I use another copy of [ocrd_tesserocr's page_update_higher_textequiv_levels()](https://github.com/OCR-D/ocrd_tesserocr/blob/master/ocrd_tesserocr/recognize.py#L1295) here, the same I used in ocrd_calamari. I'd like to have this in OCR-D's standard library.