暂无知识库，请先创建知识库
+                        </div>
+                        <div class="text-center">
+                            <button class="btn btn-primary" onclick="switchSection('knowledge-base')">
+                                创建知识库
+                            </button>
+                        </div>
+                    `;
+                } else {
+                    knowledgeListElement.innerHTML = '';
+                    
+                    allKnowledgeBases.forEach(knowledge => {
+                        const item = document.createElement('div');
+                        item.className = 'plugin-option';
+                        item.setAttribute('data-id', knowledge.id);
+                        item.setAttribute('onclick', 'toggleKnowledgeSelection(this)');
+                        
+                        item.innerHTML = `
+                            <div class="plugin-icon">
+                                <i class="bi bi-journal-text"></i>
                             </div>
-                            <div class="text-center">
-                                <button class="btn btn-primary" onclick="switchSection('knowledge-base')">
-                                    创建知识库
-                                </button>
+                            <div class="plugin-content">
+                                <div class="plugin-title">${knowledge.name}</div>
+                                <div class="plugin-description">
+                                    <small>${knowledge.fileCount || 0}个文件</small>
+                                </div>
+                            </div>
+                            <div class="checkbox">
+                                <input type="checkbox" class="form-check-input">
                             </div>
                         `;
-                    } else {
-                        knowledgeListElement.innerHTML = '';
                         
-                        allKnowledgeBases.forEach(knowledge => {
-                            const item = document.createElement('div');
-                            item.className = 'plugin-option';
-                            item.setAttribute('data-id', knowledge.id);
-                            item.setAttribute('onclick', 'toggleKnowledgeSelection(this)');
-                            
-                            item.innerHTML = `
-                                <div class="plugin-icon">
-                                    <i class="bi bi-journal-text"></i>
-                                </div>
-                                <div class="plugin-content">
-                                    <div class="plugin-title">${knowledge.name}</div>
-                                    <div class="plugin-description">
-                                        <small>${knowledge.fileCount || 0}个文件</small>
-                                    </div>
-                                </div>
-                                <div class="checkbox">
-                                    <input type="checkbox" class="form-check-input">
-                                </div>
-                            `;
-                            
-                            knowledgeListElement.appendChild(item);
-                        });
-                    }
+                        knowledgeListElement.appendChild(item);
+                    });
                 }
-                
-                // 更新知识库管理页面的列表
-                const existingKnowledgeList = document.getElementById('existing-knowledge-list');
-                if (existingKnowledgeList) {
-                    if (allKnowledgeBases.length === 0) {
-                        existingKnowledgeList.innerHTML = `
-                            <div class="alert alert-info">
-                                <i class="bi bi-info-circle me-2"></i>
-                                暂无知识库，请创建第一个知识库
-                            </div>
-                        `;
-                    } else {
-                        existingKnowledgeList.innerHTML = '';
+            }
+            
+            // 更新知识库管理页面的列表
+            const existingKnowledgeList = document.getElementById('existing-knowledge-list');
+            if (existingKnowledgeList) {
+                if (allKnowledgeBases.length === 0) {
+                    existingKnowledgeList.innerHTML = `
+                        <div class="alert alert-info">
+                            <i class="bi bi-info-circle me-2"></i>
+                            暂无知识库，请创建第一个知识库
+                        </div>
+                    `;
+                } else {
+                    existingKnowledgeList.innerHTML = '';
+                    
+                    allKnowledgeBases.forEach(knowledge => {
+                        const item = document.createElement('div');
+                        item.className = 'knowledge-item';
                         
