Semafore : shallow semantic parser
Having Semafore running seems quite complicated; hence I'm explaing what I have done so far stepwise. The first step is having the required files downloaded and untared:
3- models
Once all these are pasted into 1 folder, there is one file to be modified. Environment variables must set manually such as SEMAFOR_HOME, MST_PARSER_HOME, MST_MACHINE, JAVA_HOME_BIN, MODEL_DIR.
For compiling the code , we use ./release/cleanAndCompile.sh . For starting the MNST server, ./relase/startMSTServer.sh . For running an example , ./release/fnParserDriver.sh sample.txt output.txt .
it requires 8gb memory and gives me Heapsize error as I'm running linux on 3gb ; so I failed at the startMSTserver.
I also tried to modify the javacode file to lessen the min heapsize for the code but it didn't work ! here is the file to manipulate startMSTserver
this line :${JAVA_HOME_BIN}/java -classpath ".:./lib/trove.jar:./lib/mallet.jar:./lib/mallet-deps.jar" -Xms8g -Xmx8g \