Table extraction

[kba]

From OCR-D/ocrd_fileformat#46

@kba:

It would be very useful to have a transformation that extracts any tables from PAGE-XML to CSV.

@bertsky:

Thoughts:

• each TableRegion needs its own CSV, so it's not immediately clear how this fits with the page→page converter paradigm
  (e.g. for page→text, one could simply paste the CSV in the middle of the plaintext, but maybe creating a multitude of output files is usually better)
• CSV may already be too coarse (no multi-span, no header distinction)
• perhaps better transfer to ocr-fileformat subrepo?