chrisdrymon / angel Goto Github PK

I'm assuming you tried this but I see you're not generating lemmas. What are your thoughts on generating lemmas?

e.g. from John 1.38 in MorphGNT SBLGNT I get:

('(ὃ', 'u--------')

things like the following should be output to stderr in my opinion, so the output of my own script can be redirected.

Loading models...
Part-of-speech models loading...
Person models loading...
Number models loading...
Tense models loading...
Mood models loading...
Voice models loading...
Gender models loading...
Case models loading...
Degree models loading...
Pre-processing text...
Text and punctuation split into 17752 individual tokens.
Angel's looking at each word by itself...
Reconsidering tags...
Studying each word in light of its context...

Calling tag() on multiple strings reloads the models each time. It would be great to load them up on the first call and then reuse them.

I wonder if, when the model needs to be downloaded, it shouldn't prompt the user first (perhaps giving the size). If the user declines, it just terminates.