revu is a comprehensive command-line tool designed to streamline the code review process. Leveraging the advanced capabilities of GPT-4 and the GitHub API, it can analyze and provide insightful reviews on pull requests, local changes, and individual files. Additionally, revu offers an intuitive commit message generator that uses local diffs and commit history to propose appropriate commit messages. Its flexible nature aims to cover various aspects of code review, offering an efficient toolset for developers.
Getting Started
Prerequisites
- You'll need to have Node.js and npm installed on your machine.
- An OpenAI API key for using GPT-4 and a GitHub token for accessing the GitHub API.
Switching to GPT-4 Model
revu is initially set to use the GPT-3.5-turbo model. If you wish to switch to GPT-4, you can do so by modifying your revu.json
config file:
- Run the config command if you haven't done so already. This will generate the
revu.json
config file:
revu config
- Locate your
revu.json
config file. By default, it is saved in the.revu
directory in your home directory (~/.revu
). - Find the
llm
section and then theopenai
subsection within it. - Change the value of
openaiModel
fromgpt-3.5-turbo
togpt-4
. - Save and close your
revu.json
config file.
Remember that using GPT-4 may result in increased API costs. Please refer to OpenAI's pricing for more information.
Installation
You can install revu globally using npm by running the following command:
npm i -g revu-cli
Alternatively, you can clone the repository and install the dependencies locally:
- Clone the repository:
git clone https://github.com/phmz/revu-cli.git
- Navigate to the project directory:
cd revu-cli
- Install dependencies:
npm install
- Build the project:
npm run build
Usage
Before using revu, you need to set up the configuration with your OpenAI API key and GitHub token. You can do this with the following command:
revu config
This will prompt you to enter your OpenAI API key and GitHub token.
For a comprehensive list of all available commands and options in revu, run the help command:
revu help
This will display a list of all the available commands, their descriptions, and options you can use with revu.
Environment Variables
revu can also be configured using environment variables. If an environment variable is not provided, revu will use the default value.
Here are the available environment variables:
-
GIT_MAX_COMMIT_HISTORY
: Maximum number of commit history entries to fetch (default: 10). -
GIT_IGNORE_PATTERNS
: A comma-separated list of regular expression patterns of files to ignore (default: []). -
GITHUB_API_URL
: Custom URL for the GitHub API (default: https://api.github.com). -
GITHUB_TOKEN
: GitHub personal access token. -
OPENAI_API_URL
: Custom URL for the OpenAI API (default: https://api.openai.com). -
OPENAI_API_KEY
: OpenAI API key for accessing the OpenAI API. -
OPENAI_MODEL
: OpenAI model to use (default: gpt-3.5-turbo). -
OPENAI_TEMPERATURE
: Temperature parameter for OpenAI model (default: 0).
Local Code Review
revu can analyze local changes in two ways:
1. Analyzing all local changes
If you want to analyze your local changes, navigate to the root directory of your local Git repository and run the following command:
revu local
revu will then analyze your local changes and provide you with a review.
2. Analyzing a specific file
If you want to analyze a specific file in your local directory, navigate to the root directory of your local Git repository and run the following command:
revu local --directory <directory> --filename <filename>
Replace <directory>
with the relative path of the directory to search and <filename>
with the name of the file to review.
Generate Commit Message
revu can propose commit messages based on local diffs and commit history. To use this feature, run the following command:
revu commit
revu will prompt you to select the files you wish to commit. Once the files are selected, revu fetches the commit history and proposes a commit message. If you agree with the suggested commit message, you can proceed to commit your changes right away. If there are unselected files left, revu will ask you if you wish to continue the commit process.
Pull Request Review
If you want to analyze a pull request, run the following command:
revu pr <repository> <pull_request_number>
Replace <repository>
with the repository to review in the format owner/repository
, and <pull_request_number>
with the number of the pull request to review. For example:
revu pr phmz/revu 42
revu will then fetch the pull request details, analyze the changes, and provide you with a review.
Ignoring Files
The revu CLI tool allows you to ignore certain files during your review process by using regular expression patterns. You can define these patterns either through a configuration file or via an environment variable. The CLI tool will ignore files that match any of the provided patterns.
Via Configuration File
You can define an array of ignorePatterns
under the git
section in your revu.json
configuration file, like so:
{
"git": {
"ignorePatterns": [".*lock.*", "another_pattern", "..."]
}
}
Via Environment Variable
Alternatively, you can use the GIT_IGNORE_PATTERNS
environment variable to define a comma-separated list of regular expression patterns:
export GIT_IGNORE_PATTERNS=.*lock.*,another_pattern,...
Development
revu is built with TypeScript. Contributions are welcome!
Code style
This project uses ESLint for linting.
You can run the linter with:
npm run lint