| 1 | package org.greenstone.server;
|
|---|
| 2 |
|
|---|
| 3 | import gate.Annotation;
|
|---|
| 4 | import gate.AnnotationSet;
|
|---|
| 5 | import gate.Corpus;
|
|---|
| 6 | import gate.CorpusController;
|
|---|
| 7 | import gate.Document;
|
|---|
| 8 | import gate.Factory;
|
|---|
| 9 | import gate.FeatureMap;
|
|---|
| 10 | import gate.Gate;
|
|---|
| 11 | import gate.GateConstants;
|
|---|
| 12 | import gate.ProcessingResource;
|
|---|
| 13 | import gate.corpora.RepositioningInfo;
|
|---|
| 14 | import gate.creole.ANNIEConstants;
|
|---|
| 15 | import gate.creole.ResourceInstantiationException;
|
|---|
| 16 | import gate.creole.SerialAnalyserController;
|
|---|
| 17 | import gate.util.GateException;
|
|---|
| 18 | import gate.util.Out;
|
|---|
| 19 | import gate.util.persistence.PersistenceManager;
|
|---|
| 20 |
|
|---|
| 21 | import java.io.File;
|
|---|
| 22 | import java.net.URL;
|
|---|
| 23 | import java.util.ArrayList;
|
|---|
| 24 | import java.util.HashSet;
|
|---|
| 25 | import java.util.Iterator;
|
|---|
| 26 | import java.util.Set;
|
|---|
| 27 |
|
|---|
| 28 | import org.apache.log4j.BasicConfigurator;
|
|---|
| 29 |
|
|---|
| 30 | public class GateScanner
|
|---|
| 31 | {
|
|---|
| 32 | private SerialAnalyserController annieController;
|
|---|
| 33 |
|
|---|
| 34 | public GateScanner()
|
|---|
| 35 | {
|
|---|
| 36 | try
|
|---|
| 37 | {
|
|---|
| 38 | BasicConfigurator.configure();
|
|---|
| 39 | Gate.setGateHome(new File("/home/sam/Desktop/Research/Gate5"));
|
|---|
| 40 | Gate.setPluginsHome(new File("/home/sam/Desktop/Research/Gate5/plugins"));
|
|---|
| 41 | Gate.setUserConfigFile(new File("/home/sam/Desktop/Research/Gate5/gate.xml"));
|
|---|
| 42 | Gate.init();
|
|---|
| 43 | Gate.getCreoleRegister().registerDirectories(new URL("file:///home/sam/Desktop/Research/Gate5/plugins/ANNIE"));
|
|---|
| 44 | Gate.getUserConfig().put(GateConstants.DOCUMENT_ADD_SPACE_ON_UNPACK_FEATURE_NAME, new Boolean(false));
|
|---|
| 45 |
|
|---|
| 46 | Out.prln("Initialising ANNIE...");
|
|---|
| 47 |
|
|---|
| 48 | annieController = (SerialAnalyserController) PersistenceManager.loadObjectFromFile(new File(new File(Gate.getPluginsHome(), ANNIEConstants.PLUGIN_DIR), ANNIEConstants.DEFAULT_FILE));
|
|---|
| 49 |
|
|---|
| 50 | Out.prln("...ANNIE loaded");
|
|---|
| 51 | }
|
|---|
| 52 | catch (Exception e)
|
|---|
| 53 | {
|
|---|
| 54 | e.printStackTrace();
|
|---|
| 55 | }
|
|---|
| 56 | }
|
|---|
| 57 |
|
|---|
| 58 | /** Tell ANNIE's controller about the corpus you want to run on */
|
|---|
| 59 | public void setCorpus(Corpus corpus)
|
|---|
| 60 | {
|
|---|
| 61 | annieController.setCorpus(corpus);
|
|---|
| 62 | } // setCorpus
|
|---|
| 63 |
|
|---|
| 64 | /** Run ANNIE */
|
|---|
| 65 | public void execute() throws GateException
|
|---|
| 66 | {
|
|---|
| 67 | Out.prln("Running ANNIE...");
|
|---|
| 68 | annieController.execute();
|
|---|
| 69 | Out.prln("...ANNIE complete");
|
|---|
| 70 | } // execute()
|
|---|
| 71 |
|
|---|
| 72 | public ArrayList<Word> classifyText(String originalText)
|
|---|
| 73 | {
|
|---|
| 74 | ArrayList<Word> words = new ArrayList<Word>();
|
|---|
| 75 | try
|
|---|
| 76 | {
|
|---|
| 77 | // create a GATE corpus and add a document for each command-line
|
|---|
| 78 | // argument
|
|---|
| 79 | Corpus corpus = (Corpus) Factory.createResource("gate.corpora.CorpusImpl");
|
|---|
| 80 | FeatureMap params = Factory.newFeatureMap();
|
|---|
| 81 | params.put("stringContent", "<html><head></head><body>" + originalText + "</body></html>");
|
|---|
| 82 | params.put("preserveOriginalContent", new Boolean(true));
|
|---|
| 83 | params.put("collectRepositioningInfo", new Boolean(true));
|
|---|
| 84 | Document doc = (Document) Factory.createResource("gate.corpora.DocumentImpl", params);
|
|---|
| 85 | corpus.add(doc);
|
|---|
| 86 |
|
|---|
| 87 | // tell the pipeline about the corpus and run it
|
|---|
| 88 | this.setCorpus(corpus);
|
|---|
| 89 | this.execute();
|
|---|
| 90 |
|
|---|
| 91 | AnnotationSet annotations = doc.getAnnotations();
|
|---|
| 92 |
|
|---|
| 93 | int index = 0;
|
|---|
| 94 | for(int j = 0; j < annotations.size(); j++)
|
|---|
| 95 | {
|
|---|
| 96 | Annotation currentAnnotation = annotations.get(j);
|
|---|
| 97 |
|
|---|
| 98 | if(currentAnnotation == null)
|
|---|
| 99 | {
|
|---|
| 100 | continue;
|
|---|
| 101 | }
|
|---|
| 102 | if(currentAnnotation.getType().equals("Token"))
|
|---|
| 103 | {
|
|---|
| 104 | Word w = new Word();
|
|---|
| 105 | w.setIndex(index++);
|
|---|
| 106 | w.setClassification((String)currentAnnotation.getFeatures().get("category"));
|
|---|
| 107 | w.setValue((String)currentAnnotation.getFeatures().get("string"));
|
|---|
| 108 |
|
|---|
| 109 | words.add(w);
|
|---|
| 110 | }
|
|---|
| 111 | }
|
|---|
| 112 | }
|
|---|
| 113 | catch(Exception ex)
|
|---|
| 114 | {
|
|---|
| 115 | ex.printStackTrace();
|
|---|
| 116 | }
|
|---|
| 117 |
|
|---|
| 118 | return words;
|
|---|
| 119 | }
|
|---|
| 120 | } |
|---|
