ask for the Stemming feature

[redstoneleo]

I'd like this feature :
https://github.com/binhetech/CyHunspell#stemming

[zverok]

The possible implementation of stemming with Spylls is demonstrated in this discussion: #19 (comment)

I'll gladly accept a PR that will make it more convenient, but unfortunately don't have time to work on this myself

[redstoneleo]

As I have tested with word wrote , it is not accurate

[zverok]

That just depends on the dictionary.

In the standard en-US dictionary, wrote is specified as a separate word form (like most of irregular verbs, I guess). Note that hunspell and its dictionaries are designed, first and foremost, as a spell-checking tool, not a full-fledged linguistic analysis package.

[redstoneleo]

I tested with https://github.com/cdhigh/chunspell
it worked as expected

[zverok]

Please show which Hunspell dictionaries you have used with both, and what was the code you have tried.

[redstoneleo]

For chunspell, by default you have the only en_US dictionaries available. --https://github.com/cdhigh/chunspell?tab=readme-ov-file#dictionaries

from hunspell import Hunspell
hunSpell = Hunspell()
print(hunSpell.stem('wrote'))#gives ('write', 'wrote')

With spylls, I used the code you gave at the comment section #19 (comment)

from spylls.hunspell import Dictionary
# en_US dictionary is distributed with spylls
# See docs to load other dictionaries
dictionary = Dictionary.from_files('en_US')
from spylls.hunspell.algo.capitalization import Type as CapType

for form in dictionary.lookuper.affix_forms('wrote', captype=CapType.NO): 
  print(form.stem))#only gives 'wrote'