Folia

Bugfix release:
* Fix in set definition parsing

* prevent creating invalid XML NCName for processor IDs

* Updated for FoLiA v2.1.0
* Implemented TextMarkupReference (t-ref)
* Wider range of nesting of structure elements
* Implemented a constraint parser for deep validation against set definition that specify constraint (proycon/folia50)
* Various fixes (e.g. for aliases, no longer lose annotator on text declarations, parsing of linebreaks with IDs, fix in test framework)

* Adapted for FoLiA v2.0.3
* Added ``Document.erase()`` method to erase annotation types (3)
* fixes for FoLiA v1 compatibility and conversion
* More flexible handling in case of missing (404) set definitions (proycon/flat147)
* Implemented JSON serialiation for relations
* Various FQL fixes
* Various other fixes

Bugfix release, implements FoLiA v2.0.2, contains various fixes. Upgrade recommended.
* various FQL fixes, adapted to new PROCESSOR statement (required for FLAT/foliadocserve)
* fix in next() method for querying multiple classes at once
* expanding Provenance data (processors) with a src and format attribute
* handle format attribute on declarations (implementation was missing)
* implemented JSON serialisation of provenance data
* Use AbstractWord interface for Morpheme and Phoneme as intended
* Added Processor.insert() and Provenance.insert()
* Ignore annotator and annotatortype assignments if a processor is already set

Minor update (minor refactoring)