missing drop-capital in GT

[bertsky]

Again, I do not know if this is systematic:

In weigel_gnothi02_1618, on page phys_0001 region TextRegion_1488379719413_342, a drop-capital is missing in the annotation, i.e. it became part of the adjacent paragraph region. Worse, its (larger) line height spilled over into tl_4, the first TextLine of the region, so it has a height of 369 pixels and overlaps tl_5 through tl_9.

[tboenig]

With fixing in #31 we will fix this too.

It's not a systematic error but also not singular. There are 108 instance of drop-capital in the GT, of which maybe 70% are to be corrected.

Will be modelled according to https://ocr-d.github.io/gt//trans_documentation/lyInitiale.html

[bertsky]

Oh wow, that is a lot!

@tboenig, are you sure you want to use the Relations annotation described in the guidelines? I knever saw this before in GT. Not that I would object from a user perspective!

@kba I think we can support this by adding a validation on drop-capital in the PAGE validator. We need to ensure that a Relations is present then, and uses @type=join. More generally, we could also detect missing drop-capitals via the new geometry validator we discussed (see OCR-D/core#252).