core-linnaeus-species-ae.2.5.0.source-code.README.txt Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of jcore-linnaeus-species-ae Show documentation

Uses the Linnaeus software (http://linnaeus.sourceforge.net/) to detect species mentions and map them to NCBI taxonomy IDs. This project contains the UIMA wrapper around the original Linnaeus software but does not include any dictionaries. Since Linnaeus works with different dictionaries (e.g. concrete species name mention recognition, species proxy/clue recognition, ...) each of the different resource has been stored in a project of its own in jcore-projects. The default mention dictionary is stored in the jcore-linneaus-species-dict project, for example.

There is a newer version: 2.6.1

Show newest version

All required resources - i.e. Linnaeus configuration, dictionary, frequency
statistics etc. - are packaged in a project of their own. This is done
because there are multiple, different dictionaries available, e.g. for pure
species name mentions or additionally for species hints (like "patient" hints
for a human).
Each such resource project brings with it a pre-configured analysis engine
descriptor that uses the resource of the respective project. Since the
resources are stored in distinct packages, different configurations may
be used in the same pipeline without resource collision problems.
However, this would cause a lot of duplicate annotations since there is no
merging algorithm.