SeineFish

A GitHub Webhook that Reviews GitHub Pull Requests

Overview

SeineFish is a GitHub webhook designed to automatically review pull requests in your GitHub repositories. It uses AI to analyze code changes, provide feedback, and summarize the review results.

Features:

Unlike traditional code review tools, SeineFish reviews the entire file content in pull requests and rates them as "GOOD", "NEEDS FURTHER TRIAGE", or "BAD". It also provides a summary of the review results to help developers quickly understand the quality of their pull requests.
SeineFish uses LangChain to select between OpenAI and WatsonX for the language model backend. This allows you to choose the best language model for your specific use case.
SeineFish handles pull_request, pull_request_review_comment, pull_request_review, and pull_request_review_thread events, providing comprehensive code review capabilities.
SeineFish is easy to set up and use. Simply create a GitHub webhook, configure the environment variables, and start the webhook server. You can then test the webhook locally and deploy it to your GitHub repository.
SeineFish is able to injest Git commit history and base on similar commits, provide feedback on the pull request.

Features

Handles pull_request, pull_request_review_comment, pull_request_review, and pull_request_review_thread events.
Uses LangChain to select between OpenAI and WatsonX for language model backend.
Reviews entire file content in pull requests.
Summarizes reviews and rates pull requests as "GOOD", "NEEDS FURTHER TRIAGE", or "BAD".

Requirements

Python 3.8 or higher
GitHub personal access token with appropriate permissions
OpenAI API key or WatsonX API key

Installation

Clone the repository:

git clone https://github.com/SeineAI/SeineFish.git
cd SeineFish

Create and activate a virtual environment:

python3 -m venv venv
source venv/bin/activate

Install the required dependencies:
```
pip install -r requirements.txt
```

Set up environment variables:

Create a .env file in the project root directory.
Add the following environment variables to the .env file:

Environment Variable	Value
GITHUB_TOKEN	<your_github_token>
OPENAI_API_KEY	<your_openai_api_key>
WATSONX_API_KEY	<your_watsonx_api_key>
WATSONX_PROJECT_ID	<your_watsonx_project_id>
REPO_OWNER	<your_github_repo_owner>
REPO_NAME	<your_github_repo_name>
REPO_PATH	<your_git_repo_path>
NGROK_WEB_ADDR	http://localhost:4041/api/tunnels
WEBHOOK_ENDPOINT	/github-webhook
WEBHOOK_PORT	8000

Run the webhook server locally for testing:
```
cd src; python main.py
```

Usage

Create the GitHub Webhook:
- Use the create_webhook.sh script in the deploy directory to create the webhook in your GitHub repository.
- Run the script and follow the prompts:
```
./deploy/create_webhook.sh
```
Test the webhook locally:
- Use the scripts in the test directory to send test payloads to the webhook endpoint.
- Test pull request events:
```
./test/test_webhook.sh
```
- Test pull request review comment events:
```
./test/test_pull_review.sh
```

For more detailed information, visit the repository.