@grishka Yay. redaktor got that too. I am interested...

Yay.
redaktor got that too. I am interested in “Alphabet doesn't matter” - can you give a pointer to code?
What I did a while ago is in
https://github.com/redaktor/widgets-preview/tree/master/src/framework/String/search
and worked on phonetic search …

PS, just in case anyone needs to detect languages, this covers 800+ … https://github.com/redaktor/languages

Like 26 Jul 2021 at 6:40 | Wall-to-wall | Open on mastodon.social

3 comments

Gregory

@sl007 https://github.com/grishka/Smithereen/blob/d1433c5531be7c513860c9a369ec138afd180e92/src/main/java/smithereen/storage/SearchStorage.java

I use this https://github.com/jirutka/unidecode to transliterate everything into Latin alphabet and store these strings in a separate table (qsearch_index). So, basically, search only ever operates on Latin strings.

26 Jul 2021 at 10:27 | Open on mastodon.social

Gregory

@sl007 so because:
- I convert names into Latin when creating and updating the search index
- I convert search queries into Latin when searching
I could type your name in Cyrillic and still find you:

26 Jul 2021 at 10:34 | Open on mastodon.social

Gregory

@sl007 also notice that there's a Sebastián with an accented á and it still found him

26 Jul 2021 at 10:36 | Open on mastodon.social

Go Up