ChatCLI is a command-line tool that uses the OpenAI GPT-3.5/GPT-4 API to provide a chat bot interface that can answer your questions. It provides a simple and intuitive way to interact with GPT-3 by asking questions and getting relevant answers.
This command is in early development and the interface and log file format are unstable and changing rapidly. I'd love to get feedback on what would make this tool more useful.
WARNING: This version supports EVALUATE blocks that allow GPT to evaluate code. This should only work if the pyeval plugin is enabled, either using "--plugin pyeval" or "--personality pyeval". There is no sandboxing, so GPT can run any code it likes if you enable this!
The following steps will guide you through the installation process:
- Clone the repository with the following command:
git clone https://github.com/cthulahoops/chatcli.git
- Navigate to the project directory:
cd chatcli
- Install Poetry (if you haven't already) by following the official Poetry installation instructions: https://python-poetry.org/docs/#installation
These instructions will guide you through the process of installing Poetry on your machine. Once installed, you can easily manage dependencies and environments for your Python projects using Poetry.
- Install ChatCLI using Poetry:
poetry shell
poetry install
Alternatively, you can install the dependencies using pip with:
pip install -r requirements.txt
The requirements.txt
may be less likely to be up to date!
-
Set your OpenAI API key in the environment variable
OPENAI_API_SECRET_KEY
. -
Run chatcli
chatcli
To ask a question, just run chatcli:
chatcli
This will start a conversation with the chat bot, which will prompt you for a question. You can also include a text file as context for your question by using the -f
or --file
option:
chatcli --file myfile.txt
You can also specify the personality that the chat bot should use with the -p
or --personality
option:
chatcli --personality concise
To continue a previous conversation, use the chat
command with the --continue
option:
chatcli --continue
To show a previous conversation, use the show
:
chatcli show
To list all the conversations that have been logged, use the log
command:
chatcli log
You can tag a conversation using the tag
command:
chatcli tag mytag
To list all the tags that have been used, use the tags
command:
chatcli tags
You can filter conversations by tag using the -t
or --tag
option:
chatcli log --tag mytag
You can remove a tag from a conversation using the untag
command:
chatcli untag mytag
To display the number of tokens used and the token cost, use the usage
command:
chatcli usage
chatcli --quick --file chatcli.py --personality code
>> Generate a README.md for this project.
- Make some changes to your code and stage them for commit:
git add -p
- Use
git diff
to see the changes you've made and pipe them to ChatGPT'schatcli.py
script to generate a commit message:
git diff --cached | chatcli -p commit
- Make a commit with the generated message:
git commit -m "$(chatcli show)"
This will use the show
command to display the last message generated by the chat bot, which is then used as the commit message.
That's it! You've successfully used ChatGPT to generate a commit message based on the changes you've made to your code.
If you wish to contribute to this project, please fork the repository and submit a pull request.