Using AggregatePlaintextFastUMLSProcessor is much faster than AggregatePlainTextProcessor, so I suggest that to start with you just use AggregatePlaintextFastUMLSProcessor.
Do you mean it is taking ~5 hours for a single file to be processed at times, or is that for a set of files? If your JVM heap space is not set large enough, you can get very slow results. Try increasing to 5G (or more) using the JVM parameter -Xmx5G For faster start up, you can also set the -Xms to the same or something close to -Xmx value. -- James On Wed, Dec 13, 2017 at 7:04 PM, Yadav, Harish <[email protected]> wrote: > Hi All, > > > > When the medical records are run with the AE as > AggregatePlaintextFastUMLSProcessor or AggregatePlainTextProcessor the > processing is very slow. It is pretty fast when the smaller files (~2 kb) > are fed as input but when I am processing with bigger files say, 2Mb, it is > very slow and the files are taking ~5 hours to process. Any pointer will be > of great help. > > > > Regards, > > Harish. >