-                        allKnowledgeBases.forEach(knowledge => {
-                            const item = document.createElement('div');
-                            item.className = 'knowledge-item';
-                            
-                            let fileListHtml = '';
-                            if (knowledge.files && knowledge.files.length > 0) {
-                                fileListHtml = '<div class="knowledge-files">';
-                                knowledge.files.forEach(file => {
-                                    const isImage = file.toLowerCase().endsWith('.jpg') || file.toLowerCase().endsWith('.png') || file.toLowerCase().endsWith('.jpeg');
-                                    const isPdf = file.toLowerCase().endsWith('.pdf');
-                                    const isDoc = file.toLowerCase().endsWith('.doc') || file.toLowerCase().endsWith('.docx');
-                                    
-                                    let iconClass = 'bi-file-text';
-                                    if (isImage) iconClass = 'bi-file-image';
-                                    if (isPdf) iconClass = 'bi-file-pdf';
-                                    if (isDoc) iconClass = 'bi-file-earmark-word';
-                                    
-                                    fileListHtml += `
-                                        <div class="knowledge-file">
-                                            <div class="knowledge-file-icon">
-                                                <i class="bi ${iconClass}"></i>
-                                            </div>
-                                            <div class="knowledge-file-info">
-                                                <div class="knowledge-file-name">${file}</div>
-                                            </div>
-                                            <button class="btn btn-sm btn-icon btn-light" onclick="downloadFile('${knowledge.id}', '${file}')">
-                                                <i class="bi bi-download"></i>
-                                            </button>
+                        let fileListHtml = '';
+                        if (knowledge.files && knowledge.files.length > 0) {
+                            fileListHtml = '<div class="knowledge-files mt-3">';
+                            knowledge.files.forEach(file => {
+                                // 使用文件名来判断文件类型
+                                const isImage = /\.(jpg|jpeg|png|gif|bmp)$/i.test(file);
+                                const isPdf = /\.pdf$/i.test(file);
+                                const isDoc = /\.(doc|docx)$/i.test(file);
+                                const isTxt = /\.txt$/i.test(file);
+                                const isMd = /\.md$/i.test(file);
+                                
+                                let iconClass = 'bi-file-text';
+                                if (isImage) iconClass = 'bi-file-image';
+                                if (isPdf) iconClass = 'bi-file-pdf';
+                                if (isDoc) iconClass = 'bi-file-earmark-word';
+                                if (isTxt) iconClass = 'bi-file-text';
+                                if (isMd) iconClass = 'bi-file-text';
+                                
+                                fileListHtml += `
+                                    <div class="d-flex align-items-center p-2 border-bottom">
+                                        <i class="bi ${iconClass} me-2" style="font-size: 1.2rem; color: var(--secondary-color);"></i>
+                                        <div class="flex-grow-1 text-truncate" title="${file}">
+                                            ${file}
                                         </div>
-                                    `;
-                                });
-                                fileListHtml += '</div>';
-                            }
-                            
-                            item.innerHTML = `
-                                <div class="knowledge-header">
-                                    <h5 class="knowledge-title">
-                                        <i class="bi bi-journal-text"></i> ${knowledge.name}
-                                    </h5>
-                                    <div>
-                                        <button class="btn btn-sm btn-light me-1" onclick="editKnowledgeBase('${knowledge.id}')">
-                                            <i class="bi bi-pencil"></i>
-                                        </button>
-                                        <button class="btn btn-sm btn-light text-danger" onclick="deleteKnowledgeBase('${knowledge.id}')">
+                                        <button class="btn btn-sm btn-icon btn-light" onclick="deleteFile('${knowledge.id}', '${encodeURIComponent(file)}')">
                                             <i class="bi bi-trash"></i>
                                         </button>
                                     </div>
+                                `;
+                            });
+                            fileListHtml += '</div>';
+                        }
+                        
+                        item.innerHTML = `
+                            <div class="knowledge-header">
+                                <h5 class="knowledge-title">
+                                    <i class="bi bi-journal-text"></i> ${knowledge.name}
+                                </h5>
+                                <div>
+                                    <button class="btn btn-sm btn-light me-1" onclick="editKnowledgeBase('${knowledge.id}')">
+                                        <i class="bi bi-pencil"></i>
+                                    </button>
+                                    <button class="btn btn-sm btn-light text-danger" onclick="deleteKnowledgeBase('${knowledge.id}')">
+                                        <i class="bi bi-trash"></i>
+                                    </button>
                                 </div>
-                                ${fileListHtml}
-                                <div class="knowledge-meta">
-                                    <div class="knowledge-meta-item">
-                                        <i class="bi bi-file-text"></i>
-                                        <span>${knowledge.fileCount || 0}个文件</span>
-                                    </div>
-                                    <div class="knowledge-meta-item">
-                                        <i class="bi bi-calendar3"></i>
-                                        <span>创建于 ${new Date(knowledge.created_at * 1000).toLocaleDateString()}</span>
-                                    </div>
+                            </div>
+                            ${fileListHtml}
+                            <div class="knowledge-meta">
+                                <div class="knowledge-meta-item">
+                                    <i class="bi bi-file-text"></i>
+                                    <span>${knowledge.fileCount || 0}个文件</span>
                                 </div>
-                                <div class="mt-3">
-                                    <form class="upload-form" data-id="${knowledge.id}">
-                                        <div class="input-group">
-                                            <input type="file" class="form-control form-control-sm" required>
-                                            <button type="submit" class="btn btn-sm btn-outline-primary">添加文件</button>
-                                        </div>
-                                    </form>
+                                <div class="knowledge-meta-item">
+                                    <i class="bi bi-calendar3"></i>
+                                    <span>创建于 ${new Date(knowledge.created_at * 1000).toLocaleDateString()}</span>
                                 </div>
-                            `;
-                            
-                            existingKnowledgeList.appendChild(item);
-                        });
+                            </div>
+                            <div class="mt-3">
+                                <form class="upload-form" data-id="${knowledge.id}">
+                                    <div class="input-group">
+                                        <input type="file" class="form-control form-control-sm" required>
+                                        <button type="submit" class="btn btn-sm btn-outline-primary">添加文件</button>
+                                    </div>
+                                </form>
+                            </div>
+                        `;
                         
