yasyf/gpt-do

GPT-powered bash commands.

gpt-do

This is a handy-dandy CLI for when you don't know wtf to do.

Instead of furiously grepping through man pages, simply use do (or ddo if on bash/zsh), and have GPT-3 do all the magic for you.

Check out the blog post here.

Demo

Click to play:

Installation

We recommend using pipx:

$ pipx install gpt-do
$ which do
~/.local/bin/do

However you can also use brew:

$ brew install yasyf/do/do
$ which do
/opt/homebrew/bin/do

Or pip:

$ pip install gpt-do
$ which do
~/.asdf/installs/python/3.11.0/bin/do

Usage

n.b. If you're on bash or zsh, do is a reserved keyword, so you'll have to use the alias ddo.

n.b. The default model used is GPT-3. Please ensure you have sufficient credits in your OpenAI account to use it.

$ export OPENAI_API_KEY=xxx # stick this in your bash_profile
$ do amend the message of my last commit to "It works!"
This command will amend the message of the last commit to 'It works!'.
git commit --amend -m 'It works!'
Do you want to continue? [y/N]: y
[main 3e6a2f6] It works!!
 Date: Thu Dec 22 01:15:40 2022 -0800
 5 files changed, 1088 insertions(+)
 create mode 100644 .gitignore
 create mode 100644 .gitmodules
 create mode 100644 README.md
 create mode 100644 poetry.lock
 create mode 100644 pyproject.toml

gpt-do's Issues

Quota error

I'm receiving openai.error.RateLimitError: You exceeded your current quota, please check your plan and billing details.

Either this is a block from openai on this prompt or something else, but I didn't exceeded the quota (only have been using chatgpt for one or to questions a day) and I do have credits on a business account.

davinci deprecated

    self._interpret_response_line(
  File "/Users/cellis/Library/Application Support/pipx/venvs/gpt-do/lib/python3.12/site-packages/openai/api_requestor.py", line 429, in _interpret_response_line
    raise self.handle_error_response(
openai.error.InvalidRequestError: The model `text-davinci-003` has been deprecated, learn more here: https://platform.openai.com/docs/deprecations```

Command doesn't run after "Do you want to continue? [y/N]: y"

On a remote machine, entering y does not execute the command. I've successfully used the tool locally.
Here is what my terminal looks like:

(base) √√ ~ % ddo echo hello world
This command will print 'hello world' to the terminal.
echo 'hello world'
Do you want to continue? [y/N]: y
(base) √√ ~ % 

I use kitty, zsh, there is a conda env activated, it's on Ubuntu 20.04, the tool was installed with pip. Not sure which additional details to provide here.

FR: confirm with one keypress

Hi, thanks for the nice tool.

I was wondering whether you could make it that the confirmation prompt accepts just on keypress, essentially only "y" or "n" instead of "y+return" or "n+return".

With shell script, you can get this behavior with read -k, but unfortunately I do not know how to do this with python (otherwise I would have made a PR myself).

Prints text in zsh before command

After pressing y to accept a command in zsh, several errors are printed to the console:

» ddo echo hello                                                                                                                                                                                   3:27PM
This command will print the string 'hello' to the terminal.
echo hello
Do you want to continue? [y/N]: y
/home/user/.zsh/oh-my-zsh/oh-my-zsh.sh:46: ZSH_CACHE_DIR: parameter not set
/home/user/.zsh/config/zshrc.d/044_functions-vcsh.zsh:3: no match
/home/user/.zsh/plugins/zsh-autosuggestions/zsh-autosuggestions.zsh:42: unmatched "
/home/user/.zsh/plugins/hhighlighter/h.sh:1: unmatched "
/home/user/.zsh/config/zshrc.d/061_plugins-hstr.zsh:4: command not found: bindkey
compinit:170: unmatched '
/home/user/.bun/_bun:630: command not found: compdef
/home/user/.zsh/oh-my-zsh/oh-my-zsh.sh:46: ZSH_CACHE_DIR: parameter not set
/home/user/.zsh/config/zshrc.d/044_functions-vcsh.zsh:3: no match
/home/user/.zsh/plugins/zsh-autosuggestions/zsh-autosuggestions.zsh:42: unmatched "
/home/user/.zsh/plugins/hhighlighter/h.sh:1: unmatched "
compinit:170: unmatched '
/home/user/.bun/_bun:630: command not found: compdef
hello

