Skip to content

Latest commit

 

History

History
201 lines (119 loc) · 7.82 KB

README.md

File metadata and controls

201 lines (119 loc) · 7.82 KB

Hackathon Logo

Sitecore Hackathon 2024

⟹ Insert your documentation here <<

Hackathon Submission Entry form

Team name

XM Boys

Category

  1. Best Module for XM Cloud (AI Cloud Service)

Description

XM Boys aims to leverage generative AI to provide a cloud service akin to Pages, Explorer, and Components, as part of Composable DXP. This AI platform, named Jarvis, will retrieve information from a selected template, allowing users to edit content similarly to Explorer. Moreover, a helper button, the AI Assistant, will open a prompt where users can provide context for the text field they wish to generate for their module. Users can also offer module or company context to ensure accurate results from the AI tool.

Finally, clicking the "Generate Content" button will create an item in Sitecore at the desired location. This initiative seeks to streamline content authoring and marketing efforts, providing an AI assistant for generative content, thereby saving time for users.

Video link

Demo Video Link in Youtube

Pre-requisites and Dependencies

XM Cloud Solution

Jarvis API (/src/jarvis-api)

  • Install Node JS 20.10.0

Jarvis UI (/src/jarvis)

  • Install Node JS 20.10.0

Installation Instructions

In order to use our module, it is necessary to install 3 different solutions:

  • XM Cloud starter kit
  • Our Jarvis API
  • Our Jarvis UI

Install XM Cloud Solution

This repository is based on the XM Cloud starter kit.

  1. In an ADMIN terminal:
.\init.ps1 -InitEnv -LicenseXmlPath "C:\path\to\license.xml" -AdminPassword "DesiredAdminPassword"
  1. Restart your terminal and run:
.\up.ps1
  1. Push Items
dotnet sitecore ser push

Install Jarvis API

The API is created in the /src/jarvis-api folder. Since it is an ExpressJS app, first you need to install the dependencies:

npm install

Also, this app uses a .env.local file to get some environment variables, so duplicate the .env file and rename it to .env.local. To set those variables, see the configuration section.

Build the solution:

npm run build

After that, you need to run the solution by executing this command:

npm run serve

This will initialize the API app at http://localhost:3001

Install Jarvis UI

The UI is a separate NextJS app. To do this, go to the /src/jarvis directory and install the dependencies:

npm install

For this project, the .env file only contains the URL for the API app. As it is http://localhost:3001, duplicate the .env file and rename it to .env.local.

Finally, Build & run the solution:

npm run build
npm run start

This will initialize the UI at http://localhost:3000

Configuration

Configure Jarvis API Environmental Variables

Variable Description
PORT Port where the services will be running (default: 3001)
GRAPH_QL_ENDPOINT Endpoint for the GraphQL API (e.g., "https://xmcloudcm.localhost/sitecore/api/authoring/graphql/v1")
GRAPH_QL_API_KEY API key for accessing the Authoring GraphQL API
AZURE_OPEN_AI_ENDPOINT Endpoint for the Azure OpenAI service (e.g., "https://{your-instance}.openai.azure.com")
AZURE_OPEN_AI_DEPLOYMENT_NAME Name of the Azure OpenAI deployment
AZURE_OPEN_AI_API_KEY API key for accessing the Azure OpenAI service
AZURE_OPEN_AI_API_VERSION API version for the Azure OpenAI service (e.g., "2024-02-15-preview")

Configure Jarvis API - Azure Open AI Service

  • Create an Azure OpenAI Service
  • Open Azure AI Studio
  • Create a Deployment
    • Choose gpt-35-turbo
    • Model version 0301

For more information, refer to How To Set Up and Configure a GPT Deployment Using the Azure OpenAI Service

Usage Instructions

Developer Configuration

  1. Create your SXA Component by following the steps for Cloning Rendering:

    • More Info
    • Go to Template Definition
    • Inherit the _Searchable on Jarvis template from Feature/Jarvis feature.

Marketer Use Case

  1. Go to the Jarvis UI at http://localhost:3000

  2. Fill in the "Please Provide a Context" field with Company Information or Component-Relevant information that needs to be used to generate the content.

    • Example: "My company sells shoes. We are running a campaign to sell new shoes for 2024, targeting teenagers."

    Fill Context

  3. Name: This input will be used to give a name to the Sitecore item.

  4. Select a Template: Choose the template to work with.

Step 4

  1. Choose Parent Location where this component will be saved.

Step 5

  1. Based on your selection, the form will be dynamically rendered according to the component's fields.

Step 6

  1. Fill out all the fields and use the AI Assistant for Generative AI.

    • Provide contextual information about the field in the prompt.

    Step 7

    Step 8

  2. Create Component Datasource.

    Step 9

  3. Connect a component in Sitecore and connect the datasource.

    Step 10 Step 11

  4. Enjoy the rest of your day!

Step 12

Comments

It is important to configure Azure Open AI properly, as well as ensuring proper network connectivity between Jarvis API, Jarvis UI, and the XM Cloud Instance (Docker containers).