PaCaBaNu

Example use of our app

Inspiration

The challenge sparked the idea of utilizing InterSystems IRIS Vector Search alongside GenAI to develop a unique holiday planning system. We aimed to integrate different similarity search techniques and RAG to streamline vacation planning for users.

What it does

Our system offers two distinct use cases. In the first scenario, users communicate their preferences—whether they prefer big cities, unny destinations, beach getaways, places with a lot of history... Leveraging this input, our program suggests the most suitable city. Not stopping there, it utilizes RAG to provide interesting facts about the recommended city and highlights nearby attractions, each accompanied by a brief description.

In the second scenario, users can opt for a visual approach by providing a reference image. The system then identifies visually similar cities as potential holiday destinations.

How we built it

This project is built primarily in Python, incorporating various libraries such as llama_index, sentence_transformers, PyTorch, and Pandas. Additionally, we utilized Electron to craft an intuitive and visually appealing interface.

What we learned

The project illuminated the versatility of RAG, demonstrating its efficacy even in unconventional applications. Additionally, we gained insights into interacting with databases, particularly those containing vectors, directly from Python, facilitating seamless integration within our system.