SamurAI application

Key features

• Virtual assistant interface for document-based information retrieval using RAG pattern
• Natural language conversation capabilities for questions and answers based on the knowledge base
• Source-linked responses - each assistant's answer includes references to source documents
• Knowledge base management:
  • Support for various document formats (focusing on text content)
  • Document content editing capabilities
  • Document management by designated moderators
• User roles and permissions:
  • Regular users - can interact with the assistant
  • Moderators - can manage documents and view user conversations
  • Administrators - can manage user roles and system access
• Conversation monitoring tools for quality assurance

Motivation

This project is a part of my engineering thesis. The goal was to create some useful system using new technologies and learn as much as possible.

Running the application with Docker

Create .env file based on .env.example file. Run the application using docker using the following command:

docker compose -f docker-compose-local.yml up --build

It will build the application and run it in docker containers along with all necessary dependencies .

App frontend will be available at http://localhost.

By default there should be some sample data in the database - you can login to the application using one of the sample users: user:user, mod:mod, admin:admin.

Backend

See samurai-be/README.md for more information.

Frontend

See samurai-fe/README.md for more information.