ai-agent-book 精选快照（<2MB 代码与文档，来自 github.com/bojieli/ai-agent-book）

chapter2/prompt-engineering/test_openrouter_direct.py

@@ -0,0 +1,89 @@
+#!/usr/bin/env python3
+"""
+Direct test of OpenRouter API connection with openai/gpt-5
+This helps isolate API connection issues from tau-bench logic
+"""
+
+import os
+import sys
+
+try:
+    from dotenv import load_dotenv
+    load_dotenv()
+except ImportError:
+    pass
+
+from litellm import completion
+
+def test_openrouter():
+    """Test direct API call to OpenRouter"""
+    
+    print("="*60)
+    print("🔍 Testing OpenRouter API directly")
+    print("="*60)
+    
+    # Check API key
+    api_key = os.environ.get("OPENROUTER_API_KEY")
+    if not api_key:
+        print("❌ OPENROUTER_API_KEY not set!")
+        print("   Please set: export OPENROUTER_API_KEY='your_key'")
+        return
+    else:
+        print(f"✅ OPENROUTER_API_KEY found (length: {len(api_key)})")
+    
+    # Test message
+    messages = [
+        {"role": "system", "content": "You are a helpful assistant."},
+        {"role": "user", "content": "Say 'Hello, I'm working!' in exactly 5 words."}
+    ]
+    
+    print("\n📤 Sending test message to OpenRouter...")
+    print(f"   Model: openai/gpt-5")
+    print(f"   Provider: openrouter")
+    print(f"   Messages: {len(messages)}")
+    
+    # Enable verbose logging
+    os.environ["LITELLM_LOG"] = "DEBUG"
+    
+    try:
+        # Make API call with reasoning_effort set to low via extra_body
+        response = completion(
+            model="openai/gpt-5",
+            custom_llm_provider="openrouter",
+            messages=messages,
+            temperature=1.0,  # gpt-5 only supports 1.0
+            # Add reasoning_effort to minimize thinking tokens via extra_body
+            extra_body={"reasoning_effort": "minimal"}  # Options: "minimal", "low", "medium", "high"
+        )
+        
+        print("\n✅ SUCCESS! Response received:")
+        print("─"*50)
+        print(f"Content: {response.choices[0].message.content}")
+        print(f"Model: {response.model}")
+        print(f"Provider: {response._hidden_params.get('custom_llm_provider', 'unknown')}")
+        if hasattr(response, 'usage'):
+            print(f"Tokens used: {response.usage}")
+        print("─"*50)
+        
+    except Exception as e:
+        print(f"\n❌ ERROR: {type(e).__name__}")
+        print(f"   {str(e)}")
+        
+        # Check common issues
+        if "401" in str(e) or "Unauthorized" in str(e):
+            print("\n💡 This looks like an authentication issue.")
+            print("   Check that your OPENROUTER_API_KEY is valid.")
+        elif "404" in str(e):
+            print("\n💡 This might mean the model 'openai/gpt-5' is not available.")
+            print("   Check OpenRouter's model list for available models.")
+        elif "429" in str(e):
+            print("\n💡 Rate limit exceeded. Wait a bit and try again.")
+        elif "timeout" in str(e).lower():
+            print("\n💡 Connection timeout. Check your network connection.")
+        
+        import traceback
+        print("\nFull traceback:")
+        traceback.print_exc()
+
+if __name__ == "__main__":
+    test_openrouter()