Making pull requests less painful with an AI agent

CodiumAI PR-Agent is an open-source tool aiming to help developers review pull requests faster and more efficiently. It automatically analyzes the pull request and can provide several types of PR feedback:

Auto-Description: Automatically generating PR description - title, type, summary, code walkthrough and labels.
Auto Review: Adjustable feedback about the PR main theme, type, relevant tests, security issues, score, and various suggestions for the PR content.
Question Answering: Answering free-text questions about the PR.
Code Suggestions: Committable code suggestions for improving the PR.
Update Changelog: Automatically updating the CHANGELOG.md file with the PR changes.

Example results:

/describe:

/review:

/reflect_and_review:

/ask:

/improve:

• Overview
• Try it now
• Installation
• Configuration
• How it works
• Why use PR-Agent
• Roadmap
• Similar projects

Overview

PR-Agent offers extensive pull request functionalities across various git providers:

		GitHub	Gitlab	Bitbucket	CodeCommit
TOOLS	Review	✅	✅	✅	✅
	⮑ Inline review	✅	✅
	Ask	✅	✅	✅
	Auto-Description	✅	✅		✅
	Improve Code	✅	✅
	⮑ Extended	✅	✅
	Reflect and Review	✅
	Update CHANGELOG.md	✅
USAGE	CLI	✅	✅	✅	✅
	App / webhook	✅	✅
	Tagging bot	✅
	Actions	✅
CORE	PR compression	✅	✅	✅
	Repo language prioritization	✅	✅	✅
	Adaptive and token-aware file patch fitting	✅	✅	✅
	Multiple models support	✅	✅	✅
	Incremental PR Review	✅

Examples for invoking the different tools via the CLI:

• Review: python cli.py --pr_url=<pr_url> review
• Describe: python cli.py --pr_url=<pr_url> describe
• Improve: python cli.py --pr_url=<pr_url> improve
• Ask: python cli.py --pr_url=<pr_url> ask "Write me a poem about this PR"
• Reflect: python cli.py --pr_url=<pr_url> reflect
• Update Changelog: python cli.py --pr_url=<pr_url> update_changelog

"<pr_url>" is the url of the relevant PR (for example: Codium-ai#50).

In the configuration file you can select your git provider (GitHub, Gitlab, Bitbucket), and further configure the different tools. Options that are available in the configuration file can be specified at run time when calling actions. Two examples:

• /review --pr_reviewer.extra_instructions="focus on the file: ..."
• /describe --pr_description.add_original_user_description=false -pr_description.extra_instructions="make sure to mention: ..."

Try it now

Try GPT-4 powered PR-Agent on your public GitHub repository for free. Just mention @CodiumAI-Agent and add the desired command in any PR comment! The agent will generate a response based on your command.

To set up your own PR-Agent, see the Installation section

---

Installation

To get started with PR-Agent quickly, you first need to acquire two tokens:

1. An OpenAI key from here, with access to GPT-4.
2. A GitHub personal access token (classic) with the repo scope.

There are several ways to use PR-Agent:

• Method 1: Use Docker image (no installation required)
• Method 2: Run as a GitHub Action
• Method 3: Run from source
• Method 4: Run as a polling server
  • Request reviews by tagging your GitHub user on a PR
• Method 5: Run as a GitHub App
  • Allowing you to automate the review process on your private or public repositories
• Method 6: Deploy as a Lambda Function
• Method 7: AWS CodeCommit

How it works

The following diagram illustrates PR-Agent tools and their flow:

Check out the PR Compression strategy page for more details on how we convert a code diff to a manageable LLM prompt

Why use PR-Agent?

A reasonable question that can be asked is: "Why use PR-Agent? What make it stand out from existing tools?"

Here are some advantages of PR-Agent:

• We emphasize real-life practical usage. Each tool (review, improve, ask, ...) has a single GPT-4 call, no more. We feel that this is critical for realistic team usage - obtaining an answer quickly (~30 seconds) and affordably.
• Our PR Compression strategy is a core ability that enables to effectively tackle both short and long PRs.
• Our JSON prompting strategy enables to have modular, customizable tools. For example, the '/review' tool categories can be controlled via the configuration file. Adding additional categories is easy and accessible.
• We support multiple git providers (GitHub, Gitlab, Bitbucket, CodeCommit), multiple ways to use the tool (CLI, GitHub Action, GitHub App, Docker, ...), and multiple models (GPT-4, GPT-3.5, Anthropic, Cohere, Llama2).
• We are open-source, and welcome contributions from the community.

Roadmap

• Support additional models, as a replacement for OpenAI (see here)
• Develop additional logic for handling large PRs (see here)
• Add additional context to the prompt. For example, repo (or relevant files) summarization, with tools such a ctags
• PR-Agent for issues, and just for pull requests
• Adding more tools. Possible directions:
  • PR description
  • Inline code suggestions
  • Reflect and review
  • Rank the PR (see here)
  • Enforcing CONTRIBUTING.md guidelines
  • Performance (are there any performance issues)
  • Documentation (is the PR properly documented)
  • ...

Similar Projects

• CodiumAI - Meaningful tests for busy devs (although various capabilities are much more advanced in the CodiumAI IDE plugins)
• Aider - GPT powered coding in your terminal
• openai-pr-reviewer
• CodeReview BOT
• AI-Maintainer