添加前端输出

deepsearcher/backend/templates/index.html

@@ -14,6 +14,7 @@
             --success-color: #10b981;
             --error-color: #ef4444;
             --warning-color: #f59e0b;
+            --info-color: #3b82f6;
         }
 
         * {
@@ -191,6 +192,29 @@
             line-height: 1.6;
         }
 
+        /* 进度日志样式 */
+        .progress-log {
+            background-color: #1e1e1e;
+            color: #d4d4d4;
+            border-radius: 6px;
+            padding: 16px;
+            font-family: 'Courier New', monospace;
+            font-size: 14px;
+            margin: 16px 0;
+            max-height: 300px;
+            overflow-y: auto;
+            display: none;
+        }
+
+        .progress-log.visible {
+            display: block;
+        }
+
+        .progress-log-entry {
+            margin: 4px 0;
+            line-height: 1.4;
+        }
+
         footer {
             text-align: center;
             margin-top: 30px;
@@ -266,6 +290,10 @@
                 </div>
                 <button id="queryBtn">执行查询</button>
                 <div id="queryStatus" class="status"></div>
+                
+                <!-- 进度日志显示区域 -->
+                <div id="progressLog" class="progress-log"></div>
+                
                 <div id="queryResult" class="result-container">
                     <h3>查询结果:</h3>
                     <div class="query-result" id="resultText"></div>
@@ -319,6 +347,37 @@
             resultElement.classList.remove('visible');
         }
 
+        // 工具函数：显示进度日志
+        function showProgressLog(messages) {
+            const logElement = document.getElementById('progressLog');
+            if (messages && messages.length > 0) {
+                logElement.innerHTML = messages.map(msg => 
+                    `<div class="progress-log-entry">${escapeHtml(msg)}</div>`
+                ).join('');
+                logElement.classList.add('visible');
+            } else {
+                logElement.classList.remove('visible');
+            }
+        }
+
+        // 工具函数：隐藏进度日志
+        function hideProgressLog() {
+            const logElement = document.getElementById('progressLog');
+            logElement.classList.remove('visible');
+        }
+
+        // 工具函数：转义HTML特殊字符
+        function escapeHtml(text) {
+            const map = {
+                '&': '&amp;',
+                '<': '&lt;',
+                '>': '&gt;',
+                '"': '&quot;',
+                "'": '&#039;'
+            };
+            return text.replace(/[&<>"']/g, function(m) { return map[m]; });
+        }
+
         // 工具函数：设置按钮加载状态
         function setButtonLoading(button, loading) {
             if (loading) {
@@ -347,6 +406,7 @@
             setButtonLoading(button, true);
             showStatus('loadStatus', '正在加载文件...', 'loading');
             hideResult();
+            hideProgressLog();
             
             try {
                 const response = await fetch('/load-files/', {
@@ -392,6 +452,7 @@
             setButtonLoading(button, true);
             showStatus('webLoadStatus', '正在加载网站内容...', 'loading');
             hideResult();
+            hideProgressLog();
             
             try {
                 const response = await fetch('/load-website/', {
@@ -439,6 +500,7 @@
             setButtonLoading(button, true);
             showStatus('queryStatus', '正在处理查询...', 'loading');
             hideResult();
+            hideProgressLog();
             
             try {
                 const response = await fetch(`/query/?original_query=${encodeURIComponent(queryText)}&max_iter=${maxIter}`, {
@@ -455,6 +517,9 @@
                     document.getElementById('resultText').textContent = data.result;
                     document.getElementById('tokenInfo').textContent = `消耗Token数: ${data.consume_token}`;
                     showResult();
+                    
+                    // 显示进度日志
+                    showProgressLog(data.progress_messages);
                 } else {
                     showStatus('queryStatus', `查询失败: ${data.detail}`, 'error');
                 }

deepsearcher/utils/log.py

@@ -1,5 +1,6 @@
 import logging
-
+import io
+from typing import List
 from termcolor import colored
 
