Update app.py

app.py

@@ -173,47 +173,63 @@ class SessionManager:
 class ImageService:
     @staticmethod
     def generate_image(prompt: str) -> str:
-        """Generate image using grok-latest-image model and return markdown URL"""
         try:
-            # Generate image using grok-latest-image
+            logging.info(f"开始生成图片，提示词: {prompt}")
+            
             response = requests.post(
                 IMAGE_MODEL_URL,
-                headers={'Authorization': f'Bearer {IMAGE_MODEL_KEY}'},
+                headers={
+                    'Content-Type': 'application/json',
+                    'Authorization': f'Bearer {IMAGE_MODEL_KEY}'
+                },
                 json={
                     "model": "grok-latest-image",
-                    "prompt": prompt
+                    "messages": [{
+                        "role": "user",
+                        "content": prompt
+                    }],
+                    "stream": False
                 }
             )
+            
+            logging.info(f"图片生成服务响应状态码: {response.status_code}")
             response.raise_for_status()
+            
             result = response.json()
+            logging.info(f"图片生成服务响应内容: {json.dumps(result, ensure_ascii=False)}")
             
-            if not result.get('url'):
-                raise ValueError("No image URL in response")
-
-            return f"![image]({result['url']})"
+            if not result.get('choices') or not result['choices'][0].get('message', {}).get('content'):
+                raise ValueError("Invalid response format")
+                
+            image_url = result['choices'][0]['message']['content']
+            logging.info(f"成功获取图片URL: {image_url}")
+            
+            return image_url
         except Exception as e:
             logging.error(f"Image generation error: {str(e)}")
             raise
 
     @staticmethod
     def get_media_id(image_url: str) -> str:
-        """Download image and upload to WeChat media storage"""
         try:
-            # Download image
+            logging.info(f"开始下载图片: {image_url}")
             image_response = requests.get(image_url)
             image_response.raise_for_status()
             image_data = image_response.content
-
-            # Upload to WeChat
+            
+            logging.info("开始上传图片到微信服务器")
             upload_url = f'https://api.weixin.qq.com/cgi-bin/media/upload?access_token={TOKEN}&type=image'
             files = {'media': ('image.jpg', image_data, 'image/jpeg')}
             response = requests.post(upload_url, files=files)
             response.raise_for_status()
+            
             result = response.json()
+            logging.info(f"微信服务器响应: {json.dumps(result, ensure_ascii=False)}")
             
             if 'media_id' not in result:
                 raise ValueError(f"Failed to get media_id: {result}")
                 
+            logging.info(f"成功获取media_id: {result['media_id']}")
             return result['media_id']
         except Exception as e:
             logging.error(f"WeChat media upload error: {str(e)}")
@@ -269,6 +285,7 @@ def generate_response_xml(to_user, from_user, content, response_type='text', med
     nonce = ''.join(random.choices(string.ascii_letters + string.digits, k=10))
     
     if response_type == 'image' and media_id:
+        logging.info(f"生成图片消息响应，media_id: {media_id}")
         xml_content = f'''
         <xml>
             <ToUserName><![CDATA[{to_user}]]></ToUserName>
@@ -282,6 +299,7 @@ def generate_response_xml(to_user, from_user, content, response_type='text', med
         '''
     else:
         formatted_content = convert_markdown_to_wechat(content)
+        logging.info(f"生成文本消息响应: {formatted_content}")
         xml_content = f'''
         <xml>
             <ToUserName><![CDATA[{to_user}]]></ToUserName>
@@ -333,24 +351,25 @@ def append_status_message(content, has_pending_parts=False, is_processing=False)
 
 def process_ai_response(messages):
     try:
+        logging.info("开始处理AI响应")
         completion = client.chat.completions.create(
             model="o3-mini",
             messages=messages,
             tools=TOOLS,
             tool_choice="auto"
         )
+        logging.info("收到AI响应")
         
         # Handle tool calls if present
         if completion.choices[0].message.tool_calls:
+            logging.info("检测到工具调用")
             for tool_call in completion.choices[0].message.tool_calls:
                 if tool_call.function.name == "generate_image":
                     try:
+                        logging.info("开始执行图片生成")
                         args = json.loads(tool_call.function.arguments)
                         # Generate image and get markdown URL
-                        image_markdown = ImageService.generate_image(args['prompt'])
-                        
-                        # Extract image URL from markdown
-                        image_url = re.search(r'\((.*?)\)', image_markdown).group(1)
+                        image_url = ImageService.generate_image(args['prompt'])
                         
                         # Get WeChat media_id
                         media_id = ImageService.get_media_id(image_url)
@@ -366,7 +385,7 @@ def process_ai_response(messages):
                             "media_id": media_id
                         }
                     except Exception as e:
-                        logging.error(f"Image generation process failed: {str(e)}")
+                        logging.error(f"图片生成过程失败: {str(e)}")
                         return {
                             "type": "text",
                             "content": f"抱歉，图片生成失败：{str(e)}",
@@ -374,6 +393,7 @@ def process_ai_response(messages):
                         }
         
         # Handle normal text response
+        logging.info("处理普通文本响应")
         response_content = completion.choices[0].message.content
         messages.append({
             "role": "assistant",
@@ -390,10 +410,12 @@ def process_ai_response(messages):
 
 def handle_async_task(session, task_id, messages):
     try:
+        logging.info(f"开始处理异步任务: {task_id}")
         if task_id not in session.response_queue:
             return
             
         result = process_ai_response(messages)
+        logging.info(f"异步任务处理完成: {task_id}")
         
         if task_id in session.response_queue and not session.response_queue[task_id].is_expired():
             session.response_queue[task_id].status = "completed"
@@ -401,6 +423,7 @@ def handle_async_task(session, task_id, messages):
             session.response_queue[task_id].response_type = result.get("type")
             session.response_queue[task_id].media_id = result.get("media_id")
     except Exception as e:
+        logging.error(f"异步任务处理失败: {str(e)}")
         if task_id in session.response_queue:
             session.response_queue[task_id].status = "failed"
             session.response_queue[task_id].error = str(e)
@@ -541,21 +564,27 @@ def wechatai():
                 encrypt_type=encrypt_type
             )
 
+        # Regular message processing
+        logging.info("准备开始处理用户消息")
         session.messages.append({"role": "user", "content": user_content})
         
         task_id = str(uuid.uuid4())
         session.current_task = task_id
         session.response_queue[task_id] = AsyncResponse()
         
-        # Replace asyncio.run with direct call
-        handle_async_task(session, task_id, session.messages.copy())
+        # Submit task to executor
+        logging.info(f"提交任务到执行器: {task_id}")
+        executor.submit(handle_async_task, session, task_id, session.messages.copy())
         
-        return generate_response_xml(
+        # Return immediate response
+        initial_response = generate_response_xml(
             from_user,
             to_user,
             append_status_message(generate_initial_response(), is_processing=True),
             encrypt_type=encrypt_type
         )
+        logging.info("返回初始响应给用户")
+        return initial_response
 
     except Exception as e:
         logging.error(f"处理请求时出错: {str(e)}")