PAGE helper functions in recognize to generateDS API?

[kba]

Now that the generateDS API has been refactored to make it easier to extend, IMHO it would be useful to have these functions available for all processors:

• page_element_unicode0
• page_element_float0
• page_get_reading_order
• page_update_higher_testequiv_level

[bertsky]

Agreed!

• page_element_unicode0
• page_element_conf0

Maybe these could go as member functions get_Unicode0 and get_conf0 into GlyphType, WordType, TextLineType and TextRegionType.

• page_get_reading_order

I use this a lot, but it could be better: When in ocrd_page_generateds, then the function should

• be named get_reading_order_dict or similar (as member of PageType)
• include instantiating the first/top-level dict
• include referencing the top-level get_ReadingOrder() and its get_OrderedGroup() or get_UnorderedGroup() (all robust to empty results)

• page_update_higher_testequiv_level

Maybe we could trigger this automatically whenever a TextEquiv gets added anywhere and/or before serialization. (In a similar spirit to planned automatic coordinate sanitation.)

Anyway, the version here is the most complete so far, but it could be simplified with the new API in core.

I should also mention: page_add_to_reading_order