AI Code Reviewer is a GitHub Action that leverages OpenAI's large language models to provide intelligent feedback and suggestions on your pull requests. This powerful tool helps improve code quality and saves developers time by automating the code review process.
- Reviews pull requests using OpenAI's large language models.
- Provides intelligent comments and suggestions for improving your code.
- Filters out files that match specified exclude patterns.
- Supports custom prompting to hone the analysis in on specific things, and avoid others.
- Easy to set up and integrate into your GitHub workflow.
-
To use this GitHub Action, you need an OpenAI API key. If you don't have one, sign up for an API key at OpenAI.
-
Add the OpenAI API key as a GitHub Secret in your repository with the name
OPENAI_API_KEY
. You can find more information about GitHub Secrets here. -
Create a
.github/workflows/main.yml
file in your repository and add the following content:
name: AI Code Reviewer
on:
pull_request:
types:
- opened
- synchronize
- ready_for_review
permissions: write-all
jobs:
review:
if: '! github.event.pull_request.draft' # Only run on non-draft pull requests. Must be coupled with the ready_for_review event above.
runs-on: ubuntu-latest
steps:
- name: Checkout Repo
uses: actions/checkout@v3
- name: AI Code Reviewer
uses: your-username/ai-codereviewer@main
with:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} # Keep this line verbatim. GitHub sets secrets.GITHUB_TOKEN automatically, it just needs to be plumbed into the action. See https://docs.github.com/en/actions/security-guides/automatic-token-authentication#about-the-github_token-secret for more information.
OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }} # Your OpenAI key. Create one at https://platform.openai.com/api-keys. Set it in your repository under Settings -> Secrets and variables -> Actions -> Repository Secrets.
OPENAI_API_MODEL: "gpt-4" # Optional: The OpenAI model (https://platform.openai.com/docs/models) to use. Defaults to "gpt-4". Prefer models with JSON Mode (https://platform.openai.com/docs/guides/text-generation/json-mode) support.
exclude: "yarn.lock, dist/**, **/*.json, **/*.md, **/*.yaml, **/*.xml" # Optional: File patterns, separated by commas, to exclude from analysis.
custom_prompts: | # Optional: Custom commands to add to the prompt used to analyze code. This is a multiline string, where each line is an individual command.
Do not worry about the verbosity of variable names, as long as they are somewhat descriptive.
Be sure to call out potential null pointer exceptions.
Be sure to call out concurrency issues and potential race conditions.
-
Replace
your-username
with your GitHub username or organization name where the AI Code Reviewer repository is located. If using the original, replace it withfreeedcom
. -
Customize the
exclude
input if you want to ignore certain file patterns from being reviewed. -
Customize or remove the
custom_prompts
input. -
Commit the changes to your repository, and AI Code Reviewer will start working on your future pull requests.
The AI Code Reviewer GitHub Action retrieves the pull request diff, filters out excluded files, and sends code chunks to the OpenAI API. It then generates review comments based on the AI's response and adds them to the pull request.
- The
gpt-4
model is powerful, but relatively expensive and does not support JSON Mode at the time of writing. Consider using one of the turbo models — such asgpt-4-turbo-preview
orgpt-3.5-turbo
— to find the right balance between price and performance. - Use
custom_prompts
to hone the action's analysis in on things you care about, and to avoid things you don't. There is no point, for example, in having the action repeat what your static analysis tooling does for you in real-time while coding. - Use an
if: '! github.event.pull_request.draft'
job conditional andready_for_review
trigger to control costs and ensure that you are only running the action on pull requests that are truly ready to be reviewed.
Contributions are welcome! Please feel free to submit issues or pull requests to improve the AI Code Reviewer GitHub Action.
Let the maintainer generate the final package (yarn build
& yarn package
).
This project is licensed under the MIT License. See the LICENSE file for more information.