Closed M3ssman closed 8 months ago
Currently, the list of Tesseract language models which require special text orientation, is a hard-coded list in odem_commons.p:71 as RTL_LANGUAGES.
odem_commons.p:71
RTL_LANGUAGES
This doesn't work of course for any self-trained model which is not known forehand.
Please integrate a config option like model_rtl = ara, fas, heb and respect it's values at processing_ocrd.py#get_recognition_level , if set.
model_rtl = ara, fas, heb
processing_ocrd.py#get_recognition_level
Otherwise use the common constant (see above) as default value.
check it
einspunktnull
commented
8 months ago einspunktnull
commented
8 months ago
Description
Currently, the list of Tesseract language models which require special text orientation, is a hard-coded list in
odem_commons.p:71asRTL_LANGUAGES.This doesn't work of course for any self-trained model which is not known forehand.
Please integrate a config option like
model_rtl = ara, fas, heband respect it's values atprocessing_ocrd.py#get_recognition_level, if set.Otherwise use the common constant (see above) as default value.