-
-
Main Home Page
-
Example Insights Page
-
Example Suggestions
Inspiration
With the rapid rise of AI-driven code generation, the need for effective, automated code review has never been greater. Frequent reviews can cause developer fatigue and increase the risk of missed vulnerabilities. This project aims to make the code review process more efficient, customizable, and accurate.
What it does
Code Review Analysis ChecK is an automated tool that takes in a GitHub PR to analyze code for vulnerabilities, suggest improvements, and prioritize issues based on importance. It enhances developer workflows by providing tailored review feedback directly within their repositories.
How we built it
We built Code Review Analysis ChecK using Python for backend logic, leveraging Flask for the web interface and integrating with GitHub’s API for seamless interaction with repositories. The machine learning model for the OpenAI API identifies vulnerabilities and critical code changes. The application runs on a Heroku-deployed server and uses custom algorithms to flag issues and add comments in pull requests. Frontend components are designed with HTML and CSS, optimizing the UI for ease of review and productivity.
Challenges we ran into
We faced challenges around finding and using robust ML models for varied code bases and languages. After spending many hours trying to get a larger AI model to run on a small gpu, we settled with the OpenAI API. We struggled a bit on getting the suggestions to actually be helpful.
Accomplishments that we're proud of
We're proud of creating a solution that brings effective vulnerability detection to developers' fingertips, improves review accuracy, and reduces review fatigue by focusing on high-priority issues. Were happy with how the user interface turned out since it simple but effective.
What we learned
We gained valuable insights into integrating APIs like GitHub’s into applications, setting up a reliable web application environment, and designing user-friendly interfaces for streamlined interaction. We also learned about the complexities of implementing machine learning for code review automation and the importance of balancing automation with user control.
What's next for Code Review Analysis ChecK
Next, we plan to enhance automation and customizability, including developing advanced models for more precise vulnerability detection and prioritization. Expanding GitHub integration will allow continuous updates and feedback within the developer’s workflow. Additionally, we aim to improve the UI with new features that increase productivity and allow organizations to focus on critical review areas tailored to their needs.
Log in or sign up for Devpost to join the conversation.