Upload app.py

app.py

@@ -37,11 +37,23 @@ session = create_session_with_retry()
 # 初始化 Hugging Face Inference Client
 # 使用环境变量或者免费的公开模型
 HF_TOKEN = os.getenv("HF_TOKEN", None)  # 可选：如果需要访问私有模型
-try:
-    client = InferenceClient(token=HF_TOKEN)
-except Exception as e:
-    print(f"Warning: Failed to initialize Hugging Face client: {e}")
-    client = None
+
+# 尝试初始化 client
+if HF_TOKEN:
+    try:
+        client = InferenceClient(token=HF_TOKEN)
+        print("✅ Hugging Face client initialized successfully with token")
+    except Exception as e:
+        print(f"⚠️ Warning: Failed to initialize Hugging Face client with token: {e}")
+        client = None
+else:
+    # 没有 token，使用公开访问（有速率限制）
+    try:
+        client = InferenceClient()
+        print("⚠️ Using Hugging Face client without token (rate limited)")
+    except Exception as e:
+        print(f"❌ Warning: Failed to initialize Hugging Face client: {e}")
+        client = None
 
 # 定义可用的 MCP 工具
 MCP_TOOLS = [
@@ -666,6 +678,7 @@ Always be helpful, accurate, and cite the data sources when providing financial
             try:
                 # 检查 client 是否可用
                 if client is None:
+                    print("⚠️ LLM client not available, using fallback")
                     return fallback_chatbot_response(message)
                 
                 response = client.chat_completion(
@@ -676,17 +689,22 @@ Always be helpful, accurate, and cite the data sources when providing financial
                     temperature=0.7
                 )
                 
+                print(f"✅ LLM response received (iteration {iteration})")
+                
                 choice = response.choices[0]
                 
                 # 检查是否有工具调用
                 if choice.message.tool_calls:
                     # 有工具调用
+                    print(f"🔧 Tool calls detected: {len(choice.message.tool_calls)}")
                     messages.append(choice.message)
                     
                     for tool_call in choice.message.tool_calls:
                         tool_name = tool_call.function.name
                         tool_args = json.loads(tool_call.function.arguments)
                         
+                        print(f"  → Calling tool: {tool_name} with args: {tool_args}")
+                        
                         # 记录工具调用
                         tool_calls_log.append({
                             "name": tool_name,
@@ -696,6 +714,8 @@ Always be helpful, accurate, and cite the data sources when providing financial
                         # 调用 MCP 工具
                         tool_result = call_mcp_tool(tool_name, tool_args)
                         
+                        print(f"  ← Tool result received")
+                        
                         # 将工具结果添加到消息列表
                         messages.append({
                             "role": "tool",
@@ -708,11 +728,23 @@ Always be helpful, accurate, and cite the data sources when providing financial
                     continue
                 else:
                     # 没有工具调用，直接返回回答
+                    print(f"💬 Final response generated")
                     response_text = choice.message.content
                     break
                     
+            except ValueError as ve:
+                # API Key 相关错误
+                error_msg = str(ve)
+                if "api_key" in error_msg.lower() or "token" in error_msg.lower():
+                    print(f"❌ LLM API authentication error: {ve}")
+                    print("ℹ️ Falling back to simple response logic")
+                    return fallback_chatbot_response(message)
+                else:
+                    raise
             except Exception as e:
-                # 如果 LLM API 失败，退回到简单逻辑
+                # 其他 LLM API 错误
+                print(f"❌ LLM API error: {e}")
+                print("ℹ️ Falling back to simple response logic")
                 return fallback_chatbot_response(message)
         
         # 构建最终响应
@@ -794,6 +826,30 @@ def query_with_status(company, query_type):
 with gr.Blocks(title="SEC Financial Data Query Assistant") as demo:
     gr.Markdown("# 🤖 SEC Financial Data Query Assistant")
     
+    # 显示 HF_TOKEN 状态提示
+    if not HF_TOKEN:
+        gr.Markdown("""
+        <div style='padding: 15px; background: #fff3cd; border-left: 4px solid #ffc107; margin: 10px 0; border-radius: 4px;'>
+            <strong>⚠️ Note:</strong> AI Assistant is running in <strong>rate-limited mode</strong> without HF_TOKEN. 
+            For better performance, set the <code>HF_TOKEN</code> environment variable in your Space settings.
+            <br><br>
+            <strong>How to get your token:</strong>
+            <ol>
+                <li>Go to <a href="https://huggingface.co/settings/tokens" target="_blank">Hugging Face Tokens</a></li>
+                <li>Create a new token (Read access is sufficient)</li>
+                <li>Add it as a Secret in your Space settings: Settings → Repository secrets → New secret</li>
+                <li>Name: <code>HF_TOKEN</code>, Value: <code>your_token_here</code></li>
+                <li>Restart the Space</li>
+            </ol>
+        </div>
+        """)
+    else:
+        gr.Markdown("""
+        <div style='padding: 15px; background: #d4edda; border-left: 4px solid #28a745; margin: 10px 0; border-radius: 4px;'>
+            <strong>✅ AI Mode Active:</strong> Full LLM capabilities enabled with Qwen/Qwen2.5-72B-Instruct model.
+        </div>
+        """)
+    
     with gr.Tab("AI Assistant"):
         # 使用 Gradio ChatInterface（兼容 4.44.1）
         chat = gr.ChatInterface(
@@ -801,10 +857,16 @@ with gr.Blocks(title="SEC Financial Data Query Assistant") as demo:
             examples=[
                 "Show me Apple's latest financial data",
                 "What's NVIDIA's 3-year financial trend?",
-                "Get Microsoft's 5-year financial trends",
-                "How is Tesla performing financially?"
+                "Compare Tesla's revenue with its expenses",
+                "How is Microsoft performing financially?",
+                "What can you tell me about Amazon's recent earnings?",
+                "Give me Alibaba's 5-year financial overview",
+                "Hello! What can you help me with?",
+                "What's the weather like today?",  # 测试非财务问题
             ],
-            cache_examples=False
+            cache_examples=False,
+            title="🤖 Intelligent Financial Assistant",
+            description="Ask me anything! I can help with financial data queries or general conversations. When you ask about companies, I'll automatically fetch real-time SEC EDGAR data."
         )
     
     with gr.Tab("Direct Query"):