When using ChatGPT: ModuleNotFoundError: No module named 'playwright'

When executing the following command with the chatgpt model: ddo whats the highest CPU process --model=chatgpt

I receive the following error:

Traceback (most recent call last):
  File "/opt/homebrew/bin/ddo", line 8, in <module>
    sys.exit(do())
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/click/core.py", line 1130, in __call__
    return self.main(*args, **kwargs)
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/click/core.py", line 1055, in main
    rv = self.invoke(ctx)
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/click/core.py", line 1404, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/click/core.py", line 760, in invoke
    return __callback(*args, **kwargs)
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/gpt_do/cli.py", line 49, in do
    do = get_doer(model)(debug=debug)
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/gpt_do/cli.py", line 8, in get_doer
    from gpt_do.doers.pywright_doer import PywrightDoer
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/gpt_do/doers/pywright_doer.py", line 2, in <module>
    from gpt_do.vendor.chatgpt_wrapper.chatgpt_wrapper import ChatGPT
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/gpt_do/vendor/chatgpt_wrapper/chatgpt_wrapper/__init__.py", line 1, in <module>
    from .chatgpt import ChatGPT
  File "/opt/homebrew/Cellar/do/0.1.16/libexec/lib/python3.10/site-packages/gpt_do/vendor/chatgpt_wrapper/chatgpt_wrapper/chatgpt.py", line 19, in <module>
    from playwright.sync_api import sync_playwright
ModuleNotFoundError: No module named 'playwright'

I assume the installation scripts of gpt-do is missing playwright in the requirements.

ModuleNotFoundError: No module named 'gpt_do.vendor'

I have installed gpt-do with pipx and when using --model=chatgpt the scripts raises an error

$ ddo echo hello to file hello.txt --model=chatgpt                                                                                    
Traceback (most recent call last):
  File "/home/tester/.local/bin/ddo", line 8, in <module>
    sys.exit(do())
  File "/home/tester/.local/pipx/venvs/gpt-do/lib/python3.10/site-packages/click/core.py", line 1130, in __call__
    return self.main(*args, **kwargs)
  File "/home/tester/.local/pipx/venvs/gpt-do/lib/python3.10/site-packages/click/core.py", line 1055, in main
    rv = self.invoke(ctx)
  File "/home/tester/.local/pipx/venvs/gpt-do/lib/python3.10/site-packages/click/core.py", line 1404, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/tester/.local/pipx/venvs/gpt-do/lib/python3.10/site-packages/click/core.py", line 760, in invoke
    return __callback(*args, **kwargs)
  File "/home/tester/.local/pipx/venvs/gpt-do/lib/python3.10/site-packages/gpt_do/cli.py", line 37, in do
    do = get_doer(model)(debug=debug)
  File "/home/tester/.local/pipx/venvs/gpt-do/lib/python3.10/site-packages/gpt_do/cli.py", line 6, in get_doer
    from gpt_do.doers.pywright_doer import PywrightDoer
  File "/home/tester/.local/pipx/venvs/gpt-do/lib/python3.10/site-packages/gpt_do/doers/pywright_doer.py", line 2, in <module>
    from gpt_do.vendor.chatgpt_wrapper.chatgpt_wrapper import ChatGPT
ModuleNotFoundError: No module named 'gpt_do.vendor'

The vendor directory is actually missing (I guess the git submodule is not populated by pipx)