v1.5.0

• Upgrade to Jakarta Servlets 6.0
• Require a servlet container compatible with Servlets 6.0, e.g., Tomcat 10
• Require Java 17

v1.4.0

• Update various libraries
• Require Java 11
• Re-training of all existing taggers is required due to major WEKA version update

v1.3.0

• Add WEKA classifier API
• Update various libraries
• Require Java 8

v1.2.1

One bugfix.

• Patch maui.war to turn off unhelpful logging to ./out/learning.log

v1.2.0

More configuration options, and small bugfixes.

• Make probability threshold for suggested tags configurable
• Fix a regression where training/cross-validation jobs with exception were mark
  ed complete
• Don't automatically delete training model when vocabulary is updated

v1.1.0

New cross-validation API, various small improvements and bugfixes.

• Add cross-validation API to compute precision and recall
• Make maximum number of recommendations returned per document configurable
• Add "version" key to / resource
• Bugfix: Problems with tagger ID capitalization (#1)
• Empty text to suggest API now responds 400 instead of 200
• Make logging more informative and less verbose
• Require Java 7

v1.0.1

Minor bugfixes and compatibility improvements.

• Disallow \ and / in tagger ID to avoid a Tomcat issue
• Fixed: Home resource wouldn't work when no trailing slash in URL
• Better error message when uploaded vocabulary is empty
• Updated dependencies

v1.0.0

First stable release.