2.1.0

Bug fixes.

Install instructions are available on our wiki.

More-comprehensive models are available from https://athena.ahc.umn.edu/biomedicus-downloads/ for verified UMLS licensees.

2.0.1

Small bugfix release because a label property "type" broke UIMA's JCasGen.

Install instructions are available on our wiki.

More-comprehensive models are available from https://athena.ahc.umn.edu/biomedicus-downloads/ for verified UMLS licensees.

2.0.0

• Beta Social History module
• Bug fixes and improvements
• Architecture and framework improvements
• Type-system revamped (auto-adapted types)
• TagEx improvements
• Added UMLS source and SUI to UmlsConcept output.

Install instructions are available on our wiki.

More-comprehensive models are available from https://athena.ahc.umn.edu/biomedicus-downloads/ for verified UMLS licensees.

1.8.4

New features and bug fixes.

Switched models from being stored with MapDB to RocksDB. Added a disk based model for acronyms.
Changed defaults from using in-memory models to using in-disk models. BioMedICUS will now run on less than 2gb of heap size.
Added a numbers detector and a units of measurement detector.
More-comprehensive models are available from http://athena.ahc.umn.edu/biomedicus-downloads/ for verified UMLS licensees.

1.7.0

• Option to use disk-based models for normalization, concepts, part of speech
• Updated Penn Treebank reader

More-comprehensive models are available from http://athena.ahc.umn.edu/biomedicus-downloads/ for verified UMLS licensees.

1.6.0

• Improvements to acronym detection.
• Improvements to tokenization pipeline.
• Improvements to concept detection.
• Label patterns.
• Social history module.
• Numerous performance improvements and bug fixes in other areas.
• Architectural work for future development.

More-comprehensive models are available from http://athena.ahc.umn.edu/biomedicus-downloads/ for verified UMLS licensees.

1.5.0

A large number of updates.

• Acronym detection.
• Concept detection updates, including use of norm vector based model and use of acronym expansions.
• Rtf parsing fixes.
• Type system updates.
• Rule based modification detection (historical, probable, negated) speed improvements.

Documentation and Wiki pages are very much works-in-progress, but I expect to make a lot of progress in July on this specifically.

Developer changes:

• Introduction of a new generic "Annotations"/"Features"/"Labels" system, work on the nomenclature / apis for this is ongoing and major changes may be made.
• Tighter integration and use of Guice for providing dependencies to Document Processors.

Models containing SNOMED CT and other restricted UMLS sources are available for UMLS licensees here.