A sophisticated basketball analysis chatbot built with LangChain, Hugging Face Llama models, and Pinecone vector database. This application provides intelligent responses to basketball-related questions using advanced AI and vector search capabilities.

• 🤖 AI-Powered Responses: Uses Hugging Face language models for natural conversation
• 🔍 Vector Search: Pinecone integration for semantic search of basketball knowledge
• 📚 Comprehensive Knowledge Base: Built-in basketball rules, positions, strategies, and statistics
• 🌐 Modern Web Interface: Beautiful Streamlit-based UI with real-time chat
• ⚡ Quick Actions: Pre-built buttons for common basketball questions
• 🔄 Real-time Updates: Dynamic knowledge base refresh capabilities

• LangChain: Conversation management and AI orchestration
• Hugging Face: Language models and transformers
• Pinecone: Vector database for semantic search
• Streamlit: Web application framework
• Sentence Transformers: Text embeddings
• Python: Core programming language

• Python 3.8 or higher
• Hugging Face account and API token
• Pinecone account and API key (optional but recommended)
• Git (for cloning the repository)

git clone <repository-url>
cd basketball-chatbot

python setup.py

The setup script will:

• Check Python version compatibility
• Install all required dependencies
• Guide you through API key configuration
• Test the installation
• Create sample basketball knowledge

If you prefer manual setup:

pip install -r requirements.txt

Create a .env file in the project root:

# Hugging Face API Token
HUGGINGFACE_API_TOKEN=your_huggingface_token_here

# Pinecone API Key (optional)
PINECONE_API_KEY=your_pinecone_api_key_here
PINECONE_ENVIRONMENT=your_pinecone_environment_here
PINECONE_INDEX_NAME=basketball-analysis

# Model Configuration
MODEL_NAME=microsoft/DialoGPT-medium
EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2

streamlit run app.py

The application will open in your default browser at http://localhost:8501.

1. Go to Hugging Face Settings
2. Create a new token
3. Copy the token and add it to your .env file

1. Go to Pinecone Console
2. Create an account and get your API key
3. Note your environment (e.g., us-east-1-aws)
4. Add both to your .env file

Note: The chatbot will work without Pinecone, but vector search capabilities will be limited.

The chatbot includes comprehensive basketball knowledge covering:

• Basic basketball rules
• Scoring system
• Game duration and timing
• Fouls and violations

• Point Guard (PG)
• Shooting Guard (SG)
• Small Forward (SF)
• Power Forward (PF)
• Center (C)

• Pick and roll
• Zone defense
• Man-to-man defense
• Fast break
• Half-court offense

• Points per game (PPG)
• Rebounds per game (RPG)
• Assists per game (APG)
• Field goal percentage (FG%)
• Three-point percentage (3P%)
• Player efficiency rating (PER)

• "What are the rules of basketball?"
• "How many points is a three-pointer worth?"
• "What is a pick and roll?"

• "What are the responsibilities of a point guard?"
• "How do you calculate field goal percentage?"
• "What makes a good shooting guard?"

• "What is zone defense?"
• "How does a fast break work?"
• "What are the different offensive strategies?"

basketball-chatbot/
├── app.py                 # Main Streamlit application
├── basketball_chatbot.py  # Core chatbot logic
├── basketball_knowledge.py # Basketball knowledge base
├── vector_store.py        # Pinecone vector database operations
├── config.py             # Configuration management
├── setup.py              # Setup script
├── requirements.txt      # Python dependencies
├── env_example.txt       # Environment variables template
├── README.md            # This file
└── .env                 # Environment variables (created during setup)

You can customize the language model in config.py:

MODEL_NAME = "meta-llama/Llama-2-7b-chat-hf"  # For Llama-2 (requires access)
MODEL_NAME = "microsoft/DialoGPT-medium"      # Default (no access required)

VECTOR_DIMENSION = 384    # Embedding dimension
METRIC = "cosine"        # Similarity metric
CHUNK_SIZE = 1000        # Text chunk size
CHUNK_OVERLAP = 200      # Chunk overlap

You can extend the basketball knowledge base by modifying basketball_knowledge.py:

def get_custom_knowledge(self):
    return [
        {
            "title": "Your Custom Topic",
            "content": "Your custom basketball knowledge here."
        }
    ]

To use different models, update the configuration:

# For Llama-2 (requires Hugging Face access)
MODEL_NAME = "meta-llama/Llama-2-7b-chat-hf"

# For other models
MODEL_NAME = "gpt2"  # Smaller model
MODEL_NAME = "EleutherAI/gpt-neo-125M"  # Alternative

1. Import Errors: Make sure all dependencies are installed

   pip install -r requirements.txt

2. API Key Errors: Verify your API keys are correctly set in .env

   cat .env  # Check your environment variables

3. Model Loading Issues: Try using the default DialoGPT model

   MODEL_NAME = "microsoft/DialoGPT-medium"

4. Pinecone Connection Issues: Check your environment and API key

   PINECONE_ENVIRONMENT = "us-east-1-aws"  # Verify your environment

• Use smaller models for faster responses
• Reduce MAX_LENGTH for quicker generation
• Adjust TEMPERATURE for more focused responses

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

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

• LangChain for conversation management
• Hugging Face for language models
• Pinecone for vector database
• Streamlit for the web interface
• Basketball community for knowledge and insights

If you encounter any issues or have questions:

1. Check the troubleshooting section
2. Review the configuration options
3. Open an issue on GitHub
4. Check the documentation

Happy Basketball Analysis! 🏀