Changelogs » Medcat

PyUp Safety actively tracks 282,845 Python packages for vulnerabilities and notifies you when to upgrade.



Negative examples are shown in documents, so annotation projects that use the same underlying model (i.e. standard NLP double annotation projects)
  Bug fixes / more documentation


Patch release for mainly bug fixes, small improvements.


Bug fixes, documentation, small improvements


Patch release for lots of bug fixes, improvements, small additional features and more documentation.


Major release of MedCATTrainer since the 0.1 prototype release.
  The entire app has been rewritten, to support:
  - A user permission model, annotors
  - Datasets and project models allow for simultaneous annotation projects with multiple distinct users to be  permissioned correctly with one shared deployment.
  - New meta task and configurable value models.
  - Entirely new UI with NHS style theme / colours.
  - Synonyms for concepts can be searched within a CDB and added directly to the document.
  - ReSTful API for all functions the UI performs
  - Bulk creating datasets, projects, users can be performed from an http client.
  - Downloading of multiple projects worth of annotations is now possible.
  - Infra improvements include dockerization with nginx to front the django web server.


Added support for specifying CUI filter list on start up to use when reporting back concepts.


Key updates include:
  - update MedCAT version to 0.3.x
  - added support for reporting meta-annotations

MedCAT is a Medical Concept Annotation Tool that uses unsupervised machine learning to recognise and link medical concepts with clinical terminologies such as UMLS. MedCAT, like similar tools, uses a concept database to find and link concept mentions inside of biomedical documents. In addition, it has disambiguation, spell-checking and the option for supervised and active learning for improved disambiguation.


Prototype release of MedCATTrainer.
  The 1.0 release plans to rebuild large portions of the app for production deployment.


Initial stable version release.