pyChatGPTLoop 0.1.10

Added backtracking chat on the basis of [terry3041/pyChatGPT]

pyChatGPTLoop

Added backtracking chat on the basis of terry3041/pyChatGPT

Extra Features

• Implement asynchronous use
• can go back to a conversational moment in the past
• increase the function of initializing personality
• set local chromedriver
• optimized the output, removing two newlines
• added reply failure handling
• added GPT-ToolBoxGPT-ToolBox

backtracking chat

initializing personality

Getting Started

This library is using only the undetected_chromedriver to bypass Cloudflare's anti-bot protection. Please make sure you have Google Chrome / Chromium before using this wrapper.

Installation

git clone https://github.com/nek0us/pyChatGPTLoop.git
pip install .

# or

pip install pyChatGPTLoop --upgrade

Usage

Obtaining session_token

1. Go to https://chat.openai.com/chat and open the developer tools by F12.
2. Find the __Secure-next-auth.session-token cookie in Application > Storage > Cookies > https://chat.openai.com.
3. Copy the value in the Cookie Value field.

Interactive mode

python3 -m pyChatGPTLoop

Import as a module

from pyChatGPTLoop.pyChatGPTLoop import ChatGPT
import asyncio
session_token = 'abc123'  # `__Secure-next-auth.session-token` cookie from https://chat.openai.com/chat
api = ChatGPT(session_token)  # auth with session token
api = ChatGPT(session_token, conversation_id='some-random-uuid')  # specify conversation id
api = ChatGPT(session_token, proxy='https://proxy.example.com:8080')  # specify proxy
api = ChatGPT(session_token, chrome_args=['--window-size=1920,768'])  # specify chrome args
api = ChatGPT(session_token, moderation=False)  # disable moderation
api = ChatGPT(session_token, verbose=True)  # verbose mode (print debug messages)

api = ChatGPT(session_token, driver_path="C:\\yourdriverpath\\chromedriver.exe") # set local chromedriver if you don't wish it update in windows
api = ChatGPT(session_token, driver_path="/yourpath/chromedriver") # set local chromedriver if you don't wish it update in linux
api = ChatGPT(session_token, personality_definition=your List) # the words list of init personality
api = ChatGPT(session_token, toolbox=True) # use GPT-ToolBox

# auth with google login
api = ChatGPT(auth_type='google', email='[e-mail]', password='password')
# auth with microsoft login
api = ChatGPT(auth_type='microsoft', email='[e-mail]', password='password')
# auth with openai login (captcha solving using speech-to-text engine)
api = ChatGPT(auth_type='openai', email='[e-mail]', password='password')
# auth with openai login (manual captcha solving)
api = ChatGPT(
    auth_type='openai', captcha_solver=None,
    email='[e-mail]', password='password'
)
# auth with openai login (2captcha for captcha solving)
api = ChatGPT(
    auth_type='openai', captcha_solver='2captcha', solver_apikey='abc',
    email='[e-mail]', password='password'
)
# reuse cookies generated by successful login before login,
# if `login_cookies_path` does not exist, it will process logining  with `auth_type`, and save cookies to `login_cookies_path`
# only works when `auth_type` is `openai` or `google`
api = ChatGPT(auth_type='openai', email='[e-mail]', password='password',
    login_cookies_path='your_cookies_path',
)


api.reset_conversation()  # reset the conversation
api.clear_conversations()  # clear all conversations

# send message 
resp = asyncio(api.async_send_message('Hello, world!'))
print(resp['message'])

# if in an async function
resp = await api.async_send_message('Hello, world!',msg_type="loop")
print(resp['message'])

# refresh the chat page
await api.refresh_chat_page()  

# return the loop_text conversation
loop_text = "some words"
resp = await api.async_send_message(loop_text,msg_type="loop")

#You can pass the conversation id to the right
resp = await api.async_send_message('Hello, world!',conversation_id,msg_type="msg")

personality_definition = [{"content":r'Now you are going to pretend to be a math teacher called "nothing" to help me with my math',"AI_verify":True}]
await api.init_personality(True,"",personality_definition)

#or 

resp = await api.async_send_message(personality_definition,conversation_id,msg_type="init")

#update session_toekn when chrome running
resp = await api.async_send_message(new_cookie,conversation_id,msg_type="update_cookie")

#all of resp:

resp: dict == {"status":bool,any}

Frequently Asked Questions

How do i initializing personality?

# An example of initializing the vocabulary format, the vocabulary content is not representative
personality_definition = [
        {
            "content":r'Now you are going to pretend to be a math teacher called "nothing" to help me with my math',
            "AI_verify":True
            },
        {
            "content":r"You will be very strict in pointing out my mistakes",
            "AI_verify":False
            }
    ]

new_conversation = True
await api.backtrack_chat(new_conversation,"",personality_definition)
# new_conversation : Whether to open a new session for personality initialization, the default is true
# personality_definition: Personality initialization phrase is a list, a single element is a dict, 
# content is the content of the phrase, and AI_verify is a successful detection of personality
# returns a boolean result

How do i go back to the conversation i had?

# You said some words to chatGPT
loop_text = 'some words'

# Go back to the dialogue where these words last appeared
await api.backtrack_chat(loop_text)
#You can pass the conversation id to the right of loop_text
#return an boolen result

# if you use an async method
resp = await api.async_send_message(loop_text,msg_type="loop")

How do I get it to work on headless linux server?

# install chromium & X virtual framebuffer
sudo apt install chromium-browser xvfb

# start your script
python3 your_script.py

How do I get it to work on Google Colab?

It is normal for the seession to be crashed when installing dependencies. Just ignore the error and run your script.

# install dependencies
!apt install chromium-browser xvfb
!pip install -U selenium_profiles pyChatGPT

# install chromedriver
from selenium_profiles.utils.installer import install_chromedriver
install_chromedriver()

# start your script as normal
!python3 -m pyChatGPTLoop

Insipration

This project is inspired by

• pyChatGPT
• ChatGPT
• chatgpt-api

Disclaimer

This project is not affiliated with OpenAI in any way. Use at your own risk. I am not responsible for any damage caused by this project. Please read the OpenAI Terms of Service before using this project.

License

This project is licensed under the GPLv3 License - see the LICENSE file for details.