Skip to content

Latest commit

 

History

History
65 lines (42 loc) · 2.88 KB

setup.md

File metadata and controls

65 lines (42 loc) · 2.88 KB

Setup Instructions 🤖

Shippie 🚢 is a NodeJS package that uses LLMs to provide feedback on code. It is designed to be used in a CI environment to provide feedback on pull requests.

Prerequisites

  • Node 18+ or Bun 1.0+
  • Git
  • Github or Gitlab CLI (optional for configure tool)

Easy Setup in CI 🚀

In the root of your git repository run:

npx shippie configure --platform=github

The setup script will ask for your OpenAI API key. You can:

  1. Leave it blank - Uses GitHub Models (free) with the built-in GITHUB_TOKEN to access GitHub's AI models at https://models.github.ai/inference
  2. Provide an OpenAI API key - Uses OpenAI's API and automatically adds the OPENAI_API_KEY secret to your repo

More info on GitHub Actions secrets can be found here.

See templates for the example yaml files. You can copy and paste them to perform a manual setup or have a look at the action configuration options.

Package Commands

  • npx shippie review - Runs the code review on the staged files.
  • npx shippie configure - Runs a setup tool to configure the application.

Configure Options

  • Setup Target - The platform you are using eg. github, gitlab or azure devops.

Review Options

Shippie supports a bunch of setup options. This is a work in progress so check out the code here for the latest options.

  • Review Language - The language you want to review the code in.
  • Platform - The platform you are using eg. github, gitlab, azure devops, local.
  • Model String - The model you want to use eg. openai:gpt-4o, azure:gpt-4o, anthropic:claude-3-5-sonnet-20240620. When using GitHub platform with GitHub Models, use openai:gpt-4o-mini with baseUrl=https://models.github.ai/inference.
  • (optional) Max Steps - The maximum number of steps the bot will take. defaults to 25.
  • (optional) Base URL - The base URL for the AI provider. Use https://models.github.ai/inference for GitHub Models, or change this to use OpenAI compatible providers like DeepSeek or local models with LM Studio or Ollama.
  • (optional) Ignore - A list of globs to ignore when reviewing the code. Defaults to dist/**, node_modules/**, **/*.d.ts, **/*.lock, **/package-lock.json.
  • (optional) Telemetry - Toggle anonymous telemetry. Defaults to True
  • (optional) Debug - Toggle debug logging. Defaults to False.

Run npx shippie --help to see all the options available.

Local Usage 🌈

Shippie 🚢 also works locally to review files staged for commit. Just add some files to the staging area.

Export your OPENAI_API_KEY to the shell

export OPENAI_API_KEY=<your-api-key>

and run:

npx shippie review