Examples of edu.umd.cloud9.io.map.HMapIFW

edu.umd.cloud9.io.map.HMapIFW

Writable representing a map where keys are ints and values are floats.

One notable feature of this class is the ability to support lazy decoding, controlled by the {@link #setLazyDecodeFlag(boolean)} method. In lazydecoding mode, when an object of this type is deserialized, key-value pairs are not inserted into the map, but rather held in arrays. The reduces memory used in cases where random access to values is not required. In lazy decoding mode, the raw keys and values may be fetched by the {@link #getKeys()} and{@link #getValues()} methods, respectively. The map can be subsequentlypopulated with the {@link #decode()} method.
@author Jimmy Lin

   *     mapping from F-terms to their df values
   * @return
   *     mapping from E-terms to their computed df values
   */
  public static HMapIFW translateDFTable(Vocab eVocabSrc, Vocab fVocabTrg, TTable_monolithic_IFAs e2f_probs, HMapSIW dfs){
    HMapIFW transDfTable = new HMapIFW();
    for(int e=1;e<eVocabSrc.size();e++){
      int[] fS = null;
      try {
        fS = e2f_probs.get(e).getTranslations(0.0f);
      } catch (Exception e1) {
        e1.printStackTrace();
      }
      float df=0;
      for(int f : fS){
        float probEF = e2f_probs.get(e, f);
        String fTerm = fVocabTrg.get(f);
        if(!dfs.containsKey(fTerm)){  //only if word is in the collection, can it contribute to the df values.
          continue;
        }      
        float df_f = dfs.get(fTerm);
        df+=(probEF*df_f);
      }
      transDfTable.put(e, df);
    }
    return transDfTable;
  }

View Full Code Here

    Tokenizer tokenizer = TokenizerFactory.createTokenizer(eLang, 
        eTokenizerModelFile, true, eStopwordsFile, eStopwordsFile + ".stemmed", null);


    // translate doc texts here
    for (HMapSIW deDoc : docs) {
      HMapIFW tfS = new HMapIFW();
      int docLen = 0;
      try {
        docLen = CLIRUtils.translateTFs(deDoc, tfS, eVocabSrc, eVocabTrg, fVocabSrc,
            fVocabTrg, e2f_Probs, f2e_Probs, tokenizer , null);   // tokenizer just for stopword list
      } catch (IOException e) {

View Full Code Here

      if (numTerms < MIN_SIZE) {
        reporter.incrCounter(Docs.SHORT, 1);
        return;
      }


      HMapIFW tfS = new HMapIFW();
      // We simply use the source-language doc length since the ratio of doc length to average doc
      // length is unlikely to change significantly (not worth complicating the pipeline)
      int docLen = CLIRUtils.translateTFs(doc, tfS, eVocabSrc, eVocabTrg, fVocabSrc, fVocabTrg,
          e2f_Probs, f2e_Probs, tokenizer, LOG);

View Full Code Here

      }  


      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);
      for(MapIF.Entry term : transDfTable.entrySet()){
        reporter.incrCounter(DF.TransDf, 1);
        writer.append(new IntWritable(term.getKey()), new FloatWritable(term.getValue()));
      }
      writer.close();
    }

View Full Code Here

    Configuration conf = IntegrationUtils.getBespinConfiguration();
    FileSystem fs = FileSystem.get(conf);


    SequenceFile.Reader reader;
    IntWritable key = new IntWritable();
    HMapIFW map = new HMapIFW();
    WeightedIntDocVector value = new WeightedIntDocVector();


    reader = new SequenceFile.Reader(fs.getConf(),
        SequenceFile.Reader.file(new Path(enwikiEn + "/test_wt-int-doc-vectors/part-00000")));


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(enIntDocVector1, value);


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(enIntDocVector2, value);
    reader.close();
  }

View Full Code Here

    Configuration conf = IntegrationUtils.getBespinConfiguration();
    FileSystem fs = FileSystem.get(conf);


