Get Hands-On: Build an Intelligent Data Application powered by the SAP HANA Cloud Vector Engine

Estimated read time 5 min read

With the release of SAP HANA Cloud’s Vector Engine, our development team built an exciting Movie Insight Demo App, which showcase the engine’s powerful vector-based querying and storage capabilities. This demo illustrates the vector engine’s role in retrieval-augmented generation (RAG), enhancing LLM responses with contextual data. Now, the entire demo is available on GitHub, providing a hands-on experience for developers & tech enthusiasts to explore the Vector Engine capabilities and practical application.

Before getting into the details, I want to give a shoutout to our awesome development team: Igor Pessoa Rocha, Wenzel Svojanovsky, Wanping Chen and Christian Schuetz for their contribution in building this demo from scratch.

What Does the Demo App Offer?

The Movie Insight Demo App enables users to understand how vector embeddings enhance Large Language Models (LLMs) by providing context-aware responses. The demo illustrates a Retrieval-Augmented Generation workflow, where additional context is retrieved via embedding vectors before a request is sent to the LLM. The similarity between two vectors is calculated through cosine similarity & L2Distance functions, which measures the semantic likeness of two texts, even across different languages.

Reference Architecture

Key Features:

Multiple Scenarios: Allows users to explore movie datasets from 1977 to 2024.RAG Mode: Toggle between direct LLM requests or context-augmented queries using vector embeddings.Expert and Standard Views: Offers a simplified view for beginners or an expert mode for advanced users, showing process steps and allowing custom selections.Interactive Elements: Visualize process durations, browse movie datasets, and customize LLM prompts.

The Movie Insight App – An Intelligent Data Application

Role of SAP HANA Cloud vecor engine in the app:

In this demo scenario, the movie titles, their release years, and plots are all stored in SAP HANA Cloud, with the plots being converted & stored as vector embeddings. These embeddings are crucial for providing contextual information to the LLM, by storing data as vectors, the vector engine enhances the LLM’s ability to understand and respond to user queries with context, thus improving accuracy and relevance.

Get Started

To try out the Movie Insight Demo App, check out the full repository on GitHub. The code provides everything needed to build an intelligent data app using SAP HANA Cloud’s Vector Engine. Whether you are a developer looking to understand retrieval-augmented generation or want to explore how cosine similarity can enrich similarity search results, this demo app offers an in-depth, hands-on experience.

If you would like to start using the SAP HANA Cloud vector engine, you can spin up a free trial instance of SAP HANA Cloud in SAP BTP Cockpit by following the steps in this tutorial.

 

​ With the release of SAP HANA Cloud’s Vector Engine, our development team built an exciting Movie Insight Demo App, which showcase the engine’s powerful vector-based querying and storage capabilities. This demo illustrates the vector engine’s role in retrieval-augmented generation (RAG), enhancing LLM responses with contextual data. Now, the entire demo is available on GitHub, providing a hands-on experience for developers & tech enthusiasts to explore the Vector Engine capabilities and practical application.Before getting into the details, I want to give a shoutout to our awesome development team: Igor Pessoa Rocha, Wenzel Svojanovsky, Wanping Chen and Christian Schuetz for their contribution in building this demo from scratch.What Does the Demo App Offer?The Movie Insight Demo App enables users to understand how vector embeddings enhance Large Language Models (LLMs) by providing context-aware responses. The demo illustrates a Retrieval-Augmented Generation workflow, where additional context is retrieved via embedding vectors before a request is sent to the LLM. The similarity between two vectors is calculated through cosine similarity & L2Distance functions, which measures the semantic likeness of two texts, even across different languages.Reference ArchitectureKey Features:Multiple Scenarios: Allows users to explore movie datasets from 1977 to 2024.RAG Mode: Toggle between direct LLM requests or context-augmented queries using vector embeddings.Expert and Standard Views: Offers a simplified view for beginners or an expert mode for advanced users, showing process steps and allowing custom selections.Interactive Elements: Visualize process durations, browse movie datasets, and customize LLM prompts.The Movie Insight App – An Intelligent Data ApplicationRole of SAP HANA Cloud vecor engine in the app:In this demo scenario, the movie titles, their release years, and plots are all stored in SAP HANA Cloud, with the plots being converted & stored as vector embeddings. These embeddings are crucial for providing contextual information to the LLM, by storing data as vectors, the vector engine enhances the LLM’s ability to understand and respond to user queries with context, thus improving accuracy and relevance.Get StartedTo try out the Movie Insight Demo App, check out the full repository on GitHub. The code provides everything needed to build an intelligent data app using SAP HANA Cloud’s Vector Engine. Whether you are a developer looking to understand retrieval-augmented generation or want to explore how cosine similarity can enrich similarity search results, this demo app offers an in-depth, hands-on experience.If you would like to start using the SAP HANA Cloud vector engine, you can spin up a free trial instance of SAP HANA Cloud in SAP BTP Cockpit by following the steps in this tutorial.   Read More Technology Blogs by SAP articles 

#SAP

#SAPTechnologyblog

You May Also Like

More From Author