Examples of opennlp.model.AbstractModel

opennlp.model.AbstractModel

      int iterations = 100;
      if (args.length > ai) {
        cutoff = Integer.parseInt(args[ai++]);
        iterations = Integer.parseInt(args[ai++]);
      }
      AbstractModel mod;
      if (dict != null) {
        buildDictionary(dict, inFile, cutoff);
      }
      if (sequence) {
        POSSampleSequenceStream ss;

View Full Code Here

    }
    ParserEventTypeEnum etype = null;
    boolean fun = false;
    int ai = 0;
    Dictionary dict = null;
    AbstractModel model = null;


    while (ai < args.length && args[ai].startsWith("-")) {
      if (args[ai].equals("-build")) {
        etype = ParserEventTypeEnum.BUILD;
      }
      else if (args[ai].equals("-attach")) {
        etype = ParserEventTypeEnum.ATTACH;
      }
      else if (args[ai].equals("-chunk")) {
        etype = ParserEventTypeEnum.CHUNK;
      }
      else if (args[ai].equals("-check")) {
        etype = ParserEventTypeEnum.CHECK;
      }
      else if (args[ai].equals("-tag")) {
        etype = ParserEventTypeEnum.TAG;
      }
      else if (args[ai].equals("-fun")) {
        fun = true;
      }
      else if (args[ai].equals("-dict")) {
        ai++;
        dict = new Dictionary(new FileInputStream(args[ai]));
      }
      else if (args[ai].equals("-model")) {
        ai++;
        model = (new SuffixSensitiveGISModelReader(new File(args[ai]))).getModel();
      }
      else {
        System.err.println("Invalid option " + args[ai]);
        System.exit(1);
      }
      ai++;
    }
    HeadRules rules = new opennlp.tools.parser.lang.en.HeadRules(args[ai++]);
    if (fun) {
      Parse.useFunctionTags(true);
    }
    opennlp.model.EventStream es = new ParserEventStream(new ParseSampleStream(new PlainTextByLineStream(new java.io.InputStreamReader(System.in))), rules, etype, dict);
    while (es.hasNext()) {
      Event e = es.next();
      if (model != null) {
        System.out.print(model.eval(e.getContext())[model.getIndex(e.getOutcome())]+" ");
      }
      System.out.println(e);
    }
  }

View Full Code Here

          Collections.addAll(dictTags, posDict.getTags(word)); 
        }
        
        Set<String> modelTags = new HashSet<String>();
        
        AbstractModel posModel = getPosModel();
        
        for  (int i = 0; i < posModel.getNumOutcomes(); i++) {
          modelTags.add(posModel.getOutcome(i));
        }
        
        if (!modelTags.containsAll(dictTags)) {
          throw new InvalidFormatException("Tag dictioinary contains tags " +
              "which are unkown by the model!");

View Full Code Here


    String lang = args[1];
    String packageName = args[2];
    String modelName = args[3];


    AbstractModel chunkerModel = new GenericModelReader(
        new BinaryFileDataReader(new FileInputStream(modelName))).getModel();


    ChunkerModel packageModel = new ChunkerModel(lang, chunkerModel);
    packageModel.serialize(new FileOutputStream(packageName));
  }

View Full Code Here


    Map<String, String> manifestInfoEntries = new HashMap<String, String>();


    EventStream es = new ChunkerEventStream(in, factory.getContextGenerator());


    AbstractModel maxentModel = TrainUtil.train(es, mlParams.getSettings(),
        manifestInfoEntries);


    return new ChunkerModel(lang, maxentModel, manifestInfoEntries, factory);
  }

View Full Code Here

    
    Map<String, String> manifestInfoEntries = new HashMap<String, String>();
    
    EventStream es = new ChunkerEventStream(in, contextGenerator);
    
    AbstractModel maxentModel = TrainUtil.train(es, mlParams.getSettings(), manifestInfoEntries);
    
    return new ChunkerModel(lang, maxentModel, manifestInfoEntries);
  }

View Full Code Here

        }
        else {
          es = new RealBasicEventStream(new PlainTextByLineDataStream(datafr));
        }
        GIS.SMOOTHING_OBSERVATION = SMOOTHING_OBSERVATION;
        AbstractModel model;
        if (type.equals("maxent")) {
        
          if (!real) {
            model = GIS.trainModel(es,USE_SMOOTHING);
          }

View Full Code Here


      File outputFile = new File(modelFileName);


