Livedata option dropdown has no grouping semantics

As a user with vision, when you see this dropdown

You understand that there's a list of options/ actions, and that this list is split into subcategories. This is the semantics we share through style.
The semantics we share through structure of the HTML should be similar. Before this PR, it was:

A list of options/ actions. We did not convey any information about the groups in the HTML.

For the sake of users of assistive tools, and overall accessibility of XWiki, you want to make sure the semantics you convey through your HTML are roughly the same as the semantics you convey through your style. This is [adaptability](https://www.w3.org/WAI/fundamentals/accessibility-principles/#adaptable).

This PR makes sure we represent those subcategories in the HTML structure. Additionally, it removes the "I am one item of the list" semantics from the separator, because a separator is not an item anyone would name when asked to list the elements of this dropdown (meaning it has no semantics and should be showed using non semantic elements/ no element style only).
____
Before AT would approx read:
"List with 6 items - item 1:Actions - item 2: Refresh - item 3:separator - item 4: Layouts - item 5:Table - item6: Cards [...]"
With the new structure they would read:
"List with 2 items - item 1: Actions, List with 1 item - item 1: Refresh - item 2: Layouts, List with 2 items - item 1: Table - item 2: Cards"