Package ivory.core.preprocess

Examples of ivory.core.preprocess.ComputeGlobalTermStatistics$MyMapper

91
92
93
94
95
96
97
98
99
100
101
    conf.setInt(Constants.DocnoOffset, 0); // docnos start at 1
    conf.setInt(Constants.MinDf, 10);
    conf.setInt(Constants.MaxDf, Integer.MAX_VALUE);

    new BuildTermDocVectors(conf).run();
    new ComputeGlobalTermStatistics(conf).run();
    new BuildDictionary(conf).run();
    new BuildIntDocVectors(conf).run();

    new BuildIntDocVectorsForwardIndex(conf).run();
    new BuildTermDocVectorsForwardIndex(conf).run();
View Full Code Here
105
106
107
108
109
110
111
112
113
114
115
    conf.setInt("Ivory.MinDf", 2); // toss away singleton terms
    conf.setInt("Ivory.MaxDf", Integer.MAX_VALUE);
    conf.setInt("Ivory.TermIndexWindow", 8);

    new BuildTermDocVectors(conf).run();
    new ComputeGlobalTermStatistics(conf).run();
    new BuildDictionary(conf).run();
    new BuildIntDocVectors(conf).run();

    new BuildIntDocVectorsForwardIndex(conf).run();
    new BuildTermDocVectorsForwardIndex(conf).run();
View Full Code Here
103
104
105
106
107
108
109
110
111
112
113
    conf.setInt(Constants.MinDf, 2); // toss away singleton terms
    conf.setInt(Constants.MaxDf, Integer.MAX_VALUE);
    conf.setInt(Constants.TermIndexWindow, 8);

    new BuildTermDocVectors(conf).run();
    new ComputeGlobalTermStatistics(conf).run();
    new BuildDictionary(conf).run();
    new BuildIntDocVectors(conf).run();

    new BuildIntDocVectorsForwardIndex(conf).run();
    new BuildTermDocVectorsForwardIndex(conf).run();
View Full Code Here
194
195
196
197
198
199
200
201
202
203
204
205
    LOG.info("Job finished in "+(System.currentTimeMillis()-startTime)/1000.0+" seconds");

    // Get CF and DF counts
    startTime = System.currentTimeMillis();
    LOG.info("Counting terms...");
    ComputeGlobalTermStatistics termCountWithDfAndCfTool = new ComputeGlobalTermStatistics(conf);
    termCountWithDfAndCfTool.run();
    LOG.info("TermCount = "+env.readCollectionTermCount()+"\nJob finished in "+(System.currentTimeMillis()-startTime)/1000.0+" seconds");

    // Build a map from terms to sequentially generated integer term ids
    startTime = System.currentTimeMillis();
    conf.setInt("Ivory.TermIndexWindow", TermIndexWindow);
View Full Code Here
103
104
105
106
107
108
109
110
111
112
113
    conf.setInt(Constants.MinDf, 10);
    conf.setInt(Constants.MaxDf, Integer.MAX_VALUE);
    conf.setInt(Constants.TermIndexWindow, 8);

    new BuildTermDocVectors(conf).run();
    new ComputeGlobalTermStatistics(conf).run();
    new BuildDictionary(conf).run();
    new BuildIntDocVectors(conf).run();

    new BuildIntDocVectorsForwardIndex(conf).run();
    new BuildTermDocVectorsForwardIndex(conf).run();
View Full Code Here
93
94
95
96
97
98
99
100
101
102
103
    conf.setInt(Constants.MinDf, 50);
    conf.setInt(Constants.MaxDf, Integer.MAX_VALUE);
    conf.setInt(Constants.TermIndexWindow, 8);

    new BuildTermDocVectors(conf).run();
    new ComputeGlobalTermStatistics(conf).run();
    new BuildDictionary(conf).run();
    new BuildIntDocVectors(conf).run();

    new BuildIntDocVectorsForwardIndex(conf).run();
    new BuildTermDocVectorsForwardIndex(conf).run();
View Full Code Here
97
98
99
100
101
102
103
104
105
106
107
    conf.setInt(Constants.MinDf, 10);
    conf.setInt(Constants.MaxDf, Integer.MAX_VALUE);
    conf.setInt(Constants.TermIndexWindow, 8);

    new BuildTermDocVectors(conf).run();
    new ComputeGlobalTermStatistics(conf).run();
    new BuildDictionary(conf).run();
    new BuildIntDocVectors(conf).run();

    new BuildIntDocVectorsForwardIndex(conf).run();
    new BuildTermDocVectorsForwardIndex(conf).run();
View Full Code Here
103
104
105
106
107
108
109
110
111
112
113
    conf.setInt(Constants.MinDf, 10);
    conf.setInt(Constants.MaxDf, Integer.MAX_VALUE);
    conf.setInt(Constants.TermIndexWindow, 8);

    new BuildTermDocVectors(conf).run();
    new ComputeGlobalTermStatistics(conf).run();
    new BuildDictionary(conf).run();
    new BuildIntDocVectors(conf).run();

    new BuildIntDocVectorsForwardIndex(conf).run();
    new BuildTermDocVectorsForwardIndex(conf).run();
View Full Code Here
232
233
234
235
236
237
238
239
240
241
242
    }

    // Get CF and DF counts.
    startTime = System.currentTimeMillis();
    LOG.info("Counting terms...");
    exitCode = new ComputeGlobalTermStatistics(conf).run();
    LOG.info("TermCount = " + env.readCollectionTermCount());
    if (exitCode >= 0) {
      LOG.info("Job ComputeGlobalTermStatistics finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");
    }else {
      LOG.info("Error: ComputeGlobalTermStatistics. Terminating...");
View Full Code Here
245
246
247
248
249
250
251
252
253
254
255
    }

    // Get CF and DF counts.
    startTime = System.currentTimeMillis();
    LOG.info("Counting terms...");
    exitCode = new ComputeGlobalTermStatistics(conf).run();
    LOG.info("TermCount = " + env.readCollectionTermCount());
    if (exitCode >= 0) {
      LOG.info("Job ComputeGlobalTermStatistics finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");
    }else {
      LOG.info("Error: ComputeGlobalTermStatistics. Terminating...");
View Full Code Here
TOP

Related Classes of ivory.core.preprocess.ComputeGlobalTermStatistics$MyMapper

  • edu.umd.cloud9.collection.clue.ClueWarcForwardIndex
  • edu.umd.cloud9.collection.clue.ClueWarcRecord
  • edu.umd.cloud9.collection.wikipedia.WikipediaPage
  • edu.umd.cloud9.io.array.ArrayListOfIntsWritable
  • edu.umd.cloud9.io.FSLineReader
  • edu.umd.cloud9.io.map.HMapIFW
  • edu.umd.cloud9.io.map.HMapIIW
  • edu.umd.cloud9.io.map.HMapSIW
  • edu.umd.cloud9.io.pair.PairOfIntLong
  • edu.umd.cloud9.io.pair.PairOfInts
    • 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.