CodeMentor AI

Your Intelligent DSA Pair Programmer & Algorithm Visualizer

CodeMentor AI is an interactive learning and debugging platform designed to help developers and students master Data Structures & Algorithms (DSA). Chat naturally, analyze your code conceptually, inspect failing counterexamples, and visualize step-by-step execution traces with AI-generated diagrams.

---

What You Can Do with CodeMentor AI

1. Chat & Learn Naturally

• Easy Problem Setup: Simply mention a problem by name (e.g., "How do I solve 3Sum?", "Explain Trapping Rain Water") or paste your code. CodeMentor automatically extracts the problem description, sets up your code editor, and prepares test cases.
• Conceptual Mentorship: Get clear, encouraging explanations tailored to your skill level without dense compiler errors or robotic jargon.

2. Deep Code Analysis & Instant Feedback

• Approach & Logic Breakdown: Understand the algorithmic pattern behind your code (Two Pointers, Sliding Window, Dynamic Programming, etc.).
• Complexity Derivations: View clear asymptotic Time and Space Complexity explanations for your solution.
• Compare Approaches: Compare your current solution against brute-force and optimal approaches side-by-side.

3. Pinpoint Bug Diagnosis & Counterexamples

• Why Does It Fail?: Discover logic bugs, off-by-one errors, and boundary issues with plain-English explanations.
• Concrete Failing Test Cases: See exact counterexample inputs where your solution breaks, comparing what your code returns vs. the expected answer.
• Clean Fixes: View clean, corrected code with non-intrusive annotations explaining the fix.

4. Visual Dry-Run Traces

• On-Demand Execution Diagrams: Ask for a dry run anytime ("Show dry run", "Trace execution") to generate a visual diagram illustrating data structures, pointer movements, and variable updates.
• Full-Screen Lightbox: Zoom in on execution diagrams or download them for offline study.

5. In-Place Code Editor & Multi-Test Manager

• Interactive Code Editor: Syntax-highlighted code editor supporting C++, Python, Java, JavaScript, TypeScript, Go, and Rust.
• Dual-Pane Workspace: Chat comfortably on the left while keeping your active problem, code, and test cases accessible on the right.
• Full-Screen Mode: Expand the code editor or problem statement into a focused full-screen popup modal anytime.
• Custom Test Cases: Add, edit, or delete multiple test cases directly in the workspace panel.

6. Interactive Action Chips

• One-Click Next Steps: Dynamic action buttons appear after each explanation (e.g., "Show Dry Run", "Why is it wrong?", "Show Optimal Solution") to guide your learning journey seamlessly.

7. Saved Sessions & History

• Personalized Account: Sign in securely with Google.
• Multi-Session History: Switch between past problem discussions or start fresh sessions anytime from the sidebar drawer.

---

Project Structure

├── backend/
│   ├── agents/               # AI Agent
│   │   ├── planner.py        # Orchestrates conversations, tool routing & context syncing
│   │   ├── prompts.py        # System instructions and Agent prompt
│   │   ├── tool.py           # Mark functions as agent tool
│   │   ├── tools.py          # 10 specialized DSA analysis and illustration tools
│   │   └── schemas.py        # Pydantic data schemas and intent definitions
│   ├── api/
│   │   └── routes/           # FastAPI API endpoints (auth, query, sessions)
│   ├── db/
│   │   └── mongodb.py        # MongoDB connection manager
│   ├── services/             # Session management, chat history, and dry-run image generation
│   ├── tests/                # Integration test suite
│   ├── main.py               # FastAPI Entry point
│   └── requirements.txt      # Python dependencies
│
├── frontend/
│   ├── src/
│   │   ├── components/
│   │   │   ├── auth/         # Google auth modal
│   │   │   ├── cards/        # Visual UI cards
│   │   │   ├── chat/         # Chat stream, input area, and dynamic action chips
│   │   │   ├── common/       # Monaco code editor and Markdown renderer
│   │   │   ├── home/         # landing page
│   │   │   └── workspace/    # Dual-pane layout and history drawer
│   │   ├── context/          # User state provider
│   │   ├── hooks/            # Custom React hooks
│   │   ├── registry/         # Component registry mapping AI structured outputs
│   │   ├── services/         # API communication
│   │   ├── types/            # TypeScript interfaces
│   │   └── App.tsx           # Main application router and split-view manager
│   ├── package.json          # Node dependencies and scripts
│   └── vite.config.ts        # Vite configuration
│
└── README.md                 # Project documentation

---

Tech Stack

Layer	Technologies
Backend	FastAPI
Agent Engine	Google ADK, Google GenAI SDK
Database	MongoDB
Data Validation	Pydantic
Frontend	React (Vite), TypeScript, Tailwind CSS

---

How to Get Started

Prerequisites

• Python 3.10+ (Python 3.12 recommended)
• Node.js 18+ & npm
• MongoDB (local installation or MongoDB Atlas free tier)
• Gemini API Key from Google AI Studio

---

Step 1: Start the Backend

1. Navigate to the project root and create a virtual environment:

   # Windows
   python -m venv venv
   .\venv\Scripts\Activate.ps1
   
   # macOS / Linux
   python3 -m venv venv
   source venv/bin/activate

2. Install dependencies:

   pip install -r backend/requirements.txt

3. Create a backend/.env file:

   GEMINI_API_KEY=your_gemini_api_key
   MONGODB_URL=mongodb://localhost:27017
   DB_NAME=codementor_db
   JWT_SECRET=your_jwt_secret_key
   GOOGLE_CLIENT_ID=your_google_oauth_client_id
   GOOGLE_CLIENT_SECRET=your_google_oauth_client_secret

4. Start the backend server:

   python -m uvicorn backend.main:app --host 0.0.0.0 --port 8000 --reload

---

Step 2: Start the Frontend

1. Navigate to the frontend directory:

   cd frontend
   npm install

2. Create a frontend/.env file:

   VITE_API_URL=http://localhost:8000/api
   VITE_GOOGLE_CLIENT_ID=your_google_oauth_client_id

3. Start the application:

   npm run dev

4. Open http://localhost:5173 in your browser and enjoy learning with CodeMentor AI!