'kana' field differs between the raw MeCab output and the Fugashi tagger output, returning "体" about fugashi HOT 2 CLOSED

Apologies, I made this issue a little hastily at the end of a long day staring at code, I should have done a more careful analysis first.

My problem was actually the other way around, with my parsed MeCab output reading the wrong field. That was my misunderstanding of MeCab's output - indeed, the confusing nature of MeCab is exactly why I'm switching to fugashi anyway 😄

Closing this issue, thanks for your time.

from fugashi.

Because your code is incomplete I'm not entirely sure what's going on.

import fugashi
tagger = fugashi.Tagger()
nodes = tagger("年")
print(nodes[0].feature.kana)
# => ネン

On my system the above code prints ネン, using fugashi==1.3.0 and unidic-lite==1.0.8.

• what dictionary are you using? Use tagger.dictionary_info to check.
• what version of fugashi and your dictionary are you using? (are you using unidic-lite?)
• can you provide example code I can run as a complete (short) file? Something like the example above.

from fugashi.