I have a list-of-list of word groups in Turkish. I want to apply stemming and I found turkishnlp package. Although it has some shortcomings, it often returns the right word. However, when I apply this to the list, I don't want the structure of my list to change and I want the words that he doesn't know to stay the same.
For example, I have this list: mylist = [['yolda','gelirken','kopek', 'gördüm'],['cok', 'tatlıydı']]
And I wrote this function:
from trnlp import TrnlpWord
def tr_stemming(x):
obj = TrnlpWord()
obj.setword(x) if isinstance(x, str) else type(x)(map(tr_stemming, x))
return obj.get_stem if isinstance(x, str) else type(x)(map(tr_stemming, x))
This function returns this list:
tr_stemming(mylist)
[['yol', 'gelir', '', 'gör'], ['', 'tatlı']]
However, I want to get this as the output: [['yol', 'gelir', 'kopek', 'gör'], ['cok', 'tatlı']]
How can I update my function? Thank you for your helps!
CodePudding user response:
IIUC, you could modify your function to:
def tr_stemming(x):
if isinstance(x, str):
obj = TrnlpWord()
obj.setword(x)
stem = obj.get_stem
return stem if stem else x
elif isinstance(x, list):
return [tr_stemming(e) for e in x]
out = tr_stemming(mylist)
output:
[['yol', 'gelir', 'kopek', 'gör'], ['cok', 'tatlı']]