You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Also went down a bit of a side quest, capturing the results here so I don't forget:
I had expected node-diacritics to normalize "BİM" to "bim" but it wasn't doing this. (this is important for 2 things: so our fuzzy matcher can match on either version, also used to coax our generated ids into ascii)
node-diacritics has a handful of open issues and hasn't seen an update in 4 years. The Turkish "İ" is not mentioned as an issue, but I thought it might be worth upgrading to a newer library for doing this.
I tried out normalize-diacritics which does mostly the same thing but is newer and receives updates
normalize-diacritics does correctly normalize "BİM" to "bim", but unfortunately the library is just too slow! Building the match index went from around 250ms to around 18sec.
So I am sticking with the older node-diacritics library for now but I added some code to simplify.js to handle the Turkish "İ".
I might look into adding similar workarounds in simplify.js for the other open issues in node-diacritics ('л', and stuff here)
I've recently seen that iD is lacking PTT (the post service company - https://en.wikipedia.org/wiki/PTT_(Turkey) ) and İş Bankası (https://en.wikipedia.org/wiki/%C4%B0%C5%9Fbank) presets. Also, the Bim preset should be replaced with capitalized letters, BİM, because it's branded as that. (https://nsi.guide/index.html?t=brands&k=shop&v=supermarket&tt=bim)
The text was updated successfully, but these errors were encountered: