Python 调用大模型 API 接口开发指南

选择建议：初学者从 OpenAI/Groq 开始（SDK 简单）。生产用 LiteLLM 统一多提供商。

2. 基本调用

核心：发送提示（prompt），获取响应。以下是单行/简单示例。

2.1 OpenAI 示例

import os
from openai import OpenAI
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role":"user","content":"解释量子计算，简短点。"}]
)
print(response.choices[0].message.content)

输出示例：量子计算利用量子比特进行并行计算，能解决经典计算机难题，如因子分解。

2.2 Anthropic (Claude) 示例

import os
from anthropic import Anthropic
client = Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])
response = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=100,
    messages=[{"role":"user","content":"写一首关于 AI 的短诗。"}]
)
print(response.content[0].text)

2.3 Groq 示例

import os
from groq import Groq
client = Groq(api_key=os.environ["GROQ_API_KEY"])
response = client.chat.completions.create(
    model="llama3-70b-8192",
    messages=[{"role":"user","content":"Python 如何处理异常？"}]
)
print(response.choices[0].message.content)

2.4 Hugging Face Inference API 示例

import os
from huggingface_hub import InferenceClient
client = InferenceClient(token=os.environ["HF_TOKEN"])
response = client.text_generation(
    "meta-llama/Llama-3.2-3B-Instruct",
    "描述 Hugging Face 的作用。",
    max_tokens=50
)
print(response)

通用提示：用 try-except 处理错误（如限速、密钥无效）。例如 except Exception as e: print(e)。

3. 高级技巧

3.1 流式响应（Streaming）

实时输出（如聊天 App）。

stream = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role":"user","content":"讲个笑话。"}],
    stream=True
)
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

3.2 异步调用（Async，高并发）

用 asyncio 并发多个请求。

import asyncio
from openai import AsyncOpenAI

async_client = AsyncOpenAI(api_key=os.environ["OPENAI_API_KEY"])

async def async_call(prompt):
    response = await async_client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role":"user","content": prompt}]
    )
    return response.choices[0].message.content

async def main():
    tasks = [async_call("提示 1"), async_call("提示 2")]
    results = await asyncio.gather(*tasks)
    print(results)

asyncio.run(main())

3.3 工具调用（Function Calling）

让模型调用自定义函数（如计算器）。OpenAI/Anthropic 支持。

# OpenAI 工具示例
tools = [{"type":"function","function":{"name":"add_numbers","description":"加两个数","parameters":{"type":"object","properties":{"a":{"type":"number"},"b":{"type":"number"}},"required":["a","b"]}}}}]
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role":"user","content":"3+5=?"}],
    tools=tools
)
# 如果调用工具，执行函数
if response.choices[0].message.tool_calls:
    tool_call = response.choices[0].message.tool_calls[0]
    if tool_call.function.name == "add_numbers":
        args = json.loads(tool_call.function.arguments)
        result = args["a"] + args["b"]
        print(result)

3.4 多模态（图像/音频）

OpenAI GPT-4o 支持。

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role":"user","content":[{"type":"text","text":"描述这张图。"},{"type":"image_url","image_url":{"url":"https://example.com/image.jpg"}}]}])
print(response.choices[0].message.content)

4. 统一多提供商：LiteLLM SDK

LiteLLM 支持 100+ LLM，一键切换。

from litellm import completion
import os
os.environ["OPENAI_API_KEY"]="..."
os.environ["ANTHROPIC_API_KEY"]="..."

# 调用 OpenAI
response = completion(model="gpt-4o-mini", messages=[{"role":"user","content":"Hi!"}])

# 切换到 Claude
response = completion(model="claude-3-5-sonnet-20241022", messages=[{"role":"user","content":"Hi!"}])
print(response.choices[0].message.content)

优势：成本跟踪、负载均衡、日志。

5. 集成框架：LangChain（链式/代理开发）

构建复杂应用。

from langchain_groq import ChatGroq
from langchain.prompts import PromptTemplate
from langchain.chains import LLMChain

llm = ChatGroq(model="llama3-70b-8192", api_key=os.environ["GROQ_API_KEY"])
prompt = PromptTemplate(input_variables=["topic"], template="解释 {topic}。")
chain = LLMChain(llm=llm, prompt=prompt)
print(chain.run("区块链"))

6. 性能优化与最佳实践

• 常见错误：401（密钥错）、429（限速）、超时（加 timeout=30）。
• 本地运行：用 Ollama/Hugging Face Transformers 跑开源模型（无 API 费）。例如 from transformers import pipeline; pipe = pipeline("text-generation", model="meta-llama/Llama-3.2-3B")。

7. 完整项目实战：Streamlit 聊天机器人

用 Groq 构建 Web UI。

app.py：

import streamlit as st
from groq import Groq
import os

client = Groq(api_key=os.environ["GROQ_API_KEY"])
st.title("Groq 聊天机器人")

if "messages" not in st.session_state:
    st.session_state.messages = []

for message in st.session_state.messages:
    with st.chat_message(message["role"]):
        st.markdown(message["content"])

prompt = st.chat_input("说点什么？")
if prompt:
    st.session_state.messages.append({"role":"user","content": prompt})
    with st.chat_message("user"):
        st.markdown(prompt)
    with st.chat_message("assistant"):
        stream = client.chat.completions.create(
            model="mixtral-8x7b-32768",
            messages=[{"role": m["role"],"content": m["content"]} for m in st.session_state.messages],
            stream=True
        )
        response = st.write_stream([chunk.choices[0].delta.content or "" for chunk in stream])
        st.session_state.messages.append({"role":"assistant","content":"".join(response)})

运行：streamlit run app.py。访问 localhost:8501。

8. 进阶学习路线