Package ivory.core.data.dictionary

Examples of ivory.core.data.dictionary.DefaultFrequencySortedDictionary

415
416
417
418
419
420
421
422
423
424
425
        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
71
72
73
74
75
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
        String termidsFile = env.getIndexTermIdsData();
        String idToTermFile = env.getIndexTermIdMappingData();

        // 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));

          dictionary = new DefaultFrequencySortedDictionary(pathMapping.get(termsFile),
              pathMapping.get(termidsFile), pathMapping.get(idToTermFile),
              FileSystem.getLocal(context.getConfiguration()));
        }
      } catch (Exception e) {
        e.printStackTrace();
View Full Code Here
127
128
129
130
131
132
133
134
135
136
137
      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
158
159
160
161
162
163
164
165
166
167
168
        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
420
421
422
423
424
425
426
427
428
429
430
        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
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
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.