This commit installs version 1.21.0 of @edx/frontend-lib-learning-assistant. This commit also refactors the Chat tests to assert on whether the Xpert component is rendered by the Chat component, not whether Xpert actually renders. This is because Xpert now has its own logic to determine whether to render.
This release uses a new GET endpoint published on the Learning Assistant backend to determine whether the Learning Assistant feature is enabled. If the features is not enabled, the Learning Assistant is not rendered, and vice-versa.
This commit modifies the Course component to pass the unitId prop, which presents the unit usage key in which the Xpert is being rendered, to the Chat component. The Chat component passes this unit ID as a prop to the Xpert component.
In order to diagnose low usage rates, we're temporarily adding Segment events for rendering the chat and for closing the call-to-action message. This will help us determine whether Xpert is being successfully rendered and whether Xpert is being ignored by tracking whether learners close the call-to-action message at a high rate.
At a minimum, we expect to remove the Segment event for rendering the Xpert, because this will be a very noisy Segment event. We plan to leave it in for a few days just to accumulate data. We will evaluate whether to keep the call-to-action dismissal Segment event based on the data.
This commit adds the Segment event for rendering the Xpert. This commit also installs version 1.11.1 of frontend-lib-learning-assistant, which introduces a Segment event for for closing the call-to-action message.
This commit installs the Xpert chatbot feature from the frontend-lib-learning-assistant repository into the frontend-lib-learning application.
This component is rendered by the Course component. The component is only rendered when a few conditions are satisfied.
