Package ivory.core.data.dictionary

Examples of ivory.core.data.dictionary.DefaultFrequencySortedDictionary

470
471
472
473
474
475
476
477
    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
28
29
30
31
32
33
34
35
36
37
38
    Path termsFilePath = new Path(env.getIndexTermsData());
    Path termIDsFilePath = new Path(env.getIndexTermIdsData());
    Path idToTermFilePath = new Path(env.getIndexTermIdMappingData());

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

    for (int i=1; i<=200; i++) {
      System.out.println(String.format("%d\t%s\t%d", i, termIDMap.getTerm(i), dfs.getDf(i)));
    }
  }
View Full Code Here
39
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
    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
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
        idToTermFile = idToTermFile.substring(idToTermFile.lastIndexOf("/") + 1);

        LOG.info("Looking for the following files in dcache: " + termsFile + ", " + termidsFile + ", " + idToTermFile);
        // Take a different code path if we're in standalone mode.
        if (conf.get("mapred.job.tracker").equals("local")) {
          dictionary = new DefaultFrequencySortedDictionary(new Path(termsFile),
              new Path(termidsFile), new Path(idToTermFile), FileSystem.getLocal(conf));
        } else {
          // We need to figure out which file in the DistributeCache is which...
          Map<String, Path> pathMapping = Maps.newHashMap();
          Path[] localFiles = DistributedCache.getLocalCacheFiles(context.getConfiguration());
          for (Path p : localFiles) {
            LOG.info("In DistributedCache: " + p);
            if (p.toString().contains(termsFile)) {
              pathMapping.put(termsFile, p);
            } else if (p.toString().contains(termidsFile)) {
              pathMapping.put(termidsFile, p);
            } else if (p.toString().contains(idToTermFile)) {
              pathMapping.put(idToTermFile, p);
            }
          }

          LOG.info(" - terms: " + pathMapping.get(termsFile));
          LOG.info(" - id: " + pathMapping.get(termidsFile));
          LOG.info(" - idToTerms: " + pathMapping.get(idToTermFile));

    String s = localFiles.length + " " + localFiles[0].toString() + " " + localFiles[1].toString() + " " + localFiles[2].toString();
    if (pathMapping.get(termsFile) == null ) {

        throw new RuntimeException(s);
    }

          dictionary = new DefaultFrequencySortedDictionary(pathMapping.get(termsFile),
              pathMapping.get(termidsFile), pathMapping.get(idToTermFile), FileSystem.getLocal(conf));
        }
      } catch (Exception e) {
        e.printStackTrace();
        throw new RuntimeException("Error initializing data!", e);
View Full Code Here
133
134
135
136
137
138
139
140
141
142
143
      LOG.info(" - idToTerms: " + pathMapping.get(idToTermFile));
      LOG.info(" - df data: " + pathMapping.get(dfFile));
      LOG.info(" - dl data: " + pathMapping.get(dlFile));

      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
58
59
60
61
62
63
64
65
66
67
68
        gs.loadDFStats(localFiles[1], fs);
        gs.loadCFStats(localFiles[2], fs);

        String indexPath = job.get("Ivory.IndexPath");
        sLogger.info("loading TermIdMap from " + indexPath);
        mTermIdMap = new DefaultFrequencySortedDictionary(localFiles[3], localFiles[4],
            localFiles[5], fs);
      } catch (IOException e) {
        e.printStackTrace();
        throw new RuntimeException("Error loading global term stats!");
      }
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
    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
208
209
210
211
212
213
214
215
216
217
218
    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);

    //for backward compatibility
    //    String indexTermsFile = localFiles[12].toString();
    //    String dfTableFile = localFiles[0].toString();
View Full Code Here
482
483
484
485
486
487
488
489
    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
153
154
155
156
157
158
159
160
161
162
163
        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
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.