mirror of
https://github.com/Nick-Hogo/AgentSeed
synced 2026-11-16 04:04:22 +08:00
32 lines
1.0 KiB
Python
32 lines
1.0 KiB
Python
import httpx
|
|
|
|
base_url = "https://api.example.com/v1/messages"
|
|
api_key = "sk-xxxx"
|
|
model = "claude-fable-5"
|
|
|
|
while True: # 持续接收用户输入，实现最基础的命令行对话。
|
|
user_input = input("You> ")
|
|
|
|
# 按照 Anthropic Messages API 协议向模型服务发送 HTTP 请求。
|
|
response = httpx.post(
|
|
base_url,
|
|
headers={
|
|
"x-api-key": api_key, # Anthropic 格式的认证信息
|
|
"anthropic-version": "2023-06-01", # Anthropic API 版本
|
|
},
|
|
json={
|
|
"model": model, # 选择的模型名称
|
|
"messages": [
|
|
{"role": "user", "content": user_input}, # 用户输入的消息
|
|
],
|
|
},
|
|
timeout=60,
|
|
)
|
|
|
|
# 将 JSON 响应转换为字典，然后取出模型生成的文本。
|
|
data = response.json()
|
|
print("Assistant>", end=" ")
|
|
for block in data["content"]: # 循环解析LLM返回信息
|
|
if block["type"] == "text":
|
|
print(block["text"], end="")
|
|
print() |