Ignore diacritics in search field

[sbraz]

Is your feature request related to a problem? Please describe.
Hi,
Would it be please be possible to ignore diacritics in the search field? My device names have plenty of accents and other French diacritics but I don't always feel like typing them when I'm filtering.

Describe the solution you'd like
Searching for "general" would return a device whose name contains "général". Same for "ca" and "ça".

Describe alternatives you've considered
Typing the diacritics, not adding diacritics to friendly names.

Additional context
I assume this would mean pre-processing all columns using some kind of "unidecode" package? Or just converting to Unicode NFD and removing combining characters?

[Nerivec]

We can do something like this:

"général".normalize("NFD").replace(/\p{Diacritic}/gu, "")

https://onecompiler.com/javascript/43yx5x69h
Should be fast enough to not need pre-processing. But I'm not sure all languages would agree with this. There might be a few cases where this would not be entirely correct in matching. I guess we can put it behind a toggle in frontend settings.

[sbraz]

That would be lovely, thanks :)
In any case, you would just pre-process the column, not the user input? So users wishing to use diacritics in their query could still do so? And the match would be done both on the unprocessed column name and the pre-processed name?

What I mean is if I looked for général, you'd match it against général and general, only the former would match.
And If I looked for general, only the latter would match?

[Nerivec]

The filtering is handled by the package used by WindFront, and has a specific behavior depending on the column (text, number, fuzzy, strict, builtin, custom, etc.)
https://tanstack.com/table/latest/docs/guide/global-filtering
Which means this is not a small update. And have to be careful of the impact on perf (mobile, low cpu) and on various languages (might fit for FR, might not for another - if I'm not mistaken, just removing diacritic doesn't work for some languages).

[sbraz]

Which means this is not a small update. And have to be careful of the impact on perf

Sure, that makes sense. Maybe using a dedicated package to perform the transliteration would be preferable for other languages?

just removing diacritic doesn't work for some languages

Even French has characters that would not be handled by the NFD form decomposition, for instance œ which should be translated to oe. And one obvious German example comes to mind too: ß which should be translated as ss.

However, if we stick to doing things in a way similar to e.g. Firefox's "Match Diacritics", I think the NFD trick would be enough (searching for ss in Firefox with "Match Diacritics" unticked doesn't match ß nor does oe match œ). Perhaps adding this option would be good first step if it's not too much work? And if users wish to have this improved to support more complex conversions, you could look into this later?

If you want to put this on hold, that's absolutely fine by me too. Thanks for the awesome work on the frontend :)

[Nerivec]

I have to check some things in the package used, see what kind of matching it does exactly. Maybe we can add/enhance a filter in the package directly, instead of customizing in WindFront, as needed.

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.

[sbraz]

Not stale, unless @Nerivec wants to reject this.

[Nerivec]

Haven't had the time to look into this yet. Need to integrate the behavior with the library, and probably provide a config to turn it on/off.