Package edu.umd.cloud9.io

Examples of edu.umd.cloud9.io.FSLineReader

98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
      }

      if(localFiles != null && localFiles.length > 0){
        samplesMap = new HMapIIW();
        try {
          FSLineReader reader = new FSLineReader(conf.get("Ivory.SampleFile"), FileSystem.get(conf));
          Text t = new Text();
          while(reader.readLine(t)!=0){
            int docno = Integer.parseInt(t.toString());
            sLogger.info(docno + " --> sample");
            samplesMap.put(docno, 1);
          }
          reader.close();
        } catch (IOException e1) {
        }
        sLogger.info(samplesMap.size()+" sampled");
      }else{
        sLogger.info("samples file does not exist");
View Full Code Here
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
      }

      if(localFiles != null && localFiles.length > 0){
        samplesMap = new HMapIIW();
        try {
          FSLineReader reader = new FSLineReader(localFiles[0], FileSystem.getLocal(job));
          Text t = new Text();
          while(reader.readLine(t)!=0){
            int docno = Integer.parseInt(t.toString());
            sLogger.info(docno + " --> sample");
            samplesMap.put(docno, 1);
          }
          reader.close();
        } catch (IOException e1) {
        }
        sLogger.info(samplesMap.size()+" sampled");
      }else{
        sLogger.info("samples file not specified in local cache");
View Full Code Here
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
      }

      if(localFiles != null && localFiles.length > 0){
        samplesMap = new HMapIIW();
        try {
          FSLineReader reader = new FSLineReader(localFiles[0], FileSystem.getLocal(job));
          Text t = new Text();
          while(reader.readLine(t)!=0){
            int docno = Integer.parseInt(t.toString());
            sLogger.info(docno + " --> sample");
            samplesMap.put(docno, 1);
          }
          reader.close();
        } catch (IOException e1) {
        }
        sLogger.info(samplesMap.size()+" sampled");
      }else{
        sLogger.info("samples file not specified in option SampleDocnosFile");
View Full Code Here
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
     
      sampleDocnosFile = job.get("SampleDocnosFile");
      if(sampleDocnosFile!=null){
        samplesMap = new HMapIIW();
        try {
          FSLineReader reader = new FSLineReader(sampleDocnosFile);
          Text t = new Text();
          while(reader.readLine(t)!=0){
            int docno = Integer.parseInt(t.toString());
            samplesMap.put(docno, 1);
          }
          reader.close();
        } catch (IOException e1) {
        }
      }
     
      mDocMapping = new WikipediaDocnoMapping();
View Full Code Here
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
      title2Docno = SequenceFileUtils.readFileIntoMap(new Path(job.get("TitleDocnoFile")));
      sampleDocnosFile = job.get("SampleDocnosFile");
      if(sampleDocnosFile!=null){
        samplesMap = new HMapIIW();
        try {
          FSLineReader reader = new FSLineReader(sampleDocnosFile);
          Text t = new Text();
          while(reader.readLine(t)!=0){
            int docno = Integer.parseInt(t.toString());
            samplesMap.put(docno, 1);
          }
          reader.close();
        } catch (IOException e1) {
        }
        sLogger.info("Loaded "+samplesMap.size()+" samples");
      }else{
        sLogger.info("No sample file read.");
View Full Code Here
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
     
      sampleDocnosFile = job.get("SampleDocnosFile");
      if(sampleDocnosFile!=null){
        samplesMap = new HMapIIW();
        try {
          FSLineReader reader = new FSLineReader(sampleDocnosFile);
          Text t = new Text();
          while(reader.readLine(t)!=0){
            String[] docnos = t.toString().split("\t");
            germandocnos.add(Integer.parseInt(docnos[0]));
            englishdocnos.add(Integer.parseInt(docnos[1]));
          }
          reader.close();
        } catch (IOException e1) {
        }
      }
     
      mDocMapping = new WikipediaDocnoMapping();
View Full Code Here
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
   *     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
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
   *     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
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
   *     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
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
        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
TOP

Related Classes of edu.umd.cloud9.io.FSLineReader

  • 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
    • 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.