Read Me

SAMPLE USAGE

java -Xmx1024m -jar scopefinder-1.0-jar-with-dependencies.jar -t negation -i ./resources/sample/input.txt -o ./resources/sample/output.txt

Arguments:
 -f <arg>   The rules files to use (alternatively use default rules file by
            specifying the scope detection type: -t negation or -t speculation)
 -t <arg>   the scope type <negation|speculation> (uses default rules
            files from "./resources/rules/")
 -i <arg>   input file (the input text is split into sentences/tokens using the OpenNLP package)
 -o <arg>   output file (outputs sentences with xml scope tags)

To enable logging add JVM parameter: -Djava.util.logging.config.file=logging.properties



BUILD

Prerequisites: maven, jdk 1.5 or later

Check out the source code: svn co https://scopefinder.svn.sourceforge.net/svnroot/scopefinder scopefinder

Executing "./installLocalJars.sh" will install into your local maven repository OpenNLP and Stanford Parser jars.

To create an executable jar with dependencies run "mvn assembly:assembly"



NOTES

ScopeFinder uses the OpenNLP sentence detector and tokenizer. It also uses the Stanford Parser to parse input text sentences.
Due to Standford parser restrictions, the code can only run single-threaded (one Stanford LexicalizedParser per JVM).
Also, the parser requires increasing the default java heap size to -Xmx1024m.