Ingesting content from an external source with Pega Knowledge Loader

In this topic, learn how to ingest PDF files from a SharePoint site. This process involves creating a Knowledge Loader, and then verifying whether Knowledge Buddy can reference the ingested files in its answers. While this topic focuses mainly on SharePoint, the differences between a SharePoint Loader and a Confluence loader are highlighted where appropriate. Knowledge Buddy functions the same, regardless of which type of Loader you use to ingest content.

The SharePoint sample site used in this demonstration contains the Guides folder, with two subfolders: Implementation Guide Install guide. Each folder contains one or more PDF documents. The goal is to have a Knowledge Buddy answer questions based on those files.

The ingested documents go into the Product Guide data source. Currently, there is no content in this data source, which you verify by logging in to the Knowledge Buddy Portal, and then clicking the Completed Content, In progress content, or Failed content tabs of the data source.

You begin the process by logging in to the Knowledge Loader Portal. Click Create, and then click SharePoint Loader. Write an appropriate name and description for your loader in the Label and Description fields respectively, and then click Submit.

Next, configure the loader with the correct values. You plan to ingest content into the Product Guide data source, which is in the Knowledge collection. For this action, you pull documents from the site folder Shared Documents/Guides, and then secure it with the KnowledgeBuddy:Public role.

Click Show optional settings. Because you plan to ingest only PDF files, in the File types to include field, enter pdf. Configure any other optional settingsthat you consider appropriate, and then click Submit to start the ingestion scheduler.

Now, return to the Knowledge Buddy Portal.

Because the ingestion happens periodically, the Knowledge Buddy is unable to answer questions immediately after you create the SharePoint Loader. However, you can return to the data source to confirm when the ingestion happens successfully. In the navigation pane of Knowledge Buddy, click Data source, and then click the ProductGuide data source. The Completed content tab displays the status, object ID of the ingested content, and any possible errors.

You can see that the Buddy successfully ingested the three documents from your SharePoint site.

To verify that the Buddy is referencing the correct article, you can configure Knowledge Buddy to return the URL of the referenced article in its response.

Reopen the Buddy at the Prompt Stage, click Actions, and then click Edit. In the Response attributes list, select URL. Click Submit to apply the changes.

Now, when you ask the question how to create a buddy implementation application? Knowledge Buddy can respond.

You can click Review response to see the references. By highlighting the URLs, you can see that Knowledge Buddy is referring to the ingested data with URLs that point to your documents on the SharePoint site.