-                        // 添加上传表单事件监听
-                        document.querySelectorAll('.upload-form').forEach(form => {
-                            form.addEventListener('submit', function(event) {
-                                event.preventDefault();
-                                addFileToKnowledgeBase(form);
-                            });
+                        existingKnowledgeList.appendChild(item);
+                    });
+                    
+                    // 添加上传表单事件监听
+                    document.querySelectorAll('.upload-form').forEach(form => {
+                        form.addEventListener('submit', function(event) {
+                            event.preventDefault();
+                            addFileToKnowledgeBase(form);
                         });
-                    }
+                    });
                 }
-            } else {
-                console.error('加载知识库失败:', result.message);
             }
-        } catch (error) {
-            console.error('加载知识库出错:', error);
+        } else {
+            console.error('加载知识库失败:', result.message);
         }
+    } catch (error) {
+        console.error('加载知识库出错:', error);
     }
+}
 
     // 创建知识库
-    async function createKnowledgeBase(event) {
-        event.preventDefault();
-        
-        const nameInput = document.getElementById('knowledge-name');
-        const fileInput = document.getElementById('knowledge-file');
-        const progressBar = document.getElementById('upload-progress');
+async function createKnowledgeBase(event) {
+    event.preventDefault();
+    
+    const nameInput = document.getElementById('knowledge-name');
+    const fileInput = document.getElementById('knowledge-file');
+    const progressBar = document.getElementById('upload-progress');
+    const submitBtn = event.submitter || document.querySelector('#knowledge-form button[type="submit"]');
+    
+    if (!nameInput.value || !fileInput.files[0]) {
+        alert('请填写知识库名称并选择文件');
+        return;
+    }
+    
+    // 获取文件扩展名
+    const fileName = fileInput.files[0].name;
+    const fileExt = fileName.substring(fileName.lastIndexOf('.')).toLowerCase();
+    
+    // 检查文件类型
+    const supportedExts = ['.txt', '.pdf', '.md', '.jpg', '.jpeg', '.png', '.gif', '.bmp'];
+    if (!supportedExts.includes(fileExt)) {
+        alert(`不支持的文件类型: ${fileExt}。\n支持的文件类型: ${supportedExts.join(', ')}`);
+        return;
+    }
+    
+    // 检查文件大小
+    const fileSize = fileInput.files[0].size;
+    const fileSizeMB = fileSize / (1024 * 1024);
+    const maxFileSizeMB = 16; // 最大16MB
+    
+    if (fileSizeMB > maxFileSizeMB) {
+        alert(`文件大小(${fileSizeMB.toFixed(2)}MB)超过最大限制(${maxFileSizeMB}MB)`);
+        return;
+    }
+    
+    // 显示进度条并禁用提交按钮
+    progressBar.style.display = 'block';
+    progressBar.querySelector('.progress-bar').style.width = '0%';
+    submitBtn.disabled = true;
+    submitBtn.innerHTML = '<span class="spinner-border spinner-border-sm" role="status" aria-hidden="true"></span> 上传中...';
+    
+    const formData = new FormData();
+    formData.append('name', nameInput.value);
+    formData.append('file', fileInput.files[0]);
+    
+    try {
+        // 使用 fetch 的 timeout 设置
+        const controller = new AbortController();
+        const timeoutId = setTimeout(() => controller.abort(), 60000); // 60秒超时
         
