编程接入chatGPT

接入chatGPT三种方式

通过sdk接入后端程序

• 命令行下载openai sdk

pip install openai

• 在python中调用sdk的demo

import openai
openai.api_key = '在网页创建的API key'
prompt = '初学python有哪些建议'
msg = {'role':'user', 'content': prompt}
result = openai.ChatCompletion.create(model = 'gpt-3.5-turbo',messages=[msg])
answer = result.choices[0].messages['content']
print(answer)

通过命令行的方式

• 命令行

openai -k 你的APIkey api chat_completions.create -m gpt-3.5-turbo -g user Python有什么优点?

通过网络请求API

• python中网络请求openapi接口的demo

import requests
u = "https://api.openai.com/v1/chat/completions"
api_key = '在网页上创建API key'
h = {
    'Content-Type': 'application/json',
    'Authorization': 'Bearer'+api_key
}
prompt = '学python有哪些阶段'
msg = {'role': 'user', 'content':prompt}
d = {
    'model': 'gpt-3.5-turbo',
    'messages': [msg]
}
r = requests.post(url=u,headers=h,json=d)
r = r.json()
print(r['choices'][0]['message']['content'])