Closed koghuzhan closed 3 years ago
The machine seems to break down words such as
ko
rsatadi qo
mitasiat `.
Maybe it's only when the character used is one other than the actual Ozbek one but it still needs to be dealt with.
This should be in the spellrelax file. Something like:
[
[ ?* [ ʼ -> [ %' | ’ | %` | %´ | %‘ | %ʻ ] | ʼ ] ?* ]
]
Fixed in b9203d5.
$ echo "qo\`mitasi" | apertium -d . uzb-morph
^qo`mitasi/qoʻmita<n><px3sp><nom>/qoʻmita<n><px3sp><nom>+e<cop><aor><p3><pl>/qoʻmita<n><px3sp><nom>+e<cop><aor><p3><sg>$^./.<sent>$
The machine seems to break down words such as
at `.
Maybe it's only when the character used is one other than the actual Ozbek one but it still needs to be dealt with.