-        if (!nameInput.value || !fileInput.files[0]) {
-            alert('请填写知识库名称并选择文件');
-            return;
-        }
+        const response = await fetch('/api/knowledge', {
+            method: 'POST',
+            body: formData,
+            signal: controller.signal
+        });
         
-        // 显示进度条
-        progressBar.style.display = 'block';
-        progressBar.querySelector('.progress-bar').style.width = '0%';
+        clearTimeout(timeoutId);
         
-        const formData = new FormData();
-        formData.append('name', nameInput.value);
-        formData.append('file', fileInput.files[0]);
+        const result = await response.json();
         
-        try {
-            const response = await fetch('/api/knowledge', {
-                method: 'POST',
-                body: formData
-            });
-            
-            const result = await response.json();
-            
-            if (result.success) {
-                // 成功启动处理，轮询进度
-                pollProgress(result.task_id, progressBar);
-            } else {
-                alert('创建知识库失败: ' + result.message);
-                progressBar.style.display = 'none';
-            }
-        } catch (error) {
-            console.error('创建知识库出错:', error);
-            alert('创建知识库出错，请查看控制台日志');
+        if (result.success) {
+            // 成功启动处理，轮询进度
+            pollProgress(result.task_id, progressBar, submitBtn);
+        } else {
+            alert('创建知识库失败: ' + result.message);
             progressBar.style.display = 'none';
+            submitBtn.disabled = false;
+            submitBtn.textContent = '创建知识库';
         }
-    }
-
-    // 添加文件到知识库
-    async function addFileToKnowledgeBase(form) {
-        const knowledgeId = form.getAttribute('data-id');
-        const fileInput = form.querySelector('input[type="file"]');
-        const submitBtn = form.querySelector('button[type="submit"]');
-        
-        if (!fileInput.files[0]) {
-            alert('请选择文件');
-            return;
+    } catch (error) {
+        console.error('创建知识库出错:', error);
+        
+        // 判断是否是超时错误
+        let errorMessage = '创建知识库出错，请查看控制台日志';
+        if (error.name === 'AbortError') {
+            errorMessage = '上传请求超时，请检查网络连接或尝试上传较小的文件';
         }
         
