PAGE XML contains TextEquiv with empty Unicode

[stweil]

The PAGE XML files contain lots of text regions without text in their TextEquiv and a few text files without text in their TextEquiv:

# Text regions without text.
% git grep "^                <Unicode></Unicode>" | wc -l
   17097
# Text lines without text:
% git grep "^                    <Unicode></Unicode>" | wc -l  
       6

Text from regions with text in lines but without text in the region gets lost when the PAGE XML file is converted to pure text using ocr-transform.

[stweil]

https://github.com/UB-Mannheim/Fibeln also has 11 files which contain text regions without text. Those files where also created using Transkribus. This indicates that it might be a general problem of that software.

[tsmdt]

Solved: a65f5ba91c1ea1be13f9443a96cedbcb249b346d