      AbstractModelWriter writer;


      AbstractModel model;
      if (type.equals("maxent")) {
  GIS.SMOOTHING_OBSERVATION = SMOOTHING_OBSERVATION;


        if (!real) {
          model = GIS.trainModel(es, maxit, cutoff, sigma);

View Full Code Here

    
    // build
    System.err.println("Training builder");
    opennlp.model.EventStream bes = new ParserEventStream(parseSamples, rules, ParserEventTypeEnum.BUILD, mdict);
    Map<String, String> buildReportMap = new HashMap<String, String>();
    AbstractModel buildModel = TrainUtil.train(bes, mlParams.getSettings("build"), buildReportMap);
    mergeReportIntoManifest(manifestInfoEntries, buildReportMap, "build");
    
    parseSamples.reset();
    
    // tag
    POSModel posModel = POSTaggerME.train(languageCode, new PosSampleStream(parseSamples), 
        mlParams.getParameters("tagger"), null, null);
    
    parseSamples.reset();
    
    // chunk
    ChunkerModel chunkModel = ChunkerME.train(languageCode, 
        new ChunkSampleStream(parseSamples),
        new ChunkContextGenerator(), mlParams.getParameters("chunker"));
    
    parseSamples.reset();
    
    // check
    System.err.println("Training checker");
    opennlp.model.EventStream kes = new ParserEventStream(parseSamples, rules, ParserEventTypeEnum.CHECK);
    Map<String, String> checkReportMap = new HashMap<String, String>();
    AbstractModel checkModel = TrainUtil.train(kes, mlParams.getSettings("check"), checkReportMap);
    mergeReportIntoManifest(manifestInfoEntries, checkReportMap, "check");


    // TODO: Remove cast for HeadRules
    return new ParserModel(languageCode, buildModel, checkModel,
        posModel, chunkModel, (opennlp.tools.parser.lang.en.HeadRules) rules,

View Full Code Here

    // build
    System.err.println("Training builder");
    opennlp.model.EventStream bes = new ParserEventStream(parseSamples, rules,
        ParserEventTypeEnum.BUILD, mdict);
    Map<String, String> buildReportMap = new HashMap<String, String>();
    AbstractModel buildModel = TrainUtil.train(bes, mlParams.getSettings("build"), buildReportMap);
    opennlp.tools.parser.chunking.Parser.mergeReportIntoManifest(manifestInfoEntries, buildReportMap, "build");
    
    parseSamples.reset();
    
    // check
    System.err.println("Training checker");
    opennlp.model.EventStream kes = new ParserEventStream(parseSamples, rules,
        ParserEventTypeEnum.CHECK);
    Map<String, String> checkReportMap = new HashMap<String, String>();
    AbstractModel checkModel = TrainUtil.train(kes, mlParams.getSettings("check"), checkReportMap);
    opennlp.tools.parser.chunking.Parser.mergeReportIntoManifest(manifestInfoEntries, checkReportMap, "check");
    
    parseSamples.reset();
    
    // attach 
    System.err.println("Training attacher");
    opennlp.model.EventStream attachEvents = new ParserEventStream(parseSamples, rules,
        ParserEventTypeEnum.ATTACH);
    Map<String, String> attachReportMap = new HashMap<String, String>();
    AbstractModel attachModel = TrainUtil.train(attachEvents, mlParams.getSettings("attach"), attachReportMap);
    opennlp.tools.parser.chunking.Parser.mergeReportIntoManifest(manifestInfoEntries, attachReportMap, "attach");
    
    // TODO: Remove cast for HeadRules
    return new ParserModel(languageCode, buildModel, checkModel,
        attachModel, posModel, chunkModel,

View Full Code Here

0 1 2 3 4 5 6

TOP

Related Classes of opennlp.model.AbstractModel

CreateModel

opennlp.maxent.GISModel

opennlp.maxent.io.SuffixSensitiveGISModelReader

opennlp.maxent.MaxentPrepAttachTest

opennlp.maxent.ModelTrainer

opennlp.maxent.quasinewton.QNTrainerTest

opennlp.perceptron.PerceptronModel

opennlp.perceptron.PerceptronPrepAttachTest

opennlp.tools.chunker.ChunkerME

opennlp.tools.chunker.ChunkerModel

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.