Local Chatbot with RAG

This project implements a local chatbot using Retrieval-Augmented Generation (RAG) technology. It consists of three main components: a chatbot API, a user interface, and the infrastructure to support the chatbot system.

Ideal for developers, researchers looking to implement a customizable, privacy-focused chatbot solution with the power of large language models and the flexibility of local deployment.

Table of Contents

• Chatbot API
• Chatbot UI
• Chatbot Infrastructure
• Getting Started
• Contributing
• License

Chatbot API

The chatbot-api component is responsible for handling the core logic of the chatbot, including:

• Processing user inputs
• Generating appropriate responses using RAG technology
• Managing conversation context
• Integrating with external services and databases

Key Features

• RESTful API endpoints for chatbot interactions
• Document upload and text extraction (PDF, DOCX, TXT)
• Vector-based document search
• Integration with Ollama for language model inference
• Database storage for chat history and documents

Technologies Used

• Python 3.12
• FastAPI
• SQLAlchemy
• Alembic for database migrations
• FAISS for vector storage
• Sentence Transformers for text embedding

Chatbot UI

The chatbot-ui component provides the user interface for interacting with the chatbot. It includes:

• A responsive web interface built with React.js
• Real-time chat functionality
• Model selection from available models
• Document upload functionality
• Chat history display
• Dark mode UI using Material-UI

Technologies Used

• React.js
• Material-UI
• Axios for API communication

Chatbot Infrastructure

The chatbot-infrastructure component manages the deployment, scaling, and monitoring of the chatbot system. It includes:

• Containerization using Docker
• Docker Compose for local development and testing
• PostgreSQL database setup
• Environment configuration for API and UI components

Key Components

• Docker containers for API, UI, and database
• Makefile for easy management of services
• Integration with locally running Ollama instance

Getting Started

To get started with the project, follow these steps:

1. Clone the repository:

   git clone https://github.com/doanhat/local-chatbot-with-rag-with-fastapi-ollama-react-faiss-postgres.git
   cd local-chatbot-with-rag-with-fastapi-ollama-react-faiss-postgres

2. Set up and start the infrastructure:

   cd chatbot-infrastructure
   make run

3. Access the components:

   • API: http://localhost:8000/api
   • UI: http://localhost:3000
   • API Documentation: http://localhost:8000/docs

For detailed setup instructions for each component, refer to their individual README files:

• Chatbot API README
• Chatbot UI README
• Chatbot Infrastructure README