重構為外網連接版本

主要變更：
- 移除所有內網 IP (192.168.x.x)
- 改用外網端點 (https://llama.theaken.com)
- 新增 llama_external_api.py 專門處理外網連接
- 更新所有文檔為外網版本
- 加入備用端點自動切換機制
- 優化錯誤處理和超時設定

quick_test.py

@@ -1,54 +1,76 @@
 """
-快速測試內網 Llama API
+快速測試 Llama API 外網連接
 """
 
 from openai import OpenAI
+import sys
 
 # API 設定
 API_KEY = "paVrIT+XU1NhwCAOb0X4aYi75QKogK5YNMGvQF1dCyo="
-BASE_URL = "http://192.168.0.6:21180/v1"  # 使用第一個可用端點
+BASE_URL = "https://llama.theaken.com/v1"  # 使用外網端點
 
 def quick_test():
-    print("連接到內網 API...")
-    print(f"端點: {BASE_URL}")
+    print("="*50)
+    print("Llama API 快速測試")
+    print("="*50)
+    print(f"連接到: {BASE_URL}")
     print("-" * 50)
     
-    client = OpenAI(
-        api_key=API_KEY,
-        base_url=BASE_URL
-    )
-    
-    # 測試對話
-    test_messages = [
-        "你好，請自我介紹",
-        "1 + 1 等於多少？",
-        "今天天氣如何？"
-    ]
-    
-    for msg in test_messages:
-        print(f"\n問: {msg}")
+    try:
+        client = OpenAI(
+            api_key=API_KEY,
+            base_url=BASE_URL,
+            timeout=15.0  # 15秒超時
+        )
         
-        try:
-            response = client.chat.completions.create(
-                model="gpt-oss-120b",
-                messages=[
-                    {"role": "user", "content": msg}
-                ],
-                temperature=0.7,
-                max_tokens=200
-            )
+        # 測試對話
+        test_messages = [
+            "你好，請自我介紹",
+            "1 + 1 等於多少？",
+            "今天天氣如何？"
+        ]
+        
+        for msg in test_messages:
+            print(f"\n問: {msg}")
             
-            answer = response.choices[0].message.content
-            # 清理可能的思考標記
-            if "<think>" in answer:
-                answer = answer.split("</think>")[-1].strip()
-            if "<|channel|>" in answer:
-                answer = answer.split("<|message|>")[-1].strip()
+            try:
+                response = client.chat.completions.create(
+                    model="gpt-oss-120b",
+                    messages=[
+                        {"role": "user", "content": msg}
+                    ],
+                    temperature=0.7,
+                    max_tokens=200
+                )
                 
-            print(f"答: {answer}")
-            
-        except Exception as e:
-            print(f"錯誤: {str(e)[:100]}")
+                answer = response.choices[0].message.content
+                # 清理可能的思考標記
+                if "<think>" in answer:
+                    answer = answer.split("</think>")[-1].strip()
+                if "<|channel|>" in answer:
+                    answer = answer.split("<|message|>")[-1].strip()
+                    
+                print(f"答: {answer[:200]}")  # 限制顯示長度
+                
+            except Exception as e:
+                error_msg = str(e)
+                if "502" in error_msg:
+                    print("錯誤: 伺服器暫時無法使用 (502)")
+                elif "timeout" in error_msg.lower():
+                    print("錯誤: 請求超時")
+                else:
+                    print(f"錯誤: {error_msg[:100]}")
+                    
+        print("\n" + "="*50)
+        print("測試完成！")
+        
+    except Exception as e:
+        print(f"\n連接失敗: {str(e)[:100]}")
+        print("\n建議:")
+        print("1. 檢查網路連接")
+        print("2. 確認可以訪問 https://llama.theaken.com")
+        print("3. 稍後再試（如果是 502 錯誤）")
+        sys.exit(1)
 
 if __name__ == "__main__":
     quick_test()