Examples of ClueWarcForwardIndex

Examples of edu.umd.cloud9.collection.clue.ClueWarcForwardIndex

41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
      String inputFile = conf.get("InputFile");
      String outputFile = conf.get("OutputFile");

      ClueWarcForwardIndex[] indexes = new ClueWarcForwardIndex[10];

      indexes[0] = new ClueWarcForwardIndex();
      indexes[0].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.01.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[1] = new ClueWarcForwardIndex();
      indexes[1].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.02.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[2] = new ClueWarcForwardIndex();
      indexes[2].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.03.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[3] = new ClueWarcForwardIndex();
      indexes[3].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.04.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[4] = new ClueWarcForwardIndex();
      indexes[4].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.05.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[5] = new ClueWarcForwardIndex();
      indexes[5].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.06.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[6] = new ClueWarcForwardIndex();
      indexes[6].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.07.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[7] = new ClueWarcForwardIndex();
      indexes[7].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.08.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[8] = new ClueWarcForwardIndex();
      indexes[8].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.09.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[9] = new ClueWarcForwardIndex();
      indexes[9].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.10.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      FileSystem fs = FileSystem.get(conf);
View Full Code Here

Examples of edu.umd.cloud9.collection.clue.ClueWarcForwardIndex

40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
      String inputFile = conf.get("InputFile");
      String outputFile = conf.get("OutputFile");
      String findexFile = conf.get("ForwardIndexFile");
      String docnoMapping = conf.get("DocnoMappingFile");

      ClueWarcForwardIndex findex = new ClueWarcForwardIndex();
      findex.loadIndex(new Path(findexFile), new Path(docnoMapping), FileSystem.get(conf));

      FileSystem fs = FileSystem.get(conf);

      sLogger.info("reading " + inputFile);

      FSLineReader reader = new FSLineReader(new Path(inputFile), fs);
      FSDataOutputStream writer = fs.create(new Path(outputFile), true);

      Text line = new Text();
      while (reader.readLine(line) > 0) {
        String[] arr = line.toString().split("\\s+");

        String docid = arr[2];
        int rank = Integer.parseInt(arr[3]);

        long start = System.currentTimeMillis();
        String url = findex.getDocument(docid).getHeaderMetadataItem("WARC-Target-URI");
        long duration = System.currentTimeMillis() - start;

        reporter.incrCounter(MyCounter.Count, 1);
        reporter.incrCounter(MyCounter.Time, duration);
View Full Code Here

Examples of edu.umd.cloud9.collection.clue.ClueWarcForwardIndex

132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
        "-collection=" + collectionPathRepacked,
        "-index=" + index };

    IntegrationUtils.exec(Joiner.on(" ").join(args));

    ClueWarcForwardIndex findex = new ClueWarcForwardIndex();
    findex.loadIndex(new Path(index), new Path(mappingFile), fs);

    assertTrue(findex.getDocument(14069750).getContent()
        .contains("Vizergy: How Design and SEO work together"));
    assertTrue(findex.getDocument("clueweb09-en0008-76-19728").getContent()
        .contains("Jostens - Homeschool Yearbooks"));
    assertEquals(1, findex.getFirstDocno());
    assertEquals(50220423, findex.getLastDocno());
  }
View Full Code Here

Examples of edu.umd.cloud9.collection.clue.ClueWarcForwardIndex

41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
      String inputFile = conf.get("InputFile");
      String outputFile = conf.get("OutputFile");

      ClueWarcForwardIndex[] indexes = new ClueWarcForwardIndex[10];

      indexes[0] = new ClueWarcForwardIndex();
      indexes[0].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.01.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[1] = new ClueWarcForwardIndex();
      indexes[1].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.02.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[2] = new ClueWarcForwardIndex();
      indexes[2].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.03.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[3] = new ClueWarcForwardIndex();
      indexes[3].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.04.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[4] = new ClueWarcForwardIndex();
      indexes[4].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.05.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[5] = new ClueWarcForwardIndex();
      indexes[5].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.06.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[6] = new ClueWarcForwardIndex();
      indexes[6].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.07.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[7] = new ClueWarcForwardIndex();
      indexes[7].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.08.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[8] = new ClueWarcForwardIndex();
      indexes[8].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.09.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      indexes[9] = new ClueWarcForwardIndex();
      indexes[9].loadIndex(new Path("/shared/ClueWeb09/collection.compressed.block/findex.en.10.dat"),
               new Path("/shared/ClueWeb09/docno-mapping.dat"), FileSystem.get(conf));

      FileSystem fs = FileSystem.get(conf);
View Full Code Here

Examples of edu.umd.cloud9.collection.clue.ClueWarcForwardIndex

40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
      String inputFile = conf.get("InputFile");
      String outputFile = conf.get("OutputFile");
      String findexFile = conf.get("ForwardIndexFile");
      String docnoMapping = conf.get("DocnoMappingFile");

      ClueWarcForwardIndex findex = new ClueWarcForwardIndex();
      findex.loadIndex(new Path(findexFile), new Path(docnoMapping), FileSystem.get(conf));

      FileSystem fs = FileSystem.get(conf);

      sLogger.info("reading " + inputFile);

      LineReader reader = new LineReader(fs.open(new Path(inputFile)));
      FSDataOutputStream writer = fs.create(new Path(outputFile), true);

      Text line = new Text();
      while (reader.readLine(line) > 0) {
        String[] arr = line.toString().split("\\s+");

        String docid = arr[2];
        int rank = Integer.parseInt(arr[3]);

        long start = System.currentTimeMillis();
        String url = findex.getDocument(docid).getHeaderMetadataItem("WARC-Target-URI");
        long duration = System.currentTimeMillis() - start;

        reporter.incrCounter(MyCounter.Count, 1);
        reporter.incrCounter(MyCounter.Time, duration);
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. 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.