feat: enable to select collections as components in problems picker

[bra-i-am]

This pull request refactors and optimizes how collections and their components are indexed and filtered in the library authoring UI. The main improvements are the introduction of a centralized collection indexing context for O(1) lookups, frontend filtering for the Collections tab to enable more flexible workflows, and associated updates to component interfaces and context providers. These changes improve performance and maintainability, especially when working with large sets of collections and components.

Collection indexing and context improvements:

• Introduced CollectionIndexProvider and related hooks/types in ComponentPickerContext.tsx to precompute and share collection/component relationships and sizes across the UI, avoiding repeated computation and improving performance. This includes new interfaces and a context for accessing the precomputed data. [1] [2]
• Updated imports and wrapped relevant page components (LibraryAuthoringPage and LibraryCollectionPage) with CollectionIndexProvider to ensure all children have access to the precomputed collection index data. [1] [2] [3] [4] [5] [6]

Filtering and content display logic:

• Moved filtering of collections in the Collections tab from the backend to the frontend in LibraryContent.tsx, ensuring all components remain available in the search hits for selection workflows. Adjusted the filtering logic in LibraryAuthoringPage.tsx accordingly. [1] [2] [3]
• Updated LibraryCollectionComponents.tsx to use the default LibraryContent (with frontend filtering) instead of passing a collections-specific content type. [1] [2]

Type and interface enhancements:

• Extended selection-related interfaces to support collections, including new types for selected collections, collection status, and events that handle both components and collections. [1] [2] [3] [4]

Test data updates:

• Added block_id fields to mock collection data in library-search.json to better reflect the updated data structures and support new indexing logic. [1] [2] [3] [4] [5]

How to test

Make sure you have a collection in your Libraries ( http://apps.local.openedx.io:2001/authoring/libraries ) with questions in it

1. Have a Teak environment
2. Mount this MFE and switch to this branch
3. Install https://github.com/eduNEXT-collab/xblock-exam-question-bank in the CMS
4. Go to the Authoring MFE ( http://apps.local.openedx.io:2001/authoring/ ), create a new course or use any existing one, go to Settings > Advanced Settings and add "examquestionbank" to the list of Advanced module list
5. Go to the content outline and access a unit
6. Click Advanced in the component selector and add an Exam Question Bank
7. Click Add Problems and select a collection; it should be selected and deselected properly, and once you click Add Selected Components, they should be displayed in the xblock as will be shown in the following screencast

Screencast

Screencast.from.09-01-26.14.40.43.webm

[MaferMazu]

The functionality works fine, thanks. My only concern is the efficiency of selecting elements from the collections (because the number of problems and collections will increase). We can handle documentation that explains the client's limitations, but if we can find a way to make it more efficient, that would be great.

[MaferMazu]

Can you add more information on why we need this change? Something simple.
I won't argue about the test change, but it could be valuable to track the why.

[bra-i-am]

@MaferMazu, this PR allows selecting collections too, due to it, once in the test it is queried screen.queryAllByRole('button', { name: 'Select' }) which returns an array bigger than before (before contained 3 items, now it includes 9 items counting the collections) and taking into account that in the library-search.json the collections are before the components (and rendered before the components), the index of the components changed from 1 to 7 (with 6 collections in between)

I could change the library-search.json items order and preserve the test with index 1, what do you think?