Files
pj_llama/test_all_models.py
aken1023 c6cc91da7f 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
2025-09-19 21:38:15 +08:00

143 lines
4.6 KiB
Python

import requests
import json
import time
API_KEY = "paVrIT+XU1NhwCAOb0X4aYi75QKogK5YNMGvQF1dCyo="
BASE_URL = "https://llama.theaken.com/v1"
MODELS = [
"gpt-oss-120b",
"deepseek-r1-671b",
"qwen3-embedding-8b"
]
def test_model(model_name):
"""測試單個模型"""
print(f"\n[測試模型: {model_name}]")
print("-" * 40)
headers = {
"Authorization": f"Bearer {API_KEY}",
"Content-Type": "application/json"
}
# 測試聊天完成端點
chat_url = f"{BASE_URL}/chat/completions"
data = {
"model": model_name,
"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Say 'Hello, I am working!' if you can see this message."}
],
"temperature": 0.5,
"max_tokens": 50
}
try:
print(f"連接到: {chat_url}")
response = requests.post(chat_url, headers=headers, json=data, timeout=30)
print(f"HTTP 狀態碼: {response.status_code}")
if response.status_code == 200:
result = response.json()
if 'choices' in result and len(result['choices']) > 0:
content = result['choices'][0]['message']['content']
print(f"[SUCCESS] AI 回應: {content}")
return True
else:
print("[ERROR] 回應格式異常")
print(f"回應內容: {json.dumps(result, indent=2)}")
else:
print(f"[ERROR] 錯誤回應")
# 檢查是否是 HTML 錯誤頁面
if response.text.startswith('<!DOCTYPE'):
print("收到 HTML 錯誤頁面 (可能是 502 Bad Gateway)")
else:
print(f"回應內容: {response.text[:300]}")
except requests.exceptions.Timeout:
print("[TIMEOUT] 請求超時 (30秒)")
except requests.exceptions.ConnectionError as e:
print(f"[CONNECTION ERROR]: {str(e)[:100]}")
except Exception as e:
print(f"[UNEXPECTED ERROR]: {str(e)[:100]}")
return False
def test_api_endpoints():
"""測試不同的 API 端點"""
print("\n[測試 API 端點可用性]")
print("=" * 50)
headers = {
"Authorization": f"Bearer {API_KEY}",
"Content-Type": "application/json"
}
# 測試不同的可能端點
endpoints = [
f"{BASE_URL}/models",
f"{BASE_URL}/chat/completions",
BASE_URL
]
for endpoint in endpoints:
try:
print(f"\n測試端點: {endpoint}")
response = requests.get(endpoint, headers=headers, timeout=10)
print(f" 狀態碼: {response.status_code}")
if response.status_code == 200:
print(" [OK] 端點可訪問")
# 如果是 JSON 回應,顯示部分內容
try:
data = response.json()
print(f" 回應類型: JSON")
if 'data' in data:
print(f" 包含 {len(data['data'])} 項資料")
except:
print(f" 回應類型: {response.headers.get('content-type', 'unknown')}")
elif response.status_code == 405:
print(" [OK] 端點存在 (但不支援 GET 方法)")
elif response.status_code == 502:
print(" [ERROR] 502 Bad Gateway - 伺服器暫時無法使用")
else:
print(f" [ERROR] 無法訪問")
except Exception as e:
print(f" [ERROR]: {str(e)[:50]}")
def main():
print("=" * 50)
print("Llama API 完整測試程式")
print("=" * 50)
print(f"API 基礎 URL: {BASE_URL}")
print(f"API 金鑰: {API_KEY[:10]}...{API_KEY[-5:]}")
# 首先測試端點可用性
test_api_endpoints()
print("\n" + "=" * 50)
print("開始測試各個模型")
print("=" * 50)
success_count = 0
for model in MODELS:
if test_model(model):
success_count += 1
time.sleep(1) # 避免請求過快
print("\n" + "=" * 50)
print(f"測試結果: {success_count}/{len(MODELS)} 個模型成功連接")
if success_count == 0:
print("\n可能的問題:")
print("1. API 伺服器暫時離線 (502 錯誤)")
print("2. API 金鑰可能不正確")
print("3. 網路連接問題")
print("4. 防火牆或代理設定")
print("\n建議稍後再試,或聯繫 API 提供者確認服務狀態。")
if __name__ == "__main__":
main()