Closed kalikiana closed 4 years ago
Hi @kalikiana There is an option for "Search similar words", clicking on the magnifying glass of search entry.
Today it is quite rigid since it uses levenshtein distance up to 1 character.. and in the proposed case it should work.. :wink:
p.s: I'm seeing you continue with Midori .. I'm very happy !! :smiley:
There is an option for "Search similar words", clicking on the magnifying glass of search entry.
Today it is quite rigid since it uses levenshtein distance up to 1 character.. and in the proposed case it should work.. wink
Hmmm so I enabled that option and it works in most cases like Jö schau
, Pokëmon
, Ill Niño
or Zé Glisse
, but not for Óró sé do bheatha 'bhail
D'Oh!. 'Óró' vs 'Oro' has an levenshtein distance of 2 characters.. :sweat_smile:
As comment before, now it is very rigid. It only check differences of 1 character, and it does not differentiate between if it is the same letter with different accentuation or a completely different letter.. :disappointed:
The ideal would be to ignore when it was the same letter with different accentuation. But I see it difficult and expensive at the CPU level.. Probably the best we can do is to make it more flexible, for example using distance 2 when the needle has at least 4 characters
Hi @kalikiana Can you test the last commit?. (If you use 1.3, you can apply the patch cleanly). Thanks, and sorry for the delay ..
It would be great if I could type
Óró sé do bheatha 'bhaile
by just typingoro
without worrying about the diactrics. In technical terms the search would need to find words by their canonicalized unicode characters.