-        // 禁用提交按钮
-        submitBtn.disabled = true;
-        submitBtn.textContent = '上传中...';
+        alert(errorMessage);
+        progressBar.style.display = 'none';
+        submitBtn.disabled = false;
+        submitBtn.textContent = '创建知识库';
+    }
+}
+// 添加文件到知识库
+async function addFileToKnowledgeBase(form) {
+    const knowledgeId = form.getAttribute('data-id');
+    const fileInput = form.querySelector('input[type="file"]');
+    const submitBtn = form.querySelector('button[type="submit"]');
+    
+    if (!fileInput.files[0]) {
+        alert('请选择文件');
+        return;
+    }
+    
+    // 获取文件扩展名
+    const fileName = fileInput.files[0].name;
+    const fileExt = fileName.substring(fileName.lastIndexOf('.')).toLowerCase();
+    
+    // 检查文件类型
+    const supportedExts = ['.txt', '.pdf', '.md', '.jpg', '.jpeg', '.png', '.gif', '.bmp'];
+    if (!supportedExts.includes(fileExt)) {
+        alert(`不支持的文件类型: ${fileExt}。\n支持的文件类型: ${supportedExts.join(', ')}`);
+        return;
+    }
+    
+    // 检查文件大小
+    const fileSize = fileInput.files[0].size;
+    const fileSizeMB = fileSize / (1024 * 1024);
+    const maxFileSizeMB = 16; // 最大16MB
+    
+    if (fileSizeMB > maxFileSizeMB) {
+        alert(`文件大小(${fileSizeMB.toFixed(2)}MB)超过最大限制(${maxFileSizeMB}MB)`);
+        return;
+    }
+    
+    // 禁用提交按钮
+    submitBtn.disabled = true;
+    submitBtn.innerHTML = '<span class="spinner-border spinner-border-sm" role="status" aria-hidden="true"></span> 上传中...';
+    
+    const formData = new FormData();
+    formData.append('file', fileInput.files[0]);
+    
+    try {
+        // 使用 fetch 的 timeout 设置
+        const controller = new AbortController();
+        const timeoutId = setTimeout(() => controller.abort(), 60000); // 60秒超时
+        
+        const response = await fetch(`/api/knowledge/${knowledgeId}/documents`, {
+            method: 'POST',
+            body: formData,
+            signal: controller.signal
+        });
         
-        const formData = new FormData();
-        formData.append('file', fileInput.files[0]);
+        clearTimeout(timeoutId);
         
-        try {
-            const response = await fetch(`/api/knowledge/${knowledgeId}/documents`, {
-                method: 'POST',
-                body: formData
-            });
-            
-            const result = await response.json();
-            
-            if (result.success) {
-                alert('文件上传成功，开始处理');
-                // 轮询处理进度
-                pollProgressSimple(result.task_id, submitBtn);
-            } else {
-                alert('添加文件失败: ' + result.message);
-                submitBtn.disabled = false;
-                submitBtn.textContent = '添加文件';
-            }
-        } catch (error) {
-            console.error('添加文件出错:', error);
-            alert('添加文件出错，请查看控制台日志');
+        const result = await response.json();
+        
+        if (result.success) {
+            alert('文件上传成功，开始处理');
+            // 轮询处理进度
+            pollProgressSimple(result.task_id, submitBtn);
+        } else {
+            alert('添加文件失败: ' + result.message);
             submitBtn.disabled = false;
             submitBtn.textContent = '添加文件';
         }
+    } catch (error) {
+        console.error('添加文件出错:', error);
+        
+        // 判断是否是超时错误
+        let errorMessage = '添加文件出错，请查看控制台日志';
+        if (error.name === 'AbortError') {
+            errorMessage = '上传请求超时，请检查网络连接或尝试上传较小的文件';
+        }
+        
+        alert(errorMessage);
+        submitBtn.disabled = false;
+        submitBtn.textContent = '添加文件';
     }
-
+}
     // 删除知识库
     async function deleteKnowledgeBase(knowledgeId) {
         if (!confirm('确定要删除这个知识库吗？此操作不可恢复。')) {
@@ -1920,85 +1998,196 @@
             alert('删除知识库出错，请查看控制台日志');
         }
     }
-
-    // 轮询进度（带进度条显示）
-    function pollProgress(taskId, progressElement) {
-        const progressBar = progressElement.querySelector('.progress-bar');
-        
-        const interval = setInterval(async () => {
-            try {
-                const response = await fetch(`/api/progress/${taskId}`);
-                const data = await response.json();
+// 删除知识库中的单个文件
+async function deleteFile(indexId, encodedFileName) {
+    try {
+        // 解码文件名以在确认对话框中显示
+        const fileName = decodeURIComponent(encodedFileName);
+        
+        if (!confirm(`确定要删除文件 "${fileName}" 吗？此操作不可恢复。`)) {
+            return;
+        }
+        
+        const response = await fetch(`/api/knowledge/${indexId}/documents/${encodedFileName}`, {
+            method: 'DELETE'
+        });
+        
+        const result = await response.json();
+        
+        if (result.success) {
+            alert('文件删除成功!');
+            loadKnowledgeBases(); // 重新加载知识库列表
+        } else {
+            alert('删除文件失败: ' + result.message);
+        }
+    } catch (error) {
+        console.error('删除文件出错:', error);
+        alert('删除文件出错，请查看控制台日志');
+    }
+}
+// 轮询进度（带进度条显示）
+function pollProgress(taskId, progressElement, submitBtn) {
+    const progressBar = progressElement.querySelector('.progress-bar');
+    let retryCount = 0;
+    const maxRetries = 3;
+    
+    const interval = setInterval(async () => {
+        try {
+            const response = await fetch(`/api/progress/${taskId}`);
+            
+            if (!response.ok) {
+                throw new Error(`HTTP错误: ${response.status}`);
+            }
+            
+            const data = await response.json();
+            
+            if (data.success) {
+                const progressData = data.data;
                 
