Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/winstxnhdw/Wingman
Your offline AI pair programmer for VSCode on Windows.
https://github.com/winstxnhdw/Wingman
copilot ctransformers huggingface llm replit vscode-extension
Last synced: about 1 month ago
JSON representation
Your offline AI pair programmer for VSCode on Windows.
- Host: GitHub
- URL: https://github.com/winstxnhdw/Wingman
- Owner: winstxnhdw
- License: gpl-3.0
- Created: 2023-07-21T20:29:07.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-04-19T11:01:26.000Z (8 months ago)
- Last Synced: 2024-04-20T00:47:19.135Z (8 months ago)
- Topics: copilot, ctransformers, huggingface, llm, replit, vscode-extension
- Language: Python
- Homepage:
- Size: 611 KB
- Stars: 26
- Watchers: 4
- Forks: 2
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-AI-driven-development - Wingman - Your offline AI pair programmer for VSCode on Windows. (Uncategorized / Uncategorized)
README
# Wingman
[](https://github.com/PyCQA/pylint)
[](https://github.com/winstxnhdw/Wingman/actions/workflows/deploy.yml)
[](https://github.com/winstxnhdw/Wingman/actions/workflows/formatter.yml)
[](https://github.com/winstxnhdw/Wingman/actions/workflows/dependabot.yml)
[](https://github.com/winstxnhdw/Wingman/actions/workflows/github-code-scanning/codeql)
Monorepo for Wingman
> [!NOTE]\
> Did you know that the **Win** in **Win**gman stands for **Win**dows?
`Wingman` is a **fast CPU-first offline AI pair programmer** for [Visual Studio Code](https://code.visualstudio.com/) on Windows. `Wingman` comes with a self-contained [FastAPI](https://fastapi.tiangolo.com/) server that leverages the fast inference speeds of [ctransformers](https://github.com/marella/ctransformers) and a lightweight 4-bit quantised model to provide a seamless pair programming experience on low-end machines. You may find the used model [here](https://huggingface.co/winstxnhdw/Replit-v2-CodeInstruct-3B-ggml-4bit).
## Features
- No internet connection required
- No API keys required
- No GPU required
- Self-contained
- Fast
## Usage
> [!IMPORTANT]\
> Ensure that Visual Studio Code has already been installed.
You can download the portable executable [here](https://github.com/winstxnhdw/Wingman/releases/tag/latest).
## Development
If you would like to contribute to `Wingman`, here is some information to get you started.
### Requirements
- [Python 3.11](https://www.python.org/downloads/release/python-3110/)
- [Node.js 16](https://nodejs.org/ja/blog/release/v16.16.0)
- [Poetry](https://python-poetry.org/)
- [Yarn](https://yarnpkg.com/)
### Setup
Install all TypeScript dependencies.
```bash
yarn
```
Install all server dependencies.
```bash
poetry install
```
Download the model.
```bash
python scripts/download_model.py
```
### Build
Minify, bundle and package the extension.
```bash
yarn package
```
Human-readable bundle of your Node application. For debugging purposes.
```bash
yarn build test
```
Build the installer.
```bash
python scripts/build.py
```
### Test
Test the Visual Studio Code extension.
```bash
yarn test
```
Test the server.
```bash
pytest
```
## Credits
A special thank you to [Alyssa](https://github.com/alyssaxchua) for the amazing extension logo and banner!