I'm trying to POS-tag some sentences in Italian with Apertium's tagger. While according to the Apertium GitHub page I am supposed to get as output also the surface form in addition to the morphological analysis, I only get the analysis. I want also the surface form. I cannot infer it since the tagger doesn't necessarily tag a single token, so I cannot simply tokenize the original sentence and loop over it or zip it with the tagger's output.
According to the GitHub page:
In [1]: import apertium
In [2]: tagger = apertium.Tagger('ita')
In [3]: tagger.tag('gatti').
Out[3]: [gatti/gatto<n><m><pl>]
What I got:
In [1]: import apertium
In [2]: tagger = apertium.Tagger('ita')
In [3]: tagger.tag('gatti') # 'gatti' is the surface form
Out[3]: [gatto<n><m><pl>]
How can I get the surface form? If I provided one token at a time this would not be a problem since I would know what the token is. But in a sentence I cannot know how the tagger creates chunks.