To get the update branch running, handling of bboxes from annotations should be improved.
If invalid bboxes are recognized, the bboxes are NOT used to extract text from annotations.
This would lead to expected exceptions in the used libraries.
Instead, ignore those boxes.
This had some issues with the tests, invalid bboxes could not be tested cause they are not extracted.
The internal structure of pdfplumber changed, figures are now images. Therefore tests must be adapted.
To get the update branch running, handling of bboxes from annotations should be improved. If invalid bboxes are recognized, the bboxes are NOT used to extract text from annotations. This would lead to expected exceptions in the used libraries. Instead, ignore those boxes. This had some issues with the tests, invalid bboxes could not be tested cause they are not extracted. The internal structure of pdfplumber changed, figures are now images. Therefore tests must be adapted.