Tooltips and highlighting does not work correctly with Tesseract hOCR output

[stweil]

Compare https://digi.bib.uni-mannheim.de/periodika/reichsanzeiger/ocr/film/abbyy/hocr/001-7920/0007.hocr?overlay=yes with https://digi.bib.uni-mannheim.de/periodika/reichsanzeiger/ocr/film/tesseract-4.0.0-20181201/001-7920/0007.hocr?overlay=yes.

Tooltips and highlighting works with the hOCR file created from ABBYY ALTO (using ocr-transform), but for the Tesseract hOCR it always shows the bbox of the whole image.

[stweil]

Unrelated side note: I used the latest hocrjs from Git to build my own hocr.fullscreen.js. https://unpkg.com/hocrjs still has wrong font scaling.

[zuphilip]

@stweil In your Tesseract hocr I see that the last word has a box filling the whole page:

<span
  class="ocrx_word hocrjs-blank"
  id="word_1_2411"
  title="bbox 0 0 10160 7240; x_wconf 95"
  style="font-family: sans-serif; position: fixed; left: 0px; top: 0px; width: 10161px; height: 7241px;"
  contenteditable="true"> </span>

If you delete that word resp. the whole block it is contained in, then the experience is similar to the ABBYY hocr file.