A production-ready, Retrieval-Augmented Generation (RAG) chatbot built for Global AI Hub. It indexes Q&A content, retrieves the most relevant chunks with Chroma, and generates helpful answers using Google Gemini.

• Simple RAG pipeline: chunk → embed → store → retrieve → generate
• Google Gemini 2.0 Flash for response generation
• text-embedding-004 for high‑quality embeddings
• Flask UI, Markdown rendering, persistent Chroma vector store

• Backend: Flask
• RAG: LangChain, Chroma
• LLM: Google Gemini (gemini-2.0-flash)
• Embeddings: Google models/text-embedding-004
• Data: Markdown Q&A (data/soru_cevap.md)

• Python 3.10+
• A valid Google API key with access to Gemini models

1. Clone the repository

   git clone https://github.com/enesmanan/gaih-chatbot.git
   cd gaih-chatbot

2. Install dependencies

   pip install -r requirements.txt

3. Configure environment Create a .env file in the project root:

   GOOGLE_API_KEY=your_google_api_key_here
   

4. Create the vector database (first run or when data/model changes)

   python create_database.py

5. Start the app

   python app.py

   Then open your browser at http://localhost:5000.

• Generative model: gemini-2.0-flash (temperature=0.57)
• Embedding model: models/text-embedding-004
• Chunking: 2000 characters, 300 overlap
• Retrieval k: 10

gaih-chatbot/
├── data/
│   └── soru_cevap.md         # Q&A dataset (Markdown)
├── app.py                    # Flask app + RAG query flow
├── create_database.py        # Chunk + embed + persist to Chroma
├── chroma_db/                # Persisted vector store (auto-created)
├── static/                   # CSS and images
├── templates/                # HTML templates
├── requirements.txt          # Dependencies
├── .env                      # GOOGLE_API_KEY (not committed)
└── README.md                 # This file