Examples of edu.umd.cloud9.webgraph.data.AnchorText

edu.umd.cloud9.io.pair.PairOfIntLong

This data structure represents a line of anchor text. A line of anchor text has some text, a weight, and a set of sources (targets) associated with it. Sources (targets) are the pages a line of anchor text originates from (points to) when the underlying link is an incoming (outgoing) link.

The implemented iterator makes it possible to iterate through the source or target documents for each line of anchor text.
@author Nima Asadi

          continue;
        }
        prob = ttable.get(srcIndex, trgIndex);
        logger.debug("Found: " + trgTerm + " with " + prob);


        topTrans.add(new PairOfFloatString(prob, trgTerm));
        // keep top numTrans translations
        if (topTrans.size() > numTrans) {
          float removedProb = topTrans.pollFirst().getLeftElement();
          sumOfProbs -= removedProb;
        }

View Full Code Here

    List<Integer> sortedIndices = new ArrayList<Integer>();
    HMapIF index2ProbMap = new HMapIF();


    float sumOfProbs = 0.0f;    //only extract the top K<15 if the mass prob. exceeds MAX_probThreshold
    while(!topTrans.isEmpty() && sumOfProbs < cumProbThreshold){
      PairOfFloatString e = topTrans.pollLast();
      String term = e.getRightElement();
      float pr = e.getLeftElement()/cumProb;    // normalize
      logger.debug(term+"-->"+pr);
      int trgIndex = trgVocab.addOrGet(term);
      sumOfProbs += e.getLeftElement();         // keep track of unnormalized cumulative prob for determining cutoff
      sortedIndices.add(trgIndex);
      index2ProbMap.put(trgIndex, pr);
    }


    // to enable faster access with binary search, we sort entries by vocabulary index.

View Full Code Here

          int e2 = eVocabTrg.get(eTerm);         


          float prob2 = f2e_Probs.get(f2, e2);
          float prob = prob1*prob2;
          sumOfProbs += prob;
          topTrans.add(new PairOfFloatString(prob, fTerm));
        }
        logger.info("Adding "+eTerm);
        addToTable(e1, topTrans, sumOfProbs, table, fVocabTrg, 1.0f, stats);      
      }
      logger.info(stats);

View Full Code Here

            curIndex = prevIndex;    // revert curIndex value since we're skipping this one
            skipTerm = true;
            continue;
          }
          logger.debug("Processing: "+srcTerm+" with index: "+curIndex);      
          topTrans.add(new PairOfFloatString(prob, trgTerm));
          sumOfProbs += prob;
          logger.debug("Added to queue: "+trgTerm+" with prob: "+prob+" (sum: "+sumOfProbs+")");      
        }else if(!earlyTerminate && !skipTerm && !delims.contains(srcTerm)){  //continue adding translation term,prob pairs (except if early termination is ON)
          topTrans.add(new PairOfFloatString(prob, trgTerm));
          sumOfProbs += prob;
          logger.debug("Added to queue: "+trgTerm+" with prob: "+prob+" (sum: "+sumOfProbs+")");      


          // keep top numTrans translations
          if(topTrans.size() > numTrans){
            PairOfFloatString pair = topTrans.pollFirst();
            float removedProb = pair.getLeftElement();
            sumOfProbs -= removedProb;
            logger.debug("Removed from queue: "+pair.getRightElement()+" (sum: "+sumOfProbs+")");      
          }
        }else{
          logger.debug("Skipped line: "+line);
        }
        //        // line processed: check if early terminate

View Full Code Here

      logger.debug("Processing: "+srcTerm+" with index: "+curIndex+" ("+srcIndex+")");
      for(int trgIndex : translations){
        trgTerm = trgVocab.get(trgIndex);
        prob = ttable.get(srcIndex, trgIndex);


        topTrans.add(new PairOfFloatString(prob, trgTerm));
        // keep top numTrans translations
        if(topTrans.size() > numTrans){
          float removedProb = topTrans.pollFirst().getLeftElement();
          sumOfProbs -= removedProb;
        }

View Full Code Here

    if (featSet > 2) {
      // uppercase token matching features : find uppercased tokens that exactly appear on both sides
      // lack of this evidence does not imply anything, but its existence might indicate parallel
//      fSentence.replaceAll("([',:;.?%!])", " $1 ");
//      eSentence.replaceAll("([',:;.?%!])", " $1 ");
      PairOfFloats pair = getUppercaseRatio(fTokenizer.processContent(fSentence), eTokenizer.processContent(eSentence));
      features.add("uppercaseratio1=" + pair.getLeftElement() );
      features.add("uppercaseratio2=" + pair.getRightElement() );
    }


    if (featSet > 3) {
      // future work = count number of single/double letter words in src and trg side

View Full Code Here

    // now, read tokens in first sentence and keep track of sequences of uppercased tokens in buffer
    HashSet<String> upperCaseMap1 = getUppercaseParts(tokens1);
    HashSet<String> upperCaseMap2 = getUppercaseParts(tokens2);
    float cntUpperRatio1 = getRatio(upperCaseMap1, upperCaseMap2);
    float cntUpperRatio2 = getRatio(upperCaseMap2, upperCaseMap1);
    PairOfFloats result = new PairOfFloats(cntUpperRatio1, cntUpperRatio2);
    return result;
  }

View Full Code Here

                String term = m2.group(1);
                if ( !term.equals("NULL") ) {
                  float prob = Float.parseFloat(m2.group(2));
                  int engIndex = trgVocab.addOrGet(term);
                  logger.debug("Added: "+term+" with index: "+engIndex+" and prob:"+prob);
                  indexProbPairs.add(new PairOfIntFloat(engIndex, prob));
                  sumOfProbs += prob;
                }
              }
            }
            // if number of translations not set, we never cut-off, so all cases are long tails

View Full Code Here

    @Override
    public void reduce(Text key, Iterable<PairOfIntLong> values, Context context)
        throws IOException, InterruptedException {
      String term = key.toString();
      Iterator<PairOfIntLong> iter = values.iterator();
      PairOfIntLong p = iter.next();
      int df = p.getLeftElement();
      long cf = p.getRightElement();
      WritableUtils.writeVInt(dfByTermOut, df);
      WritableUtils.writeVLong(cfByTermOut, cf);
      if (iter.hasNext()) {
        throw new RuntimeException("More than one record for term: " + term);
      }

View Full Code Here


      // map from the id back to text
      // sLogger.info("termid: " + key);
      String term = mTermIdMap.getTerm(key.get());
      // sLogger.info("term: " + term);
      PairOfIntLong pair = gs.getStats(term);


      if (pair == null) {
        p.setCf(-1);
        p.setDf(-1);
      } else {
        p.setCf(pair.getRightElement());
        p.setDf(pair.getLeftElement());
      }


      output.collect(key, p);
    }

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of edu.umd.cloud9.webgraph.data.AnchorText

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.