Ufal-morphodita

-------------------------
- Fix regression in morpho::generate caused in 1.9.1.
- Add check in encode_dictionary that no lemma is repeated.

-------------------------
- Allow loading larger dictionaries by increasing the internal
limit of size of search structures created during loading.
- Export string_piece using correct namespace in src_lib_only/morphodita.h.


Version 1.9.0-prerelease [15 May 16]
------------------------------------
- Use t2t_docsys http://github.com/ufal/t2t_docsys as documentation system.
- Use C++ Builtem http://github.com/ufal/cpp_builtem as build system.
- Add C bindings.
- Remove support for shared library build.
- Change license from LGPL to MPL 2.0.
- Use C++ iostreams instead of C stdio.
- Start using Semantic Versioning http://semver.org/.
- Add REST server using MicroRestD http://github.com/ufal/microrestd.
- Add tagger::tag_analyzed method.
- Add derivator and derivation_formatter classes.

-------------------------
- Add support for OS X 10.7 and later.
- Use UniLib instead of MorphoDiTa specific utf8 class.
- Update Unicode in tokenizer from 6.3.0 to 7.0.0.
- Fix English tokenizer bug 1.
- Use PyTypeObject in Python bindings instead of proxy classes.

-------------------------
- Add strip_lemma_comment and strip_lemma_id tag set converters.

-------------------------
- Allow English negative prefix rules to overlap, allowing
to use also "non-" as negative prefix in additon to "non".
- Allow passing tagger model instead of morphological model
to run_morpho_analyze, run_morpho_cli and run_morpho_generate
binaries.
- Add "--version" option printing current version of binaries.
- Clarify stable/development versions.

-------------------------
- Initial public release.