Skip to content

Latest commit

 

History

History
126 lines (80 loc) · 7.52 KB

README.md

File metadata and controls

126 lines (80 loc) · 7.52 KB

Exercise 04 - Explore the SAP HANA Cloud Vector Engine

The SAP HANA Cloud is the leading in-memory database on the market. It is also the go-to database for CAP applications. SAP has added the vector engine to the SAP HANA database. The vector engine provides the data foundation for a new set of natural and intuitive capabilities. For this CodeJam, you will use the vector engine to augment large language models with business context. You will use the vector engine for two things. The engine stores relevant business data as vector embeddings and, second, searches for similar and relevant vector embeddings based on the user's prompts. The corresponding text of these vector embeddings then get passed along to a defined LLM for accurate business answers using that contextual information.

The nature of a vector engine is to seamlessly work with generative AI applications. By providing the ability to store different types of data, such as text, images, or sound, in the form of vector embeddings, the vector engine can be used to pass data along together with Retrieval Augmented Generation (RAG) requests, natural language processing, image recognition, similarity searches, and recommendation systems.

For this CodeJam, the SAP HANA Cloud instance has already been initialized and configured. For detailed instructions on how to set up SAP HANA Cloud on SAP BTP, you can follow the SAP BTP Setup Guide.

In this exercise, you will learn:

  • What vector embeddings are.
  • How to enable the SAP HANA Cloud vector engine
  • How to configure your CAP application for SAP HANA deployment
  • How to bind your CAP application to a new HDI instance

What are Vector Embeddings

"Vector embeddings are mathematical representations used to encode objects into multi-dimensional vector space. These embeddings capture the relationships and similarities between objects. SAP HANA Cloud Vector Engine will facilitate the storage and analysis of complex and unstructured vector data(embeddings) into a format that can be seamlessly processed, compared, and utilized to build various intelligent data applications and add more context in case of GenAI scenarios."

Source: Vectorize your Data: SAP HANA Cloud's Vector Engine for Unified Data Excellence

How do you enable SAP HANA Cloud Vector Engine

The SAP HANA Cloud vector engine comes bundled with the SAP HANA Cloud instance on the SAP Business Technology Platform. Within your subaccount on SAP BTP, you or the global account admin can assign entitlements to enable the SAP HANA Cloud service plan. Vector embeddings are just one of the supported data types for SAP HANA. They can be used for SQL queries and HANA operations.

Configure the project for SAP HANA deployment

You must configure your CAP project for SAP HANA deployment. You can do this by executing the cds add hana command. The command configures deployments for SAP HANA to use the hdbtable and hdbview formats. If you add the argument of --profile hybrid to the command, the configuration is added to a [hybrid] profile in your package.json.

👉 Open the project in SAP Business Application Studio (BAS) or VSCode.

👉 Open a new terminal if not already open.

👉 Execute the cds add hana --for hybrid command:

cds add hana --for hybrid

explore-sap-hana-cloud-vector-engine-add-hana

The profile hybrid relates to the hybrid testing scenario, which allows you to run your CAP service on localhost but use a real connection to a running SAP HANA Cloud instance. You will utilize this feature in a later exercise.

👉 Open the package.json file in your project.

👉 Notice how there is a new section defined containing SAP HANA-related configuration:

"cds": {
    "requires": {
      "[hybrid]": {
            "db":"hana"
      }      
    }
  }

Create your HDI container for HANA deployment

For this CodeJam each of you will create their own HDI container.

Use the Cloud Foundry CLI to create your HDI container (replace the placeholder with a meaningful name):

cf create-service hana hdi-shared <your hdi-container name>

The creation of the HDI container takes some time. You can check the process via the following command:

cf services

explore-sap-hana-cloud-vector-engine-create-hdi-instance

If the service creation is completed you can go to the next step to create the binding.

explore-sap-hana-cloud-vector-engine-create-hdi-instance

Bind local service to SAP HANA Schemas & HDI Containers instance for hybrid testing

At this point, your CAP application knows that you want to use an SAP HANA database for persistence, but it doesn't know what database instance you want to use and has no connection. You need to bind your application to the SAP HANA Cloud instance configured on SAP BTP. When you bind your application, an HDI container instance gets spun up for your application's data, and a service key is issued to authenticate against that instance.

The CDS CLI can bind your application to the SAP HANA Schemas & HDI Containers instance.

👉 Open the project in SAP Business Application Studio (BAS) or VSCode.

👉 Open a new terminal if not already open.

👉 Log into the provided Cloud Foundry instance and authenticate using the provided credentials:

cf login -a https://api.cf.us10.hana.ondemand.com

👉 Create a binding for your CAP application (replace the placeholder with a meaningful name):

cds bind -2 <your-hdi-container-name>:<your-service-key-name>

explore-sap-hana-cloud-vector-engine-binding-config

Summary

At this point, you have learned what the SAP HANA Cloud vector engine is, how vector embeddings work, how to get your project ready for SAP HANA deployment, and how to bind your application to an HDI container instance. In the following exercises, you will evolve your project into a fully functioning CAP AI service application utilizing SAP HANA Cloud vector engine, Generative AI Hub, embedding, and chat LLMs.

Further reading

Next exercise