Package org.apache.mahout.common.iterator

Examples of org.apache.mahout.common.iterator.FileLineIterable

75
76
77
78
79
80
81
82
83
84
85
    FastByIDMap<Collection<Preference>> userIDPrefMap = new FastByIDMap<Collection<Preference>>();
   
    int counter = 0;
    FilenameFilter filenameFilter = new MovieFilenameFilter();
    for (File movieFile : new File(dataDirectory, "training_set").listFiles(filenameFilter)) {
      Iterator<String> lineIterator = new FileLineIterable(movieFile, false).iterator();
      String line = lineIterator.next();
      long movieID = Long.parseLong(line.substring(0, line.length() - 1)); // strip colon
      while (lineIterator.hasNext()) {
        line = lineIterator.next();
        if (++counter % 100000 == 0) {
View Full Code Here
160
161
162
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
        = new org.apache.mahout.fpm.pfpgrowth.fpgrowth2.FPGrowthObj<String>();
      Collection<String> features = new HashSet<String>();

      try {
        fp.generateTopKFrequentPatterns(
                new StringRecordIterator(new FileLineIterable(new File(input), encoding, false), pattern),
                fp.generateFList(
                        new StringRecordIterator(new FileLineIterable(new File(input), encoding, false), pattern),
                        minSupport),
                minSupport,
                maxHeapSize,
                features,
                new StringOutputConverter(new SequenceFileOutputCollector<Text, TopKStringPatterns>(writer)),
                new ContextStatusUpdater(null));
        } finally {
          Closeables.closeQuietly(writer);
        }
    } else {
      FPGrowth<String> fp = new FPGrowth<String>();
      Collection<String> features = new HashSet<String>();
      try {
        fp.generateTopKFrequentPatterns(
                new StringRecordIterator(new FileLineIterable(new File(input), encoding, false), pattern),
                fp.generateFList(
                        new StringRecordIterator(new FileLineIterable(new File(input), encoding, false), pattern),
                        minSupport),
                minSupport,
                maxHeapSize,
                features,
                new StringOutputConverter(new SequenceFileOutputCollector<Text, TopKStringPatterns>(writer)),
View Full Code Here
83
84
85
86
87
88
89
90
91
92
93
    assertEquals("files Size: " + files.length + " is not: " + 0, 0, files.length);
    BayesFileFormatter.collapse("animal", analyzer, input, Charsets.UTF_8, new File(out, "animal"));
    files = out.listFiles();
    assertEquals("files Size: " + files.length + " is not: " + 1, 1, files.length);
    int count = 0;
    for (String line : new FileLineIterable(files[0])) {
      assertTrue("line does not start with label", line.startsWith("animal"));
      count++;
    }
    assertEquals(count + " does not equal: " + WORDS.length, count, WORDS.length);
  }
View Full Code Here
179
180
181
182
183
184
185
186
187
188
189
     */
    HadoopUtil.delete(conf, outPath);
   
    Set<String> categories = new HashSet<String>();
    if (!catFile.isEmpty()) {
      for (String line : new FileLineIterable(new File(catFile))) {
        categories.add(line.trim().toLowerCase(Locale.ENGLISH));
      }
    }
   
    DefaultStringifier<Set<String>> setStringifier =
View Full Code Here
153
154
155
156
157
158
159
160
161
162
163
             + "org.apache.hadoop.io.serializer.WritableSerialization");
    // Dont ever forget this. People should keep track of how hadoop conf
    // parameters can make or break a piece of code
   
    Set<String> categories = new HashSet<String>();
    for (String line : new FileLineIterable(new File(catFile))) {
      categories.add(line.trim().toLowerCase(Locale.ENGLISH));
    }
   
    DefaultStringifier<Set<String>> setStringifier =
        new DefaultStringifier<Set<String>>(conf, GenericsUtil.getClass(categories));
View Full Code Here
81
82
83
84
85
86
87
88
89
90
91
    indexItemIDMap = TasteHadoopUtils.readIDIndexMap(conf.get(ITEMID_INDEX_PATH), conf);

