Data Architect - Knowledge Chat Backend

A knowledge-based chat system that integrates with various data sources including GitHub repositories, SQL Schema scripts, and PDF documents.

Features

• Vector Database Integration: Uses ChromaDB for efficient similarity search
• Multiple Data Source Support:
  • GitHub repositories
  • SQL Schemas with schema analysis
  • PDF documents
• Code Analysis: Analyzes code structure and relationships
• Chat Interface: Query your knowledge base using natural language

Setup

Prerequisites

• Python 3.9+
• Ollama for embeddings
• Git

Installation

1. Clone the repository:

   git clone https://github.com/KKranthi6881/Data-Architect.git
   cd Data-Architect
   

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Start the backend server:

   uvicorn src.app:app --reload --host 0.0.0.0 --port 8000
   

5. Start the frontend (in a separate terminal):

   cd knowledge-chat-app
   npm install
   npm run dev
   

Usage

Adding Knowledge Sources

1. GitHub Repositories: Connect to GitHub repositories to analyze code
2. SQL Scripts: Upload SQL scripts to analyze schema and relationships
3. PDF Documents: Upload PDF documents for knowledge extraction

Querying Knowledge

Use the chat interface to ask questions about your connected knowledge sources.

Architecture

• Backend: FastAPI with ChromaDB vector store
• Frontend: React with Chakra UI
• Embeddings: Ollama for text embeddings
• Analysis: Custom code analyzer for SQL and programming languages

License

MIT