/*
* ProfilePRs.java
*
* Copyright (c) 1995-2012, The University of Sheffield. See the file
* COPYRIGHT.txt in the software or at http://gate.ac.uk/gate/COPYRIGHT.txt
*
* This file is part of GATE (see http://gate.ac.uk/), and is free
* software, licenced under the GNU Library General Public License,
* Version 2, June 1991 (in the distribution as file licence.html,
* and also available at http://gate.ac.uk/gate/licence.html).
*
* Kalina Bontcheva, 04/10/2001
*
* $Id: ProfilePRs.java 17886 2014-04-20 09:46:53Z markagreenwood $
*/
package gate.creole;
import java.io.File;
import java.util.*;
import gate.*;
import gate.creole.gazetteer.DefaultGazetteer;
import gate.creole.orthomatcher.OrthoMatcher;
import gate.creole.splitter.SentenceSplitter;
import gate.creole.tokeniser.DefaultTokeniser;
import gate.util.GateException;
import gate.util.Out;
import gate.util.profile.Profiler;
//import java.text.NumberFormat;
/**
* This class provides a main function that:
* <UL>
* <LI>
* initialises the GATE library, and creates all PRs
* <LI>
* takes a directory name as argument
* <LI>
* for each .html file in that directory:
* <BR> create a GATE document from the file
* <BR> run the PRs on the document
* <BR> dump some statistics in the end
* </UL>
*/
public class ProfilePRs {
/** String to print when wrong command-line args */
private static String usage =
"usage: ProfilePRs [-dir directory-name | file(s)]";
private static double totalDocLength = 0;
private static int docs = 0;
private static Profiler prof = new Profiler();
private static double maxDocLength = 0;
/** Main function */
public static void main(String[] args) throws Exception {
// say "hi"
Out.prln("processing command line arguments");
// check we have a directory name or list of files
List<File> inputFiles = null;
if(args.length < 1) throw new GateException(usage);
if(args[0].equals("-dir")) { // list all the files in the dir
if(args.length < 2) throw new GateException(usage);
File dir = new File(args[1]);
File[] filesArray = dir.listFiles();
if(filesArray == null)
throw new GateException(
dir.getPath() + " is not a directory; " + usage
);
inputFiles = Arrays.asList(filesArray);
} else { // all args should be file names
inputFiles = new ArrayList<File>();
for(int i = 0; i < args.length; i++)
inputFiles.add(new File(args[i]));
}
prof.initRun("Measuring performance on directory " + args[1]);
// prof.enable(false);
// prof.enableGCCalling(false);
// initialise GATE
prof.checkPoint("Before GATE.init()");
Gate.init();
//tell GATE we're in batch mode
// gate.Main.batchMode = true;
// create some processing resources
prof.checkPoint("Before creating the processing resources");
//create a default tokeniser
FeatureMap params = Factory.newFeatureMap();
DefaultTokeniser tokeniser = (DefaultTokeniser) Factory.createResource(
"gate.creole.tokeniser.DefaultTokeniser", params);
prof.checkPoint("Tokeniser initialised");
//create a default gazetteer
params = Factory.newFeatureMap();
DefaultGazetteer gaz = (DefaultGazetteer) Factory.createResource(
"gate.creole.gazetteer.DefaultGazetteer", params);
prof.checkPoint("Gazetteer initialised");
//create a splitter
params = Factory.newFeatureMap();
SentenceSplitter splitter = (SentenceSplitter) Factory.createResource(
"gate.creole.splitter.SentenceSplitter", params);
prof.checkPoint("Sentence splitter initialised");
//create a tagger
params = Factory.newFeatureMap();
POSTagger tagger = (POSTagger) Factory.createResource(
"gate.creole.POSTagger", params);
prof.checkPoint("POSTagger initialised");
//create a grammar
params = Factory.newFeatureMap();
ANNIETransducer transducer = (ANNIETransducer) Factory.createResource(
"gate.creole.ANNIETransducer", params);
prof.checkPoint("Grammars initialised");
//create an orthomatcher
params = Factory.newFeatureMap();
OrthoMatcher orthomatcher = (OrthoMatcher) Factory.createResource(
"gate.creole.orthomatcher.OrthoMatcher", params);
prof.checkPoint("Orthomatcher initialised");
// for each document
// create a gate doc
// set as the document for hte PRs
// run the PRs
// dump output from the doc
// delete the doc
Out.prln("\nLooping on input files list");
Iterator<File> filesIter = inputFiles.iterator();
docs = inputFiles.size();
int fileNo=0;
while(filesIter.hasNext()) {
File inFile = filesIter.next(); // the current file
fileNo++;
// set the source URL parameter to a "file:..." URL string
params.clear();
params.put(Document.DOCUMENT_URL_PARAMETER_NAME, inFile.toURI().toURL().toExternalForm());
params.put(Document.DOCUMENT_ENCODING_PARAMETER_NAME, "");
// create the document
Document doc = (Document) Factory.createResource(
"gate.corpora.DocumentImpl", params
);
totalDocLength += doc.getContent().size().longValue();
if (maxDocLength < doc.getContent().size().longValue())
maxDocLength = doc.getContent().size().longValue();
// set the document param on the PRs
tokeniser.setDocument(doc);
prof.checkPoint("Processing file " + inFile.getPath() +
", #" + fileNo + "/" + docs, new String[0], true, false, false);
tokeniser.execute();
prof.checkPoint("", new String[] {"Tokenizer", "Processing"}, false, false, false);
//run gazetteer
gaz.setDocument(doc);
gaz.execute();
prof.checkPoint("", new String[] {"Gazettier", "Processing"}, false, false, false);
//run splitter
splitter.setDocument(doc);
splitter.execute();
prof.checkPoint("", new String[] {"Splitter", "Processing"}, false, false, false);
//run the tagger
tagger.setDocument(doc);
tagger.execute();
prof.checkPoint("", new String[] {"Tagger", "Processing"}, false, false, false);
//run the transducer
transducer.setDocument(doc);
transducer.execute();
prof.checkPoint("", new String[] {"JAPE grammars", "Processing"}, false, false, false);
// run the orthomatcher
orthomatcher.setDocument(doc);
orthomatcher.execute();
prof.checkPoint("", new String[] {"Orthomatcher", "Processing"}, false, false, false);
// make the doc a candidate for garbage collection
Factory.deleteResource(doc);
} // input files loop
prof.checkPoint("Done!");
totalDocLength = totalDocLength/1024;
Out.prln("\nTotal KBytes processed: " + (long)totalDocLength);
Out.prln("\nMax document size in bytes: " + (long)maxDocLength +
" (" + (long) maxDocLength/1024 + " Kb)");
prof.printCategAvg("Processing", docs, totalDocLength, "kb");
prof.printCategAvg("Tokenizer", docs, totalDocLength, "kb");
prof.printCategAvg("Gazettier", docs, totalDocLength, "kb");
prof.printCategAvg("Splitter", docs, totalDocLength, "kb");
prof.printCategAvg("Tagger", docs, totalDocLength, "kb");
prof.printCategAvg("JAPE grammars", docs, totalDocLength, "kb");
prof.printCategAvg("Orthomatcher", docs, totalDocLength, "kb");
} // main
} // class ProfilePRs
