LanguageMachines/ucto: v0.25

[Ko van der Sloot]

• Added a test for https://github.com/LanguageMachines/ucto/issues/87
• Adapted to latest update in tokconfig-fra (uctodata 0.9)
• Deal with unknown languages (as detected by ucto), using iso-639-3 'und' (https://github.com/LanguageMachines/ucto/issues/86)
  • don't tokenize unknown languages
  • configurable sentence splitter for "und" text
  • added tests
• added code to set the separator (--seperators), so ucto can split on more than just spaces
• migrated test wrapper to Python 3 (was still on 2.7)

[Maarten van Gompel]

• Set up a Dockerfile
• Added build-deps.sh to automatically download, build and install dependencies
• Updated software metadata (codemeta.json) to latest requirements as proposed in CLARIAH
• deprecated options -f and -x, still works but no longer advertised and gives a deprecation notice (https://github.com/LanguageMachines/ucto/issues/88)
• textcat.cfg is now searched for in user config dir as well as global config; also allow running without textcat if the config is missing entirely (same as if not compiled in)
• added support for user-based configuration dirs ($XDG_CONFIG_HOME/ucto), takes precedence over global data dirs