-                if (data.success) {
-                    const progressData = data.data;
-                    
-                    // 更新进度条
-                    progressBar.style.width = progressData.progress + '%';
+                // 更新进度条
+                progressBar.style.width = progressData.progress + '%';
+                progressBar.setAttribute('aria-valuenow', progressData.progress);
+                
+                // 更新进度状态
+                if (submitBtn.tagName === 'BUTTON') {
+                    submitBtn.innerHTML = `<i class="bi bi-arrow-repeat spin me-1"></i> ${progressData.status || '处理中...'}`;
+                }
+                
+                // 处理完成或出错时
+                if (progressData.progress >= 100 || progressData.error) {
+                    clearInterval(interval);
                     
-                    // 处理完成或出错时
-                    if (progressData.progress >= 100 || progressData.error) {
-                        clearInterval(interval);
-                        
-                        if (progressData.error) {
-                            alert('处理失败: ' + progressData.status);
-                        } else {
-                            alert(`知识库处理成功! 已处理${progressData.docCount}个文档片段`);
-                        }
-                        
-                        // 隐藏进度条
-                        progressElement.style.display = 'none';
-                        
-                        // 重置表单
-                        document.getElementById('knowledge-name').value = '';
-                        document.getElementById('knowledge-file').value = '';
-                        
-                        // 重新加载知识库列表
-                        loadKnowledgeBases();
+                    if (progressData.error) {
+                        alert('处理失败: ' + progressData.status);
+                    } else {
+                        alert(`知识库处理成功! 已处理${progressData.docCount}个文档片段`);
                     }
+                    
+                    // 隐藏进度条
+                    progressElement.style.display = 'none';
+                    
+                    // 恢复按钮状态
+                    submitBtn.disabled = false;
+                    submitBtn.innerHTML = '创建知识库';
+                    
+                    // 重置表单
+                    document.getElementById('knowledge-name').value = '';
+                    document.getElementById('knowledge-file').value = '';
+                    
+                    // 重新加载知识库列表
+                    loadKnowledgeBases();
                 }
-            } catch (error) {
-                console.error('获取进度信息出错:', error);
+            } else {
+                throw new Error('轮询返回错误状态');
             }
-        }, 1000); // 每秒轮询一次
-    }
+            
+            // 重置重试计数
+            retryCount = 0;
+            
+        } catch (error) {
+            console.error('获取进度信息出错:', error);
+            retryCount++;
+            
+            if (retryCount >= maxRetries) {
+                clearInterval(interval);
+                alert('获取进度信息失败，请检查控制台日志');
+                
+                // 恢复按钮状态
+                submitBtn.disabled = false;
+                submitBtn.innerHTML = '创建知识库';
+                progressElement.style.display = 'none';
+            }
+        }
+    }, 1000); // 每秒轮询一次
+}
 
