API highlight control

[CommanderStorm]

Added the ability to control the highlighting behaviour.
As noticed in #103 It would be nice to make this customizable and not result in hacks as seen in TUM-Dev/Campus-Android#1462 (comment)
Resolves #114

Note:
I have changed the highlighting fields to <em>. I can revert this, if you'd like, but I assume this removes at least a bit of code in the webclient, so I thought this would be a better solution.

[octycs]

Codewise I think this is fine (only the HTML escaping would be good to still have). The only reason I am not sure whether to switch to MeiliSearches highlighting now, is, that they still seem to treat 'typos' in numbers. For example:

It happens only for IDs longer than 4 characters, which are mainly only in the MW and CH buildings. It's not too bad, but I don't really like it. What do you think?

Edit: in some cases also 4 chars only:

[CommanderStorm]

I am not sure whether to switch to MeiliSearches highlighting now, is, that they still seem to treat 'typos' in numbers. It's not too bad, but I don't really like it. What do you think?

Well…
But that is how the search results get produced.
The highlighting only shows where things match…

I do think that it is better to show that MS does typo-tolerance even in these circumstances…
We can potentially tweak the typo tolerance a bit (rel #115), but there is just not much we can do…

I honestly don't see why you think this is a problem… Could you explain your reasoning a bit more? 😅

[octycs]

I honestly don't see why you think this is a problem… Could you explain your reasoning a bit more? sweat_smile

When you go through the search results (in a quick manner), I think you mainly want to see as directly as possible whether this is the result you were looking for or not. And the highlighting helps you by showing you that the highlighted parts match exactly what you typed in.

I don't think that the typo-tolerance in search is bad, but I don't like it in the highlighting. I can understand that MeiliSearch does this, because it could be helpful if you search large text databases, but my problem is that it also applies to numbers.

I think our ID search is in general quite okay, but not ideal. So I don't object to merging this, it's just that I wouldn't want to leave it this way in the long run (It does also have advantages of course, especially Meilisearch doesn't highlight too much if your search contains a single digit).