workspace list-page: show label

[bertsky]

Sometimes physical page identifiers look like phys_fcd565cd-d170-4774-bb37-9aa25b68370b – which is impossible to type/memorize. But even names like PHYS_0001 are not as helpful as the @ORDERLABEL and @ORDER attributes.

Since presentation systems usually combine the latter two (@ORDER [@ORDERLABEL] IINM) we should offer something of the sort, too.

So ocrd workspace list-page could use an option (for backwards compatibility) to display these labels in extra columns.

But also, the page range operator could be made aware (somehow) of these labels, allowing me to specify, say -g 20 instead of -g phys_fcd565cd-d170-4774-bb37-9aa25b68370b.

[bertsky]

So ocrd workspace list-page could use an option (for backwards compatibility) to display these labels in extra columns.

We would have to refactor OcrdMets.physical_pages to return the complete mets:div element (instead of just the @ID string), so all attributes can be queried from it. EDIT Oh, I completely forgot about #1063, which already solves that first part (by providing an extra kwarg to get_physical_pages).

For logical div labels, we would still need something like ODEM's _log_type_for_id...

But also, the page range operator could be made aware (somehow) of these labels, allowing me to specify, say -g 20 instead of -g phys_fcd565cd-d170-4774-bb37-9aa25b68370b.

This could actually be as simple as replacing…
https://github.com/OCR-D/core/blob/742906e330d5ef1139fd18c86b73c154c0a67eae/ocrd_models/ocrd_models/ocrd_mets.py#L292C12-L295
…with the extended test…

                for page_num, page in enumerate(self._tree.getroot().xpath('//mets:div[@TYPE="page"]', namespaces=NS)):
                    page_id = page.get('ID')
                    page_nr = page.get('ORDER', str(page_num))
                    if page_id in pageId_patterns or page_nr in pageId_patterns or \
                        any([p.fullmatch(page_id) or p.fullmatch(page_nr) for p in pageId_patterns if isinstance(p, typing.Pattern)]):

[kba]

All of these features are now merged with #1063:

• OcrdMets.get_physical_pages can return the mets:div elements instead of just the ID
• ocrd workspace list-page supports ORDER, ORDERLABEL, LABEL and CONTENTIDS in addition to ID
• ocrd workspace update-page --set allows setting them
• -g ranges can refer to any of them