    SequenceFile.Reader reader;
    IntWritable key = new IntWritable();
    HMapIFW map = new HMapIFW();
    WeightedIntDocVector value = new WeightedIntDocVector();


    reader = new SequenceFile.Reader(fs.getConf(),
        SequenceFile.Reader.file(new Path(dewikiEn + "/test_wt-int-doc-vectors/part-00000")));
    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(deIntDocVector1, value);


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(deIntDocVector2, value);
    reader.close();
  }

View Full Code Here

    Configuration conf = IntegrationUtils.getBespinConfiguration();
    FileSystem fs = FileSystem.get(conf);


    SequenceFile.Reader reader;
    IntWritable key = new IntWritable();
    HMapIFW map = new HMapIFW();
    WeightedIntDocVector value = new WeightedIntDocVector();


    reader = new SequenceFile.Reader(fs.getConf(),
        SequenceFile.Reader.file(new Path(galagoIndex + "/test_wt-int-doc-vectors/part-00000")));


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(galagoIntDocVector1, value);


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(galagoIntDocVector2, value);
    reader.close();
  }

View Full Code Here

    FileSystem fs = FileSystem.get(conf);


    SequenceFile.Reader reader;
    IntWritable key = new IntWritable();
    WeightedIntDocVector value = new WeightedIntDocVector();
    HMapIFW map = new HMapIFW();


    reader = new SequenceFile.Reader(fs.getConf(),
        SequenceFile.Reader.file(new Path(opennlpIndex + "/test_wt-int-doc-vectors/part-00000")));
    reader.next(key, value);
    map = value.getWeightedTerms();
    System.out.println("*** top 10 terms ***");
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(opennlpIntDocVector1, value);


    reader.next(key, value);
    map = value.getWeightedTerms();
    System.out.println("*** top 10 terms ***");
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(opennlpIntDocVector2, value);
    reader.close();
  }

View Full Code Here

    Configuration conf = IntegrationUtils.getBespinConfiguration();
    FileSystem fs = FileSystem.get(conf);


    SequenceFile.Reader reader;
    IntWritable key = new IntWritable();
    HMapIFW map = new HMapIFW();
    WeightedIntDocVector value = new WeightedIntDocVector();


    reader = new SequenceFile.Reader(fs.getConf(),
        SequenceFile.Reader.file(new Path(enwikiEn + "/test_wt-int-doc-vectors/part-00000")));


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(enIntDocVector1, value);


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(enIntDocVector2, value);
    reader.close();
  }

View Full Code Here

    Configuration conf = IntegrationUtils.getBespinConfiguration();
    FileSystem fs = FileSystem.get(conf);


    SequenceFile.Reader reader;
    IntWritable key = new IntWritable();
    HMapIFW map = new HMapIFW();
    WeightedIntDocVector value = new WeightedIntDocVector();


    reader = new SequenceFile.Reader(fs.getConf(),
        SequenceFile.Reader.file(new Path(dewikiEn + "/test_wt-int-doc-vectors/part-00000")));
    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(deIntDocVector1, value);


    reader.next(key, value);
    System.out.println("*** top 10 terms ***");
    map = value.getWeightedTerms();
    for ( MapIF.Entry entry : map.getEntriesSortedByValue(10)) {
      System.out.println(entry.getKey() + ": " + entry.getValue());
    }
    verifyIntDocVector(deIntDocVector2, value);
    reader.close();
  }

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of edu.umd.cloud9.io.map.HMapIFW

bak.pcj.IntIterator

edu.umd.cloud9.example.bfs.BfsNodeTest

edu.umd.cloud9.example.pagerank.PageRankNodeTest

edu.umd.cloud9.io.array.ArrayListOfIntsWritable

edu.umd.cloud9.io.fastutil.Int2FloatOpenHashMapWritableTest

edu.umd.cloud9.io.fastutil.Int2IntOpenHashMapWritableTest

edu.umd.cloud9.io.fastutil.String2FloatOpenHashMapWritableTest

edu.umd.cloud9.io.fastutil.String2IntOpenHashMapWritableTest

edu.umd.cloud9.io.map.HMapIDW

edu.umd.cloud9.io.map.HMapIDWTest

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.