LeetCode RAG

A personalized AI mentor that helps you review and improve on LeetCode problems using your own study history.

🚀 About

This project implements a Retrieval-Augmented Generation (RAG) pipeline that ingests your LeetCode practice logs (from Excel) and allows you to chat with an AI that knows your specific past mistakes, confidence levels, and notes.

It uses LangChain, ChromaDB, and OpenAI to index your notes and retrieve relevant past solutions when you ask questions.

🚧 Work in Progress / Roadmap

This project is currently under active development. The current version uses Gradio, but major updates are planned:

• Retriever Optimization: Improving the SelfQueryRetriever logic to make context retrieval more reliable.
• Frontend Migration: Moving from Gradio to Chainlit for a more robust chat interface.
• History Persistence: Implementing logic to save and resume previous conversation sessions.

🛠️ Setup & Usage

1. Configure Environment: Create a .env file in the root directory:

   OPENAI_API_KEY=your_api_key
   EXCEL_PATH=path/to/your/leetcode_excel_file.xlsx

2. Data Source: The system expects an Excel file containing your practice history (columns: date, question_name, difficulty, topic, algorythm, notes_/_mistakes, confidance).

3. Run the App:

   python gradio_app.py

📚 Tech Stack

• LangChain (Orchestration)
• ChromaDB (Vector Store)
• OpenAI (Embeddings & LLM)
• Pandas (Data Ingestion)
• Gradio (Current UI)