import json from openai import OpenAI import requests import os import time from multiprocessing import Pool, Manager import threading import re import random # ================= 配置区域 ================= API_KEY = "token-abc123" BASE_URL = "http://localhost:1054/v1" MODEL_NAME = "/mnt/tidal-alsh01/usr/dawo/dawo/model_ori/Qwen/Qwen2.5-72B-Instruct" def get_response(q): client = OpenAI(api_key=API_KEY, base_url=BASE_URL) response = client.chat.completions.create( model=MODEL_NAME, messages=[ {"role": "user", "content": q} ], temperature=0.7, ) return response.choices[0].message.content class ALLIN: def __init__(self): # 使用原参考代码2的 API Key self.api_key = "9b6dce6b66eb42f49aa74635e5484de9" self.url = "https://runway.devops.rednote.life/openai/google/v1:generateContent" self.headers = { "api-key": self.api_key, "Content-Type": "application/json" } def get_response(self, messages, max_tokens=8192, temperature=0.6): """ 调用Gemini API获取响应 """ try: # 提取 system 和 user 消息 system_text = None user_text = None for m in messages: if m["role"] == "system": system_text = m["content"] elif m["role"] == "user": user_text = m["content"] # 构建请求体 data = { "contents": [ { "role": "user", "parts": [{"text": user_text or ""}] } ], "generationConfig": { "temperature": temperature, "maxOutputTokens": max_tokens, "topP": 1, "seed": 0 # , # "thinkingConfig": { # "includeThoughts": True # } } } if system_text: data["systemInstruction"] = { "parts": [{"text": system_text}] } response = requests.post(self.url, headers=self.headers, json=data, timeout=180) res_dict = response.json() # 解析返回的文本 return res_dict["candidates"][0]["content"]["parts"][0]["text"] except Exception as e: print(f"Gemini 调用失败: {str(e)}", flush=True) return None llm = ALLIN() messages = [ { "role": "system", "content": "" }, { "role": "user", "content": "你好,现在是什么时间" } ] print(llm.get_response(messages))