Initial commit: Llama API Client with full documentation

- Added complete Python client for Llama AI models
- Support for internal network endpoints (tested and working)
- Support for external network endpoints (configured)
- Interactive chat interface with multiple models
- Automatic endpoint testing and failover
- Response cleaning for special markers
- Full documentation in English and Chinese
- Complete test suite and examples
- MIT License and contribution guidelines

quick_test.py

@@ -0,0 +1,54 @@
+"""
+快速測試內網 Llama API
+"""
+
+from openai import OpenAI
+
+# 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
+
+def quick_test():
+    print("連接到內網 API...")
+    print(f"端點: {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:
+            response = client.chat.completions.create(
+                model="gpt-oss-120b",
+                messages=[
+                    {"role": "user", "content": msg}
+                ],
+                temperature=0.7,
+                max_tokens=200
+            )
+            
+            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}")
+            
+        except Exception as e:
+            print(f"錯誤: {str(e)[:100]}")
+
+if __name__ == "__main__":
+    quick_test()