bootstrap-gpt

Guide GPT to help you efficiently complete various tasks, write GPT prompts programs, and automate GPT prompts, support web api

cn

download

set api key

.env

api_key=sk-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
api_base=https://chatgpt-api.shn.hk/v1/
server_port=8080

run

cd bootstrap-gpt

pip install openai && pip install beautifulsoup4 && pip install jsonpath && pip install jinja2

python3 ./main.py

http server

> python3 ./main.py

Welcome to Bootstrap GPT
sample > use server

> server

HTTP Server running on port 8080

api

post /v1/chat/completions
{"messages":[{"role":"user","content":"v2-hot"}],"model":"gpt-3.5-turbo"}

res

{"choices": [{"message": {"role": "assistant", "content": "hello"}]}

coordination https://github.com/weekend-project-space/chatgpt-online Better eating effect

bootstrap simple

{
    "author": "ruidong",
    "date": "2023-04-17",
    "version": "0.1",
    "description": "code boot",
    "boot": [{
            "w": "Hello, this is a code template. Please select template 1. simple 2. demo",
            "r":"r0",
            "b": {
                "1": 1,
                "2": 2
            }
        },
        {
            "w": "Please enter the language type"
            "r":"r1",
        },
        {
            "w": "Please enter the requirement",
            "r":"r2",
            "p": "chat:Write a {{r2}} program for {{r1}}"
        }
    ]
}

• w: write supports template variables

• r: read

• b: Branch branch supports object field eq jump and Array sequential execution or no further execution

• p: prompt supports template variable micro instructions for guiding chatgt [chat:] to search for plain text or crawler [spider:] output, which can expand more micro instruction prompt guidance

r2 r1 is a variable with a name of type and an index of r: reader w: writer p: prompts m: msg

Generate Rule Reference interpreter.py