🔍 IssueLens

AI-powered GitHub Repository Intelligence Platform for semantic code search, repository indexing, and intelligent issue analysis.

---

📖 Overview

Modern software repositories contain thousands of files, functions, and GitHub issues. Finding the right implementation or understanding how a feature works often requires manually searching through large codebases.

IssueLens is an AI-powered code intelligence platform that indexes GitHub repositories, extracts code structure using Abstract Syntax Trees (AST), generates semantic embeddings, and enables natural-language search over source code.

Instead of searching with keywords like:

timedelta

Developers can ask:

"Where is time converted into a timedelta?"

IssueLens understands the meaning of the query and retrieves the most relevant code using vector similarity search.

---

✨ Current Features

Repository Management

• Import GitHub repositories
• Clone repositories locally
• GitHub REST API integration
• MongoDB Atlas metadata storage

Code Intelligence

• AST-based Python parser
• Function extraction
• Class extraction
• Method detection
• Qualified symbol names
• Source code chunk extraction
• Docstring extraction

AI Search

• Semantic code embeddings
• SentenceTransformer (all-MiniLM-L6-v2)
• FAISS vector indexing
• Natural language code search
• Cosine similarity ranking

Backend

• FastAPI
• MongoDB Atlas
• Modular architecture
• Environment configuration
• Async services
• Health monitoring

---

🚧 Upcoming Features

• GitHub Issue Import
• Semantic Issue Search
• Duplicate Issue Detection
• AI-generated Issue Summaries
• Repository Knowledge Graph
• RAG-based Question Answering
• Multi-language Support
• Incremental Repository Indexing
• Web Dashboard
• LLM-powered Code Explanations

---

🏗 Architecture

                GitHub Repository
                        │
                        ▼
                 Git Clone Service
                        │
                        ▼
                 AST Python Parser
                        │
                        ▼
                  Code Chunks
                        │
                        ▼
           SentenceTransformer Embeddings
                        │
                        ▼
                 FAISS Vector Index
                        │
                        ▼
             Semantic Code Search API
                        │
                        ▼
                  FastAPI Backend
                        │
                        ▼
                    React Frontend

---

🛠 Tech Stack

Backend

• FastAPI
• Python 3.11
• Motor
• HTTPX
• Pydantic v2

Database

• MongoDB Atlas

AI & Search

• Sentence Transformers
• all-MiniLM-L6-v2
• FAISS
• NumPy

Git Integration

• GitPython
• GitHub REST API

Frontend (Upcoming)

• Next.js
• React
• Tailwind CSS

---

📂 Project Structure

IssueLens/

├── backend/
│
│   ├── app/
│   │
│   ├── api/
│   │   └── routes/
│   │
│   ├── core/
│   ├── db/
│   ├── indexing/
│   │   ├── chunker.py
│   │   ├── embedder.py
│   │   ├── utils.py
│   │   └── vector_store.py
│   │
│   ├── models/
│   ├── parsers/
│   ├── repositories/
│   ├── services/
│   │
│   └── main.py
│
├── storage/
│   └── repos/
│
├── frontend/
│
└── docs/

---

🚀 Getting Started

Clone

git clone https://github.com/YOUR_USERNAME/IssueLens.git

cd IssueLens/backend

---

Install

pip install -r requirements.txt

or

pip install -e .

---

Environment

Create .env

MONGODB_URL=

DATABASE_NAME=issuelens

GITHUB_TOKEN=

---

Run Backend

uvicorn app.main:app --reload

Swagger

http://127.0.0.1:8000/docs

---

🔍 Example Semantic Search

Instead of searching:

timedelta

Developers can search:

convert seconds into timedelta

IssueLens returns:

_make_timedelta()

Similarity: 0.60

storage/repos/pallets__flask/src/flask/app.py

Lines 73-77

without relying on keyword matching.

---

📈 Roadmap

Phase 1 ✅

• FastAPI Backend
• MongoDB Atlas
• GitHub Repository Import
• Git Repository Cloning
• AST-based Python Parser
• Code Chunk Extraction
• SentenceTransformer Embeddings
• FAISS Semantic Search

---

Phase 2 🚧

• Repository Index API
• Search API
• GitHub Issue Import
• Issue Embeddings
• Similar Issue Detection
• AI Issue Summaries

---

Phase 3 🚀

• Repository Dashboard
• RAG Question Answering
• LLM Code Explanation
• Multi-language Parsing
• Incremental Re-indexing
• Analytics Dashboard

---

🎯 Motivation

Developers spend significant time:

• Searching unfamiliar codebases
• Understanding large repositories
• Finding relevant implementations
• Connecting GitHub issues to source code
• Navigating thousands of functions

IssueLens reduces this effort by combining AST parsing, vector embeddings, semantic search, and AI-powered code understanding into a unified developer platform.

---

🤝 Contributing

Contributions are welcome!

1. Fork the repository
2. Create a feature branch

git checkout -b feature/my-feature

3. Commit your changes

git commit -m "feat: add awesome feature"

4. Push

git push origin feature/my-feature

5. Open a Pull Request

---

📄 License

MIT License

---

👨‍💻 Author

Yash Bhat

Built with ❤️ using Python, FastAPI, MongoDB, FAISS, and AI.