Open bertsky opened 2 years ago
Also: at most segment types, we should convert Tag
, Property
and Link
to something appropriate instead of removing them.
IIRC Transkribus uses these to label lines as "illegible" or "abbreviated" etc. Perhaps we should first make sure we understand the semantics and schema of allowed values before we map to Labels
in PRImA.
It would be really helpful to have an example page from Transkribus which heavily uses these features. (Generally, a regression test would be nice to have...)
Perhaps we should also synchronise with DTABf concordance...
In particular: setting a correct @type
for each predefined @custom
, e.g. paragraph
for poem_lg
or other
for closer
.
The valuable information does not have to be removed. Transforming not just the attributes, but also its recursive
Property
elements intoMetadataItem/Labels/Label
is worthwhile IMO.