use cases for ocrd-merge processor, e.g.

• simply renaming segment IDs (to make them valid or more readable) – by running in a single input fileGrp:

    ocrd-merge -I OCR-D-OLD-IDs -O OCR-D-NEW-IDs
  

• trivially overlaying GT and prediction (showing both reading orders in a top unordered group)

• recombining pages that have been split up by their Border into fileGrps for left and right, e.g.

    ocrd-anybaseocr-crop -P rulerAreaMax 0 -P marginLeft 0.2 -P marginRight 0.4 -I ORIG -O CROP-L
    ocrd-anybaseocr-crop -P rulerAreaMax 0 -P marginLeft 0.6 -P marginRight 0.8 -I ORIG -O CROP-R
    ocrd-paddleocr-segment -I CROP-L -O SEG-L ...
    ocrd-paddleocr-segment -I CROP-R -O SEG-R ...
    ocrd-tesserocr-recognize -I SEG-L -O OCR-L ...
    ocrd-tesserocr-recognize -I SEG-R -O OCR-R ...
    ocrd-merge -I OCR-L,OCR-R -O OCR
  

• recombining pages that have been split up into fileGrps to reduce image size (say 2x2 patches of huge tables or newspapers)

For out-of-the-box UX and testing, would it make sense to also bundle (some? all?) of the page processing scripts you developed for ocrd-command?

Yes. I'll add

• a respective addition to the readme
• a few preset jsons for the builtin processors
• a few tests for the new builtin processors

Is this sufficient documentation in your opinion, @kba?

(Perhaps the actual workflow recipes should go into the workflow guide – there is only so much you can do here without mentioning/depending on other tools and actual problems...)

Is this sufficient documentation in your opinion, @kba?

(Perhaps the actual workflow recipes should go into the workflow guide – there is only so much you can do here without mentioning/depending on other tools and actual problems...)

Yes, excellent, many thanks. I think the list of presets are a good starting point, they are documented and bundled. Of course, the workflow guide sorely needs an update and would benefit from this but should not be prereq for this PR.

Merging and releasing now.