Open LFDM opened 10 years ago
Made a lot of progress during #172
FYI, I have some thoughts on this but would like to take a little time to review what @lichtr and @Christof have done a bit more closely first.
Don't think there is that much to see yet. @christof implemented a directive + a service in #172, that use information from inside the documentStore to retrieve a language code - and act accordingly for arabic. Needs more generalization, the direction of this development is not clear as of now.
But it will be the retrievers job to extract language information from documents (right now the language service reads directly from the treebank xml, respectively its json conversion, which couples the service to the treebank schema), the language service will then be able to access this information and do whatever it has to do.
@lichtr fixed the morphological attributes in #171 - they differ slightly for Greek and Latin (which has no article for example), but we're right now not reading in any language info to change attribute sets in the morphology plugin - but that will happen during this week.
So... happy to hear your thoughts, @balmas :)
Right now it's a little Latin heavy. We're working on Arabic and on better Greek support, which of course is already working, but lacking a bit in its morphology settings.