001/** 002 * Copyright (c) 2007-2011, Regents of the University of Colorado 003 * All rights reserved. 004 * 005 * Redistribution and use in source and binary forms, with or without 006 * modification, are permitted provided that the following conditions are met: 007 * 008 * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 009 * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 010 * Neither the name of the University of Colorado at Boulder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. 011 * 012 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 013 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 014 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 015 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE 016 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 017 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 018 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 019 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 020 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 021 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 022 * POSSIBILITY OF SUCH DAMAGE. 023 */ 024package org.cleartk.examples.documentclassification.basic; 025 026import java.io.File; 027 028import org.apache.uima.analysis_engine.AnalysisEngineProcessException; 029import org.apache.uima.collection.CollectionReader; 030import org.apache.uima.fit.component.JCasAnnotator_ImplBase; 031import org.apache.uima.fit.factory.AggregateBuilder; 032import org.apache.uima.fit.factory.AnalysisEngineFactory; 033import org.apache.uima.fit.pipeline.SimplePipeline; 034import org.apache.uima.fit.util.JCasUtil; 035import org.apache.uima.jcas.JCas; 036import org.cleartk.examples.type.UsenetDocument; 037import org.cleartk.ml.CleartkAnnotator; 038import org.cleartk.ml.jar.GenericJarClassifierFactory; 039import org.cleartk.ml.jar.JarClassifierBuilder; 040import org.cleartk.opennlp.tools.SentenceAnnotator; 041import org.cleartk.snowball.DefaultSnowballStemmer; 042import org.cleartk.token.tokenizer.TokenAnnotator; 043import org.cleartk.util.ViewUriUtil; 044import org.cleartk.util.ae.UriToDocumentTextAnnotator; 045import org.cleartk.util.cr.UriCollectionReader; 046 047import com.lexicalscope.jewel.cli.CliFactory; 048import com.lexicalscope.jewel.cli.Option; 049 050/** 051 * Copyright (c) 2012, Regents of the University of Colorado <br> 052 * All rights reserved. <br> 053 * 054 * Illustrates how to run a basic document classification annotator, once it is trained using the 055 * TrainModel class. For a more in-depth example that demonstrates ClearTK best practices including 056 * the use of more sophisticated feature extractors and the evaluation framework refer to the 057 * examples in org.cleartk.examples.document.classification.advanced. 058 * 059 * @author Lee Becker 060 * 061 */ 062public class RunModel { 063 064 public interface Options { 065 @Option( 066 longName = "test-dir", 067 description = "Specify the directory containing the documents to label.", 068 defaultValue = "data/3news-bydate/test") 069 public File getTestDirectory(); 070 071 @Option( 072 longName = "models-dir", 073 description = "specify the directory containing the trained model jar", 074 defaultValue = "target/simple_document_classification/models") 075 public File getModelsDirectory(); 076 } 077 078 public static void main(String[] args) throws Exception { 079 Options options = CliFactory.parseArguments(Options.class, args); 080 081 // //////////////////////////////////////// 082 // Create collection reader to load URIs 083 // //////////////////////////////////////// 084 CollectionReader reader = UriCollectionReader.getCollectionReaderFromDirectory( 085 options.getTestDirectory(), 086 UriCollectionReader.RejectSystemFiles.class, 087 UriCollectionReader.RejectSystemDirectories.class); 088 089 // //////////////////////////////////////// 090 // Create document classification pipeline 091 // //////////////////////////////////////// 092 AggregateBuilder builder = new AggregateBuilder(); 093 094 // Convert URIs in CAS URI View to Plain Text 095 builder.add(UriToDocumentTextAnnotator.getDescription()); 096 097 // NLP pre-processing components 098 builder.add(SentenceAnnotator.getDescription()); // Sentence segmentation 099 builder.add(TokenAnnotator.getDescription()); // Tokenization 100 builder.add(DefaultSnowballStemmer.getDescription("English")); // Stemming 101 102 // Simple document classification annotator 103 builder.add(AnalysisEngineFactory.createEngineDescription( 104 BasicDocumentClassificationAnnotator.class, 105 CleartkAnnotator.PARAM_IS_TRAINING, 106 false, 107 GenericJarClassifierFactory.PARAM_CLASSIFIER_JAR_PATH, 108 JarClassifierBuilder.getModelJarFile(options.getModelsDirectory()))); 109 110 // ////////////////////////////////////////////////////////////////////////////// 111 // Run pipeline and classify documents 112 // ////////////////////////////////////////////////////////////////////////////// 113 SimplePipeline.runPipeline( 114 reader, 115 builder.createAggregateDescription(), 116 AnalysisEngineFactory.createEngineDescription(PrintClassificationsAnnotator.class)); 117 } 118 119 public static class PrintClassificationsAnnotator extends JCasAnnotator_ImplBase { 120 121 @Override 122 public void process(JCas jCas) throws AnalysisEngineProcessException { 123 UsenetDocument document = JCasUtil.select(jCas, UsenetDocument.class).iterator().next(); 124 System.out.println("classified " + ViewUriUtil.getURI(jCas) + " as " + document.getCategory() 125 + "."); 126 } 127 } 128}