Package ivory.core.data.dictionary

Examples of ivory.core.data.dictionary.DefaultFrequencySortedDictionary

40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
    FileSystem fs = FileSystem.getLocal(new Configuration());
    Path termsFilePath = new Path("etc/trec-index-terms.dat");
    Path termIDsFilePath = new Path("etc/trec-index-termids.dat");
    Path idToTermFilePath = new Path("etc/trec-index-termid-mapping.dat");

    DefaultFrequencySortedDictionary dictionary =
        new DefaultFrequencySortedDictionary(termsFilePath, termIDsFilePath, idToTermFilePath, fs);

    assertEquals(312232, dictionary.size());
    assertEquals("page", dictionary.getTerm(1));
    assertEquals("time", dictionary.getTerm(2));
    assertEquals("will", dictionary.getTerm(3));
    assertEquals("year", dictionary.getTerm(4));
    assertEquals("nikaan", dictionary.getTerm(100000));

    assertEquals(1, dictionary.getId("page"));
    assertEquals(2, dictionary.getId("time"));
    assertEquals(3, dictionary.getId("will"));
    assertEquals(4, dictionary.getId("year"));
    assertEquals(100000, dictionary.getId("nikaan"));
   
    assertEquals(null, dictionary.getTerm(312233));

    Iterator<String> iter = dictionary.iterator();
    assertTrue(iter.hasNext());
    assertEquals("page", iter.next());
    assertTrue(iter.hasNext());
    assertEquals("time", iter.next());
    assertTrue(iter.hasNext());
    assertEquals("will", iter.next());
    assertTrue(iter.hasNext());
    assertEquals("year", iter.next());
    assertTrue(iter.hasNext());

    int cnt = 0;
    for (@SuppressWarnings("unused") String s : dictionary) {
      cnt++;
    }
    assertEquals(dictionary.size(), cnt);

    cnt = 0;
    iter = dictionary.iterator();
    while(iter.hasNext()) {
      cnt++;
      iter.next();
    }
    assertEquals(dictionary.size(), cnt);
  }
View Full Code Here
156
157
158
159
160
161
162
163
164
165
166
    postingsIndex = new IntPostingsForwardIndex(indexPath, fs);
    LOG.info(" - Number of terms: " + readCollectionTermCount());
    LOG.info("Done!");

    try {
      termidMap = new DefaultFrequencySortedDictionary(new Path(getIndexTermsData()),
          new Path(getIndexTermIdsData()), new Path(getIndexTermIdMappingData()), fs);
    } catch (Exception e) {
      throw new ConfigurationException("Error initializing dictionary!");
    }
View Full Code Here
153
154
155
156
157
158
159
160
161
162
163
    postingsIndex = new IntPostingsForwardIndex(indexPath, fs);
    LOG.info(" - Number of terms: " + readCollectionTermCount());
    LOG.info("Done!");

    try {
      termidMap = new DefaultFrequencySortedDictionary(new Path(getIndexTermsData()),
          new Path(getIndexTermIdsData()), new Path(getIndexTermIdMappingData()), fs);
    } catch (Exception e) {
      throw new ConfigurationException("Error initializing dictionary!");
    }
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
        LOG.info(" - id: " + pathMapping.get(termidsFile));
        LOG.info(" - idToTerms: " + pathMapping.get(idToTermFile));
        LOG.info(" - df data: " + pathMapping.get(dfFile));

        try{
          dict = new DefaultFrequencySortedDictionary(pathMapping.get(termsFile),
              pathMapping.get(termidsFile), pathMapping.get(idToTermFile), FileSystem.getLocal(conf));
          dfTable = new DfTableArray(pathMapping.get(dfFile), FileSystem.getLocal(conf));
        } catch (Exception e) {
          e.printStackTrace();
          throw new RuntimeException("Error loading Terms File for dictionary from "+localFiles[0]);
View Full Code Here
428
429
430
431
432
433
434
435
436
437
438
        en2DeProbs = new TTable_monolithic_IFAs(fs2, new Path(e2fttableFile), true);
      } catch (IOException e) {
        e.printStackTrace();
      } 

      DefaultFrequencySortedDictionary dict = new DefaultFrequencySortedDictionary(new Path(env.getIndexTermsData()), new Path(env.getIndexTermIdsData()), new Path(env.getIndexTermIdMappingData()), fs2);
      DfTableArray dfTable = new DfTableArray(new Path(dfByIntFile), fs2);

      HMapIFW transDfTable = CLIRUtils.translateDFTable(eVocab_e2f, fVocab_e2f, en2DeProbs, dict, dfTable);