-    // 简化版轮询进度（不显示进度条）
-    function pollProgressSimple(taskId, buttonElement) {
-        const originalText = buttonElement.textContent;
-        
-        const interval = setInterval(async () => {
-            try {
-                const response = await fetch(`/api/progress/${taskId}`);
-                const data = await response.json();
+// 简化版轮询进度（不显示进度条）
+function pollProgressSimple(taskId, buttonElement) {
+    const originalText = buttonElement.textContent;
+    let retryCount = 0;
+    const maxRetries = 3;
+    
+    buttonElement.innerHTML = '<i class="bi bi-arrow-repeat spin me-1"></i> 处理中...';
+    
+    const interval = setInterval(async () => {
+        try {
+            const response = await fetch(`/api/progress/${taskId}`);
+            
+            if (!response.ok) {
+                throw new Error(`HTTP错误: ${response.status}`);
+            }
+            
+            const data = await response.json();
+            
+            if (data.success) {
+                const progressData = data.data;
                 
-                if (data.success) {
-                    const progressData = data.data;
+                // 更新按钮文本显示进度
+                buttonElement.innerHTML = `<i class="bi bi-arrow-repeat spin me-1"></i> ${progressData.status || '处理中...'} (${progressData.progress}%)`;
+                
+                // 处理完成或出错时
+                if (progressData.progress >= 100 || progressData.error) {
+                    clearInterval(interval);
                     
-                    // 处理完成或出错时
-                    if (progressData.progress >= 100 || progressData.error) {
-                        clearInterval(interval);
-                        
-                        if (progressData.error) {
-                            alert('处理失败: ' + progressData.status);
-                        } else {
-                            alert(`处理成功! 已处理${progressData.docCount}个文档片段`);
-                        }
-                        
-                        // 恢复按钮
-                        buttonElement.disabled = false;
-                        buttonElement.textContent = originalText;
-                        
-                        // 重新加载知识库列表
-                        loadKnowledgeBases();
+                    if (progressData.error) {
+                        alert('处理失败: ' + progressData.status);
+                    } else {
+                        alert(`处理成功! 已处理${progressData.docCount}个文档片段`);
                     }
+                    
+                    // 恢复按钮
+                    buttonElement.disabled = false;
+                    buttonElement.textContent = originalText;
+                    
+                    // 重新加载知识库列表
+                    loadKnowledgeBases();
                 }
-            } catch (error) {
-                console.error('获取进度信息出错:', error);
+            } else {
+                throw new Error('轮询返回错误状态');
+            }
+            
+            // 重置重试计数
+            retryCount = 0;
+            
+        } catch (error) {
+            console.error('获取进度信息出错:', error);
+            retryCount++;
+            
+            if (retryCount >= maxRetries) {
+                clearInterval(interval);
+                alert('获取进度信息失败，请检查控制台日志');
+                
+                // 恢复按钮
+                buttonElement.disabled = false;
+                buttonElement.textContent = originalText;
             }
-        }, 1000); // 每秒轮询一次
+        }
+    }, 1000); // 每秒轮询一次
+}
+// 根据文件扩展名获取图标类
+function getFileIconClass(fileName) {
+    if (!fileName) return 'bi-file';
+    
+    const ext = fileName.split('.').pop().toLowerCase();
+    
+    switch (ext) {
+        case 'pdf': return 'bi-file-pdf';
+        case 'doc': 
+        case 'docx': return 'bi-file-earmark-word';
+        case 'xls':
+        case 'xlsx': return 'bi-file-earmark-excel';
+        case 'ppt':
+        case 'pptx': return 'bi-file-earmark-ppt';
+        case 'jpg':
+        case 'jpeg':
+        case 'png':
+        case 'gif':
+        case 'bmp': return 'bi-file-image';
+        case 'md': return 'bi-markdown';
+        case 'txt': return 'bi-file-text';
+        default: return 'bi-file';
     }
-
+}
     // 使用AI生成工作流
     async function generateAIWorkflow() {
         const description = document.getElementById('agent-description').value;

uploads/496449c167344601ba87c24743701b99.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a046fe2339ff94de24bb636ebdffbb108629131e633e3e6e9090b9550f0afeda
+size 499224

uploads/eb55dde866f64be4a1fdbc81d3d1df0e.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a046fe2339ff94de24bb636ebdffbb108629131e633e3e6e9090b9550f0afeda
+size 499224