🧭 RepoCodeNavi – AI-Powered Codebase Navigator

RepoCodeNavi is an AI-powered assistant that helps developers understand and navigate any GitHub repository instantly using RAG, Prompting, Structured Output, and Function Calling.

---

🚀 Problem Statement

Navigating large or unfamiliar codebases is challenging:

• Documentation is often outdated or incomplete.
• New contributors waste hours figuring out where features live.
• Searching code manually is slow and error-prone.

RepoCodeNavi solves this by connecting natural language questions directly to the relevant parts of the codebase.

---

💡 Features

• 🔍 Natural Language Code Search – Ask anything about the repo.
• 🗂 AI Summaries – Clear, beginner-friendly explanations of files/folders.
• 🐛 Issue-to-Code Mapping – Links GitHub issues to related files.
• 📚 Learning Suggestions – Guides you to resources for unfamiliar tech.
• 🖥 Function Calling – AI can open files, fetch issues, and trigger actions.
• 📊 Structured Output – AI responses are returned as clean, structured JSON for UI rendering.

---

🛠 Tech Stack

Layer	Technology
Frontend	React.js (Vite)
Backend	Node.js + Express
AI Layer	Python (FastAPI) + OpenAI/GPT + LangChain
Embeddings	Pinecone / FAISS
Integrations	GitHub OAuth + GitHub API
Deployment	Docker + Vercel (Frontend) + Render (Backend)

---

⚡ Why RAG, Prompting, Structured Output & Function Calling?

• RAG: Retrieves actual repo context before generating responses → less hallucination.
• Prompting: Ensures explanations are clear, concise, and relevant.
• Structured Output: Returns JSON so frontend can display results in an interactive format.
• Function Calling: AI triggers actions like opening files or fetching issue details.

---

📂 How It Works

1. Authenticate with GitHub → Choose a repository.
2. Index the Repo → Files, README, and docs are embedded into a vector DB.
3. Ask Questions → e.g., “Where is the login logic?”
4. RAG Retrieves Context → Relevant code/files fetched.
5. AI Generates Response → Uses prompting for clear explanations.
6. Structured Output + Function Calls → Data shown in UI and actions triggered.

---

📦 Installation (Local Development)

# Clone the repo
git clone https://github.com/your-username/RepoCodeNavi.git

# Navigate into the project
cd RepoCodeNavi

# Install dependencies
npm install   # Frontend
pip install -r requirements.txt  # Backend AI Service

# Start development servers
npm run dev