ohdsi-nlp-wrapper

This repository contains three separate eclipse projects: clamp-wrapper, ctakes-wrapper, metamap-lite-wrapper; Users can import them into eclipse as existing projects.

download dependencies

To run the wrapper, users must download the dependent libraries and resources and then place them in the subfolder. For example, This is a directory of ctakes wrapper:

ctakes-wrapper

--bin

--input

--output

--lib

----ctakes

------config

------desc

------lib

------resources

--resources

----org

--src

cTAKES wrapper setup

Download the latest cTAKES binary file: http://apache.osuosl.org//ctakes/ctakes-4.0.0/apache-ctakes-4.0.0-bin.zip
Unzip that file
Run mkdir -p ctakes-wrapper/lib/ctakes
Copy lib directory into classpath: cp -r apache-ctakes-4.0.0/lib <ctakes-wrapper-location>/lib/ctakes/
Copy resource files into classpath: cp -r apache-ctakes-4.0.0/resources ctakes-wrapper/lib/ctakes/
Download the official cTAKES UMLS resources: https://downloads.sourceforge.net/project/ctakesresources/sno_rx_16ab.zip
Run mkdir -p ctakes-wrapper/resources/org/apache/ctakes/dictionary/lookup/fast
Unzip that file into the resources directory: unzip sno_rx_16ab.zip -d <ctakes-wrapper-location>/resources/org/apache/ctakes/dictionary/lookup/fast/
Setup your UMLS credentials in this file: <ctakes-wrapper-location>/lib/ctakes/resources/org/apache/ctakes/dictionary/lookup/fast/sno_rx_16ab.xml
- There are two places where you need to replace CHANGE_ME with your UMLS user name and your UMLS password.
Now you should be able to run the class in the ctakes-wrapper project. This version of cTAKES makes use of large machine learning models and should be run with at least 2GB RAM. To do this in Eclipse, in the run configuration dialog, under the Arguments tab, add "-mx2g" to the VM Arguments text field.