Examples of edu.umd.cloud9.collection.DocnoMapping

edu.umd.cloud9.io.map.HMapIFW

Interface for an object that maintains a bidirectional mapping between docids and docnos. A docid is a globally-unique String identifier for a document in the collection. For many types of information retrieval algorithms, documents in the collection must be sequentially numbered; thus, each document in the collection must be assigned a unique integer identifier, which is its docno. Typically, the docid/docno mappings are stored in a mappings file, which is loaded into memory by concrete objects implementing this interface.

Unless there are compelling reasons otherwise, it is preferable to start numbering docnos from one instead of zero. This is because zero cannot be represented in many common compression schemes that are used in information retrieval (e.g., Golomb codes).
@author Jimmy Lin

    qr.runQueries();
    long end = System.currentTimeMillis();


    sLogger.info("Total query time: " + (end - start) + "ms");


    DocnoMapping mapping = qr.getDocnoMapping();


    for (String model : qr.getModels()) {
      sLogger.info("Verifying results of model \"" + model + "\"");


      Map<String, Accumulator[]> results = qr.getResults(model);

View Full Code Here

    qr.runQueries();
    long end = System.currentTimeMillis();


    sLogger.info("Total query time: " + (end - start) + "ms");


    DocnoMapping mapping = qr.getDocnoMapping();


    for (String model : qr.getModels()) {
      sLogger.info("Verifying results of model \"" + model + "\"");


      Map<String, Accumulator[]> results = qr.getResults(model);

View Full Code Here


  }


  static float eval(QueryEngine qe, Configuration conf, String setting){
    Qrels qrels = new Qrels(conf.get(Constants.QrelsPath));
    DocnoMapping mapping = qe.getDocnoMapping();
    float apSum = 0, p10Sum = 0;
    Map<String, Accumulator[]> results = qe.getResults();
    for (String qid : results.keySet()) {
      float ap = (float) RankedListEvaluator.computeAP(results.get(qid), mapping,
          qrels.getReldocsForQid(qid));

View Full Code Here

      return -1;
    }


    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);
    Path mappingFile = env.getDocnoMappingData();
    new ClueWarcDocnoMappingBuilder().build(new Path(collection), mappingFile, conf);


    conf.set(Constants.CollectionName, "ClueWeb:English:Segment" + segment);
    conf.set(Constants.CollectionPath, collection);
    conf.set(Constants.IndexPath, indexPath);
    conf.set(Constants.InputFormat, SequenceFileInputFormat.class.getCanonicalName());

View Full Code Here

          "-output_path=" + indexRootPath + "/wiki-docid-tmp",
          "-output_file=" + mappingFile.toString(),
          "-lang=" + collectionLang };
      LOG.info("Running BuildWikipediaDocnoMapping with args " + Arrays.toString(arr));


      BuildWikipediaDocnoMapping tool = new BuildWikipediaDocnoMapping();
      tool.setConf(conf);
      tool.run(arr);


      fs.delete(new Path(indexRootPath + "/wiki-docid-tmp"), true);
    }


    // Repack Wikipedia into sequential compressed block
    p = new Path(seqCollection);
    LOG.info(seqCollection + " doesn't exist, creating...");
    String[] arr = new String[] { "-input=" + rawCollection,
        "-output=" + seqCollection,
        "-mapping_file=" + mappingFile.toString(),
        "-compression_type=block",
        "-wiki_language=" + collectionLang };
    LOG.info("Running RepackWikipedia with args " + Arrays.toString(arr));


    RepackWikipedia tool = new RepackWikipedia();
    tool.setConf(conf);
    tool.run(arr);


    conf.set(Constants.CollectionName, "Wikipedia-"+collectionLang);
    conf.setInt(Constants.NumMapTasks, numMappers);
    conf.setInt(Constants.NumReduceTasks, numReducers);
    conf.set(Constants.CollectionPath, seqCollection);

View Full Code Here

        "-mapping_file=" + mappingFile.toString(),
        "-compression_type=block",
        "-wiki_language=" + collectionLang };
    LOG.info("Running RepackWikipedia with args " + Arrays.toString(arr));


    RepackWikipedia tool = new RepackWikipedia();
    tool.setConf(conf);
    tool.run(arr);


    conf.set(Constants.CollectionName, "Wikipedia-"+collectionLang);
    conf.setInt(Constants.NumMapTasks, numMappers);
    conf.setInt(Constants.NumReduceTasks, numReducers);
    conf.set(Constants.CollectionPath, seqCollection);

View Full Code Here

   *     FileSystem object
   * @return
   *     mapping from term ids to df values
   */
  public static HMapIFW readTransDfTable(Path path, FileSystem fs) {
    HMapIFW transDfTable = new HMapIFW();
    try {
      SequenceFile.Reader reader = new SequenceFile.Reader(fs, path, fs.getConf());


      IntWritable key = (IntWritable) reader.getKeyClass().newInstance();
      FloatWritable value = (FloatWritable) reader.getValueClass().newInstance();


      while (reader.next(key, value)) {
        transDfTable.put(key.get(), value.get());
        //        logger.info(key.get()+"-->"+value.get());
        key = (IntWritable) reader.getKeyClass().newInstance();
        value = (FloatWritable) reader.getValueClass().newInstance();
      }
      reader.close();

View Full Code Here

   *     ttable E-->F (i.e., Pr(f|e))
   * @return
   *     mapping from E-terms to their computed df values
   */
  public static HMapIFW translateDFTable(Vocab eVocabSrc, Vocab fVocabTrg, TTable_monolithic_IFAs e2f_probs, FrequencySortedDictionary dict, DfTableArray dfTable){
    HMapIFW transDfTable = new HMapIFW();
    for(int e=1;e<eVocabSrc.size();e++){
      int[] fS = e2f_probs.get(e).getTranslations(0.0f);
      float df=0;
      for(int f : fS){
        float probEF = e2f_probs.get(e, f);
        String fTerm = fVocabTrg.get(f);
        int id = dict.getId(fTerm); 
        if(id != -1){
          float df_f = dfTable.getDf(id);        
          df += (probEF*df_f);
        }else{
          logger.debug(fTerm+" not in dict");
        }
      }
      transDfTable.put(e, df);
    }
    return transDfTable;
  }

View Full Code Here

   *     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

        if (e1.getDocno() < mBlockStart)
          continue;
        if (e1.getDocno() >= mBlockEnd)
          break;


        HMapIFW map = new HMapIFW();


        sLogger.debug(key + ": " + e1);


        PostingsReader reader2 = postings.getPostingsReader();


        while (reader2.nextPosting(e2)) {


          sLogger.debug(key + ": " + e1 + ", " + e2);


          if (e1.getDocno() == e2.getDocno())
            continue;


          // compute partial score of similarity for a pair of
          // documents
          float weight = mModel.computeScore(e1.getScore(), e2.getScore(),
              mDocLengthTable.getDocLength(e1.getDocno()), mDocLengthTable
                  .getDocLength(e2.getDocno()));


          map.put(e2.getDocno(), weight);
        }
        output.collect(new IntWritable(e1.getDocno()), map);
      }
    }

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of edu.umd.cloud9.collection.DocnoMapping

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.