DA-Search-Engine

Overiew

Compatible Database

• Japan Search (Under testing)
• NDL (Under testing)

Purpose of this Project

1. Utilize external interfaces (APIs) from multiple databases to retrieve metadata in response to user queries to the LLM, and then use RAG (Retrieval-Augmented Generation) to output the results of the database queries in natural language.
2. Furthermore, through prompt engineering, it enables optimized output tailored to the each individual, facilitating not only for researchers and experts but also more wide range of people.
3. The goal is to utilize multimodal AI (text, images, audio, etc.) for database retrieval and output visualization, allowing AI that has learned multimodal cognition to act as a hub between users and database. This enables multimedia search and output, creating a knowledge base that can process complex multimedia informations.

Features

Coming soon!!!

Getting Started

Prerequisites

Coming soon!!!

Installation

Coming soon!!!

Configuration

Coming soon!!!

To-Do List

Tasks

Core Functionality

• Task 1
  • Query generation to SQL database through API.
  • Retrieve RDF format information.
  • Embedding to vector database.
  • Generate responce.
• Expand supported database.
• Development of front end.

Contact

For any questions, feel free to reach out to me at [[email protected]].