chiragagrawal24/cursor_2d_animation

Cursor 2D Animation Pipeline

An intelligent educational animation generation system that converts topics into high-quality Manim animations using AI-powered LangGraph pipelines, web search, and RAG (Retrieval Augmented Generation).

🎯 Overview

This project creates educational animations automatically by:

1. 1.Topic Analysis: Breaking down educational topics into digestible scenes
2. 2.AI-Powered Code Generation: Using Google Gemini LLM with web search and Manim documentation RAG
3. 3.Automated Animation: Generating complete Manim Python scripts for each scene
4. 4.Modular Architecture: Clean, maintainable code structure with separated concerns

🚀 Features

• —🧠 Intelligent Scene Generation: Automatically breaks down complex topics into animated scenes
• —🔍 Web-Enhanced Research: Uses Tavily search for current information and context
• —📚 RAG-Powered Documentation: Leverages Supabase vector store with Manim documentation
• —⚡ Parallel Processing: Concurrent scene generation for improved performance
• —🎬 Manim Integration: Generates production-ready Manim animation scripts
• —🔄 LangGraph Pipeline: Robust workflow orchestration with state management
• —📊 Progress Tracking: Real-time pipeline execution monitoring

📁 Project Structure

Cursor_2D_Animation/
├── src/                           # Core source code
│   ├── editor/                    # Animation editing components
│   ├── langraph_pipeline/         # Main AI pipeline modules
│   │   ├── config.py             # Environment and LLM configuration
│   │   ├── main_modular.py       # Clean main entry point
│   │   ├── manim_generator.py    # Manim code generation with RAG
│   │   ├── pipeline.py           # LangGraph workflow orchestration
│   │   ├── scene_generator.py    # Scene description generation
│   │   └── state.py              # Pipeline state management
│   ├── renderer/                  # Animation rendering components
│   ├── storage/                   # Data storage utilities
│   └── utils/                     # Common utility functions
├── scripts/                       # Utility scripts
│   └── RAG/                      # RAG system components
│       ├── chunk_docs.py         # Document chunking utilities
│       ├── crawl_recursive.py    # Web crawling for documentation
│       ├── create_table.sql      # Database schema
│       ├── insert_docs.py        # Document insertion pipeline
│       └── utils.py              # RAG utility functions
├── assets/                        # Static assets and resources
├── docs/                          # Project documentation
├── tests/                         # Unit and integration tests
├── out/                           # Generated animations output
├── requirements.txt               # Python dependencies
└── README.md                      # This file

🛠️ Installation

Prerequisites

• —Python 3.10+
• —FFmpeg (for video rendering)
• —Git

Setup

1. 1.Clone the repository

   git clone https://github.com/yashpinjarkar10/Cursor_2D_Animation.git
   cd Cursor_2D_Animation

1. 1.Install dependencies

   pip install -r requirements.txt

1. 1.Environment configuration Create a .env file with your API keys:

   GOOGLE_API_KEY=your_gemini_api_key
   TAVILY_API_KEY=your_tavily_api_key
   SUPABASE_URL=your_supabase_url
   SUPABASE_SERVICE_KEY=your_supabase_key
   LANGSMITH_API_KEY=your_langsmith_key  # Optional for debugging

🎮 Usage

Basic Usage

cd src/langraph_pipeline
python main_modular.py

Advanced Options

# Test mode with predefined topic
python main_modular.py --test

# Debug mode with detailed logging
python main_modular.py --debug

Example Workflow

1. 1.Run the pipeline: python main_modular.py
2. 2.Enter your educational topic (e.g., "Pythagorean Theorem")
3. 3.Watch as the AI generates scene descriptions
4. 4.Generated Manim scripts appear in generated_scenes/
5. 5.Rendered videos will be saved to out/

🔧 Key Technologies

• —🤖 AI/ML: Google Gemini 2.5 Flash, LangChain, LangGraph
• —🎬 Animation: Manim Community Edition
• —🔍 Search: Tavily Web Search API
• —📚 RAG: Supabase Vector Database, Google Embeddings
• —🕸️ Web Crawling: Crawl4AI for documentation scraping
• —🎥 Video Processing: MoviePy, FFmpeg
• —📊 Monitoring: LangSmith (optional)

🧩 Core Components

LangGraph Pipeline (src/langraph_pipeline/)

• —State Management: Typed state flow through the pipeline
• —Scene Generation: AI-powered topic breakdown into scenes
• —Code Generation: RAG-enhanced Manim script creation
• —Parallel Processing: Concurrent scene processing for efficiency

RAG System (scripts/RAG/)

• —Web Crawling: Recursive documentation crawling
• —Document Processing: Intelligent chunking and embedding
• —Vector Storage: Supabase-based retrieval system
• —Query Enhancement: Context-aware documentation search

📊 Output

The pipeline generates:

• —Manim Scripts: Complete Python files ready for rendering
• —Scene Descriptions: Human-readable animation breakdowns
• —Execution Logs: Detailed pipeline performance metrics
• —Rendered Videos: Final MP4 animations (when rendering is enabled)

🔮 Future Enhancements

• —[ ] Video Editor Integration: Automated scene assembly
• —[ ] Advanced Rendering: GPU acceleration and optimization
• —[ ] Interactive UI: Web-based interface for topic input
• —[ ] Template System: Reusable animation patterns
• —[ ] Batch Processing: Multiple topic processing
• —[ ] Quality Assessment: Automated animation quality metrics

🤝 Contributing

Contributions are welcome! Please:

1. 1.Fork the repository
2. 2.Create a feature branch
3. 3.Make your changes with tests
4. 4.Submit a pull request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• —Manim Community for the excellent animation framework
• —LangChain Team for the AI orchestration tools
• —Google for the Gemini LLM API
• —Supabase for the vector database platform

Built with ❤️ for educational content creators and AI enthusiasts ---