 
@@ -50,6 +51,44 @@ class ColoredFormatter(logging.Formatter):
         # return super().format(record)
 
 
+class StreamCaptureHandler(logging.Handler):
+    """
+    A custom logging handler that captures log messages in a list.
+    
+    This handler stores log messages in memory so they can be retrieved later,
+    which is useful for sending log messages to a frontend.
+    """
+    
+    def __init__(self):
+        super().__init__()
+        self.messages: List[str] = []
+        
+    def emit(self, record):
+        """
+        Emit a log record by storing it in the messages list.
+        
+        Args:
+            record: The log record to emit.
+        """
+        log_entry = self.format(record)
+        self.messages.append(log_entry)
+        
+    def get_messages(self) -> List[str]:
+        """
+        Get all captured log messages.
+        
+        Returns:
+            A list of captured log messages.
+        """
+        return self.messages.copy()
+        
+    def clear_messages(self):
+        """
+        Clear all captured log messages.
+        """
+        self.messages.clear()
+
+
 # config log
 dev_logger = logging.getLogger("dev")
 dev_formatter = ColoredFormatter("%(asctime)s - %(levelname)s - %(message)s")
@@ -60,8 +99,11 @@ dev_logger.setLevel(logging.INFO)
 
 progress_logger = logging.getLogger("progress")
 progress_handler = logging.StreamHandler()
+progress_capture_handler = StreamCaptureHandler()
 progress_handler.setFormatter(ColoredFormatter("%(message)s"))
+progress_capture_handler.setFormatter(logging.Formatter("%(message)s"))
 progress_logger.addHandler(progress_handler)
+progress_logger.addHandler(progress_capture_handler)
 progress_logger.setLevel(logging.INFO)
 
 dev_mode = False
@@ -158,3 +200,21 @@ def color_print(message, **kwargs):
         **kwargs: Additional keyword arguments to pass to the logger.
     """
     progress_logger.info(message)
+    progress_handler.flush()
+
+
+def get_progress_messages() -> List[str]:
+    """
+    Get all captured progress messages.
+    
+    Returns:
+        A list of captured progress messages.
+    """
+    return progress_capture_handler.get_messages()
+
+
+def clear_progress_messages():
+    """
+    Clear all captured progress messages.
+    """
+    progress_capture_handler.clear_messages()

main.py

@@ -155,11 +155,12 @@ def load_files(
         HTTPException: If loading files fails.
     """
     try:
+        # 修复batch_size为None时的问题
         load_from_local_files(
             paths_or_directory=paths,
             collection_name=collection_name,
             collection_description=collection_description,
-            batch_size=batch_size,
+            batch_size=batch_size if batch_size is not None else 256,  # 提供默认值
         )
         return {"message": "Files loaded successfully."}
     except Exception as e:
@@ -205,11 +206,12 @@ def load_website(
         HTTPException: If loading website content fails.
     """
     try:
+        # 修复batch_size为None时的问题
         load_from_website(
             urls=urls,
             collection_name=collection_name,
             collection_description=collection_description,
-            batch_size=batch_size,
+            batch_size=batch_size if batch_size is not None else 256,  # 提供默认值
         )
         return {"message": "Website loaded successfully."}
     except Exception as e:
@@ -244,8 +246,21 @@ def perform_query(
         HTTPException: If the query fails.
     """
     try:
+        # 清除之前的进度消息
+        from deepsearcher.utils.log import clear_progress_messages
+        clear_progress_messages()
+        
         result_text, _, consume_token = query(original_query, max_iter)
-        return {"result": result_text, "consume_token": consume_token}
+        
+        # 获取进度消息
+        from deepsearcher.utils.log import get_progress_messages
+        progress_messages = get_progress_messages()
+        
+        return {
+            "result": result_text, 
+            "consume_token": consume_token,
+            "progress_messages": progress_messages
+        }
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))