      SequenceFile.Writer writer = SequenceFile.createWriter(fs2, conf, new Path(transDfFile), IntWritable.class, FloatWritable.class);
View Full Code Here
207
208
209
210
211
212
213
214
    eScoreFn = (ScoringModel) new Bm25();
    eScoreFn.setAvgDocLength(lang2AvgSentLen.get(eLang));        //average sentence length = heuristic based on De-En data
    eScoreFn.setDocCount(env.readCollectionDocumentCount());

    dict = new DefaultFrequencySortedDictionary(new Path(env.getIndexTermsData()), new Path(env.getIndexTermIdsData()), new Path(env.getIndexTermIdMappingData()), fs);
    dfTable = new DfTableArray(new Path(env.getDfByTermData()), fs);
  }
View Full Code Here
473
474
475
476
477
478
479
480
    eScoreFn = (ScoringModel) new Bm25();
    eScoreFn.setAvgDocLength(lang2AvgSentLen.get(eLang));        //average sentence length = heuristic based on De-En data
    eScoreFn.setDocCount(env.readCollectionDocumentCount());

    dict = new DefaultFrequencySortedDictionary(new Path(env.getIndexTermsData()), new Path(env.getIndexTermIdsData()), new Path(env.getIndexTermIdMappingData()), localFs);
    dfTable = new DfTableArray(new Path(env.getDfByTermData()), localFs);
  }
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
        LOG.info(" - id: " + pathMapping.get(termidsFile));
        LOG.info(" - idToTerms: " + pathMapping.get(idToTermFile));
        LOG.info(" - df data: " + pathMapping.get(dfFile));

        try{
          dict = new DefaultFrequencySortedDictionary(pathMapping.get(termsFile),
              pathMapping.get(termidsFile), pathMapping.get(idToTermFile), FileSystem.getLocal(conf));
          dfTable = new DfTableArray(pathMapping.get(dfFile), FileSystem.getLocal(conf));
        } catch (Exception e) {
          e.printStackTrace();
          throw new RuntimeException("Error loading Terms File for dictionary from "+localFiles[0]);
View Full Code Here
424
425
426
427
428
429
430
431
432
433
434
        en2DeProbs = new TTable_monolithic_IFAs(fs2, new Path(e2fttableFile), true);
      } catch (IOException e) {
        e.printStackTrace();
      } 

      DefaultFrequencySortedDictionary dict = new DefaultFrequencySortedDictionary(new Path(env.getIndexTermsData()), new Path(env.getIndexTermIdsData()), new Path(env.getIndexTermIdMappingData()), fs2);
      DfTableArray dfTable = new DfTableArray(new Path(dfByIntFile), fs2);

      HMapIFW transDfTable = CLIRUtils.translateDFTable(eVocab_e2f, fVocab_e2f, en2DeProbs, dict, dfTable);

      SequenceFile.Writer writer = SequenceFile.createWriter(fs2, conf, new Path(transDfFile), IntWritable.class, FloatWritable.class);
View Full Code Here
204
205
206
207
208
209
210
211
    eScoreFn = (ScoringModel) new Bm25();
    eScoreFn.setAvgDocLength(lang2AvgSentLen.get(eLang));        //average sentence length = heuristic based on De-En data
    eScoreFn.setDocCount(env.readCollectionDocumentCount());

    dict = new DefaultFrequencySortedDictionary(new Path(env.getIndexTermsData()), new Path(env.getIndexTermIdsData()), new Path(env.getIndexTermIdMappingData()), fs);
    dfTable = new DfTableArray(new Path(env.getDfByTermData()), fs);
  }
View Full Code Here
TOP

Related Classes of ivory.core.data.dictionary.DefaultFrequencySortedDictionary

  • ivory.core.data.dictionary.DefaultFrequencySortedDictionaryTest
  • ivory.core.index.DistributeGlobalStatsToPostings$MyMapper
  • ivory.core.preprocess.BuildIntDocVectors$MyMapper
  • ivory.core.preprocess.BuildTranslatedTermDocVectors$DataWriterMapper
  • ivory.core.preprocess.BuildTranslatedTermDocVectors$MyMapperTrans
  • ivory.core.preprocess.BuildWeightedTermDocVectors$MyMapper
  • ivory.core.RetrievalEnvironment
  • ivory.core.util.PrintMostFrequentTerms
  • ivory.lsh.bitext.PreprocessHelper
  • org.apache.hadoop.conf.Configuration
    • Copyright © 2018 www.massapicom. All rights reserved.
    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.