You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The segments package provides Unicode Standard tokenization routines and orthography segmentation,
implementing the linear algorithm described in the orthography profile specification from
The Unicode Cookbook (Moran and Cysouw 2018 ).
Command line usage
Create a text file:
$ echo "aäaaöaaüaa" > text.txt
Now look at the profile:
$ cat text.txt | segments profile
Grapheme frequency mapping
a 7 a
ä 1 ä
ü 1 ü
ö 1 ö
Write the profile to a file:
$ cat text.txt | segments profile > profile.prf
Edit the profile:
$ more profile.prf
Grapheme frequency mapping
aa 0 x
a 7 a
ä 1 ä
ü 1 ü
ö 1 ö
Now tokenize the text without profile:
$ cat text.txt | segments tokenize
a ä a a ö a a ü a a
And with profile:
$ cat text.txt | segments --profile=profile.prf tokenize
a ä aa ö aa ü aa
$ cat text.txt | segments --mapping=mapping --profile=profile.prf tokenize
a ä x ö x ü x
API
>>>fromsegmentsimportProfile, Tokenizer>>>t=Tokenizer()
>>>t('abcd')
'a b c d'>>>prf=Profile({'Grapheme': 'ab', 'mapping': 'x'}, {'Grapheme': 'cd', 'mapping': 'y'})
>>>print(prf)
Graphememappingabxcdy>>>t=Tokenizer(profile=prf)
>>>t('abcd')
'ab cd'>>>t('abcd', column='mapping')
'x y'
About
Unicode Standard tokenization routines and orthography profile segmentation
