Generating shell commands and code using natural language models (OpenAI ChatGPT).
Project description
Shell Craft
Generating shell commands and code using natural language models (OpenAI ChatGPT).
Install using Pip
pip install openai-shell-craft
Build from Docker and Run
docker build . --tag shell_craft --target shell_craft
docker run shell_craft --help
Command Line Interface
All command line usage follows the following format:
shell-craft <request>
Quick Start
shell-craft "tell me how long the system has been up for"
Returns:
uptime -p
Prompt Types
Shell Craft supports many different --prompt
options. We have organized them into categories here only to aid in reference usage.
Category | Name | Argument Value |
---|---|---|
Language | Bash | bash |
Language | PowerShell | powershell |
Language | Python | python |
Language | C | c |
Language | C# | c_sharp |
Language | JavaScript | javascript |
Language | Java | java |
Language | Go | go |
Template | Bug Report | bug_report |
Template | Feature Request | feature_request |
When --prompt
is not specified, Shell Craft will use bash
unless PowerShell is used to call Shell Craft, then it uses powershell
.
Help Prompt
For additional help you can run:
shell-craft --help
Bash Example
Executing:
shell-craft --prompt bash "find all swp files and delete them if theyre not in use"
Returns a Bash command:
find . -name '*.swp' -type f ! -exec fuser -s {} \; -delete
Powershell Example
Executing:
shell-craft --prompt powershell "remove all files older than 30 days"
Returns a PowerShell command:
Get-ChildItem -Path "C:\your\path" -Recurse | Where-Object { $_.LastWriteTime -lt (Get-Date).AddDays(-30)} | Remove-Item -Force
Refactoring existing code
cat my_code.py | python -m shell_craft --prompt python --refactor
Documenting existing code
cat my_code.py | python -m shell_craft --prompt python --document
Generating tests for existing code
cat my_code.py | python -m shell_craft --prompt python --test
Feature Request Example
Executing:
shell_craft --prompt feature_request "add an icon that shows the total test coverage"
Returns Markdown:
---
name: Add Total Test Coverage Icon
about: Add an icon that shows the total test coverage
labels: Enhancement
---
**Is your feature request related to a problem? Please describe.**
There is currently no easy way to visualize the total test coverage for a project.
**Describe the solution you'd like**
I would like to have an icon on the project dashboard that, when clicked, displays the total test coverage for the project. This will allow developers to easily monitor and improve the overall test coverage for the project.
**Describe alternatives you've considered**
As an alternative, I have considered adding the total test coverage to the project README file. However, this solution would not be as easily accessible as an icon on the project dashboard.
Language Models
Support for different OpenAI models.
Currently recommended --model
options:
gpt-3.5-turbo
(default)gpt-4
gpt-4-32k
code-davinci-002
text-davinci-003
If --model
is not specified, it will use gpt-3.5-turbo
by default.
Python API
The following example shows how to use Shell Craft inside of your own Python code. The example loads your api key from an environment variable, uses the Bash prompt, and gets input from the user.
import os
from shell_craft import Service
from shell_craft.prompts import BASH_PROMPT
Service(
api_key = os.environ.get("OPENAI_API_KEY"),
prompt = BASH_PROMPT,
).query(
message = input("Enter your request: ")
)
API Key
Shell Craft uses the OpenAI API to generate the shell commands, requiring an API key to give it access to an account with OpenAI.
Warning! API keys are sensitive information and should be kept confidential. Don't share your API key with anyone who shouldn't have access to it. Failure to do so may result in charges on your OpenAI account. Shell Craft never transmits this key to any service other than the official OpenAI API. Your API key is your responsibility.
To obtain an API key from the OpenAI platform, follow these steps:
- Go to https://platform.openai.com/account/api-keys in your web browser.
- Log in to your OpenAI account or create a new account if you don't have one.
- Click the "Create new secret key" button.
- Copy the key that is generated for you.
Once you have your API key, you must pass it to Shell Craft via one of the following methods.
Option 1 (Recommended) - Configuration file:
shell-craft looks for a configuration file in the following places, in order:
- the current directory, named config.json
- the file named by $SHELLCRAFT_CONFIG
- $XDG_CONFIG_HOME/shell-craft/config.json
- $HOME/.config/shell-craft/config.json
The contents are simple json:
{
"openai_api_key": "<your secret key>"
}
Now, Shell Craft will know your API key in order to run future prompts without having to specify it or ensure it exists in the environmental variables.
Option 2 - Command Line Argument:
shell-craft "list all files in directory" --api-key API_KEY
Where API_KEY
is your OpenAI API key.
Option 3 - Environment Variable:
export OPENAI_API_KEY=<your secret key>
shell-craft "list all files in directory"
Where <your secret key>
is your OpenAI API key.
License
This project is licensed under the MIT License - see the LICENSE file for details.
Acknowledgments
Acknowledgments contain a few of the people who have helped the project along the way.
- OpenAI for their language models and API.
- Joel Fourhman for the initial idea, the challenge, and the project name.
- Chase Montgomery for testing, feedback, and estimated time savings vs manual implementation.
- Bryan Gonzalez for powershell feedback and merge reviews.
- Aaron Croissette for business use case feedback, documentation, and a lovely PyPi icon on the readme.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for openai_shell_craft-0.5.5-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 31300ee96bc73c3d87b81124ac647909ccca25b9b7e639c5137fb4d657353a50 |
|
MD5 | b6643ebff614bb84bdeac4beb6539463 |
|
BLAKE2b-256 | d966f61ea2db41ff9ec26a4c650985e174ef9ebf0db021b2742323665a13a93c |