    String itemFilePathString = conf.get(ITEMS_FILE);
    if (itemFilePathString != null) {
      itemsToRecommendFor = new FastIDSet();
      for (String line : new FileLineIterable(HadoopUtil.openStream(new Path(itemFilePathString), conf))) {
        try {
          itemsToRecommendFor.add(Long.parseLong(line));
        } catch (NumberFormatException nfe) {
          log.warn("itemsFile line ignored: {}", line);
        }
View Full Code Here
93
94
95
96
97
98
99
100
101
102
103
    }
  }

  private FastByIDMap<String> buildMapping() throws IOException {
    FastByIDMap<String> mapping = new FastByIDMap<String>();
    for (String line : new FileLineIterable(dataFile)) {
      mapping.put(toLongID(line), line);
    }
    lastModified = dataFile.lastModified();
    return mapping;
  }
View Full Code Here
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
      try {
        inputStream = fs.open(input);
        inputStreamAgain = fs.open(input);
        fp.generateTopKFrequentPatterns(
                new StringRecordIterator(new FileLineIterable(inputStream, encoding, false), pattern),
                fp.generateFList(
                        new StringRecordIterator(new FileLineIterable(inputStreamAgain, encoding, false), pattern),
                        minSupport),
                minSupport,
                maxHeapSize,
                features,
                new StringOutputConverter(new SequenceFileOutputCollector<Text, TopKStringPatterns>(writer)),
                new ContextStatusUpdater(null));
      } finally {
        Closeables.close(writer, false);
        Closeables.close(inputStream, true);
        Closeables.close(inputStreamAgain, true);
      }
    } else {
      FPGrowth<String> fp = new FPGrowth<String>();


      inputStream = fs.open(input);
      inputStreamAgain = fs.open(input);
      try {
        fp.generateTopKFrequentPatterns(
                new StringRecordIterator(new FileLineIterable(inputStream, encoding, false), pattern),
                fp.generateFList(
                        new StringRecordIterator(new FileLineIterable(inputStreamAgain, encoding, false), pattern),
                        minSupport),
                minSupport,
                maxHeapSize,
                features,
                new StringOutputConverter(new SequenceFileOutputCollector<Text, TopKStringPatterns>(writer)),
View Full Code Here
63
64
65
66
67
68
69
70
71
72
73
        Path unqualifiedUsersFilePath = new Path(usersFilePathString);
        FileSystem fs = FileSystem.get(unqualifiedUsersFilePath.toUri(), jobConf);
        usersToRecommendFor = new FastIDSet();
        Path usersFilePath = unqualifiedUsersFilePath.makeQualified(fs);
        in = fs.open(usersFilePath);
        for (String line : new FileLineIterable(in)) {
          try {
            usersToRecommendFor.add(Long.parseLong(line));
          } catch (NumberFormatException nfe) {
            log.warn("usersFile line ignored: {}", line);
          }
View Full Code Here
93
94
95
96
97
98
99
100
101
102
103
      }

      String messageId = null;
      boolean inBody = false;
      Pattern quotedTextPattern = options.getQuotedTextPattern();
      for (String nextLine : new FileLineIterable(mboxFile, options.getCharset(), false)) {
        if (options.isStripQuotedText() && quotedTextPattern.matcher(nextLine).find()) {
          continue;
        }
        for (int i = 0; i < matchers.length; i++) {
          Matcher matcher = matchers[i];
View Full Code Here
TOP

Related Classes of org.apache.mahout.common.iterator.FileLineIterable

  • mia.recommender.ch05.GenderRescorer
  • org.apache.mahout.cf.taste.example.bookcrossing.BookCrossingDataModel
  • org.apache.mahout.cf.taste.example.grouplens.BooleanDataAndDateDataModel
  • org.apache.mahout.cf.taste.example.grouplens.GroupLensDataModel
  • org.apache.mahout.cf.taste.example.kddcup.track2.TrackItemSimilarity
  • org.apache.mahout.cf.taste.example.netflix.NetflixDataModel
  • org.apache.mahout.cf.taste.example.netflix.NetflixFileDataModel
  • org.apache.mahout.cf.taste.example.netflix.TransposeToByUser
  • org.apache.mahout.cf.taste.hadoop.example.als.netflix.NetflixDatasetConverter
  • org.apache.mahout.cf.taste.hadoop.item.AggregateAndRecommendReducer
    • 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.