🤖 Code Review Comment Analyzer

🔗 Live App: https://code-comment-reviewer-kx5pfmcjc-ishita-kanaujias-projects.vercel.app/

🚀 Overview

This project is a full-stack NLP-powered web application that analyzes GitHub pull request comments and classifies them into actionable categories:

🐞 Bug
❓ Question
💡 Suggestion
🎨 Style Nitpick

It helps developers and teams understand feedback patterns and improve code quality.

🏗️ Architecture

This project follows a modern full-stack architecture:

React (Frontend) → FastAPI (Backend) → ML Model (Scikit-learn)

Frontend (React + Vite)
- User inputs GitHub PR URL
- Displays summary + categorized comments
Backend (FastAPI)
- Fetches PR comments using GitHub API
- Runs ML model for classification
- Returns structured results
ML Model
- TF-IDF + LinearSVC
- Trained on labeled code review comments

🛠️ Tech Stack

Frontend

React (Vite)
JavaScript
CSS

Backend

FastAPI
Python

Machine Learning

Scikit-learn
Pandas
NLTK
Joblib

📊 Model Performance

Accuracy: 93%

precision recall f1-score support

Bug 0.71 0.79 0.75 68
Question 0.67 0.52 0.59 23
Style 0.64 0.56 0.60 32
Suggestion 0.96 0.96 0.96 908

accuracy 0.93 1031

⚙️ Features

🔗 Analyze real GitHub PRs
📊 Summary of comment categories
🧠 ML-based classification
⚡ Fast API responses
🎯 Clean and responsive UI

🧪 How to Run Locally

1. Clone the repo

git clone https://github.com/your-username/code-comment-reviewer.git
cd code-comment-reviewer

Run Backend

cd backend
pip install -r requirements.txt
uvicorn main:app --reload

Run Frontend

cd frontend
npm install
npm run dev

Create Environment Variables

Create a .env file inside frontend/:

VITE_API_URL=http://127.0.0.1:8000

Key Learnings

Built full-stack app (React + FastAPI)
Integrated ML model into production API
Handled API communication & environment configs
Designed clean UI/UX for real users

FilesExpand file tree

readme.md

Latest commit

History