Package SevenZip

Examples of SevenZip.LzmaAlone$CommandLine

73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
      outputDirOpt).withOption(parentOpt).withOption(helpOpt).create();
    try {
     
      Parser parser = new Parser();
      parser.setGroup(group);
      CommandLine cmdLine = parser.parse(args);
     
      if (cmdLine.hasOption(helpOpt)) {
        CommandLineUtil.printHelp(group);
        return;
      }
     
      File parentDir = new File((String) cmdLine.getValue(parentOpt));
      File outputDir = new File((String) cmdLine.getValue(outputDirOpt));
      String analyzerName = (String) cmdLine.getValue(analyzerNameOpt);
      Charset charset = Charset.forName((String) cmdLine.getValue(charsetOpt));
      Analyzer analyzer = ClassUtils.instantiateAs(analyzerName, Analyzer.class);
      // parent dir contains dir by category
      if (!parentDir.exists()) {
        throw new FileNotFoundException("Can't find input directory " + parentDir);
      }
View Full Code Here
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
    Group group = gbuilder.withName("Options").withOption(dumpFileOpt).withOption(outputDirOpt).withOption(
      chunkSizeOpt).withOption(numChunksOpt).withOption(s3IdOpt).withOption(s3SecretOpt).create();
   
    Parser parser = new Parser();
    parser.setGroup(group);
    CommandLine cmdLine;
    try {
      cmdLine = parser.parse(args);
    } catch (OptionException e) {
      log.error("Error while parsing options", e);
      CommandLineUtil.printHelp(group);
      return;
    }
   
    Configuration conf = new Configuration();
    String dumpFilePath = (String) cmdLine.getValue(dumpFileOpt);
    String outputDirPath = (String) cmdLine.getValue(outputDirOpt);
   
    if (cmdLine.hasOption(s3IdOpt)) {
      String id = (String) cmdLine.getValue(s3IdOpt);
      conf.set("fs.s3n.awsAccessKeyId", id);
      conf.set("fs.s3.awsAccessKeyId", id);
    }
    if (cmdLine.hasOption(s3SecretOpt)) {
      String secret = (String) cmdLine.getValue(s3SecretOpt);
      conf.set("fs.s3n.awsSecretAccessKey", secret);
      conf.set("fs.s3.awsSecretAccessKey", secret);
    }
    // do not compute crc file when using local FS
    conf.set("fs.file.impl", "org.apache.hadoop.fs.RawLocalFileSystem");
    FileSystem fs = FileSystem.get(URI.create(outputDirPath), conf);
   
    int chunkSize = 1024 * 1024 * Integer.parseInt((String) cmdLine.getValue(chunkSizeOpt));
   
    int numChunks = Integer.MAX_VALUE;
    if (cmdLine.hasOption(numChunksOpt)) {
      numChunks = Integer.parseInt((String) cmdLine.getValue(numChunksOpt));
    }
   
    String header = "<mediawiki xmlns=\"http://www.mediawiki.org/xml/export-0.3/\" "
                    + "xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" "
                    + "xsi:schemaLocation=\"http://www.mediawiki.org/xml/export-0.3/ "
View Full Code Here
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
    Group group = gbuilder.withName("Options").withOption(dictOpt).withOption(outOpt).withOption(wordOpt)
        .withOption(inputOpt).withOption(dictTypeOpt).create();
    try {
      Parser parser = new Parser();
      parser.setGroup(group);
      CommandLine cmdLine = parser.parse(args);
     
      if (cmdLine.hasOption(helpOpt)) {
        CommandLineUtil.printHelp(group);
        return;
      }
     
      String input = cmdLine.getValue(inputOpt).toString();
      String dictFile = cmdLine.getValue(dictOpt).toString();
      int numWords = 20;
      if (cmdLine.hasOption(wordOpt)) {
        numWords = Integer.parseInt(cmdLine.getValue(wordOpt).toString());
      }
      Configuration config = new Configuration();
     
      String dictionaryType = "text";
      if (cmdLine.hasOption(dictTypeOpt)) {
        dictionaryType = cmdLine.getValue(dictTypeOpt).toString();
      }
     
      List<String> wordList;
      if ("text".equals(dictionaryType)) {
        wordList = Arrays.asList(VectorHelper.loadTermDictionary(new File(dictFile)));
      } else if ("sequencefile".equals(dictionaryType)) {
        wordList = Arrays.asList(VectorHelper.loadTermDictionary(config, dictFile));
      } else {
        throw new IllegalArgumentException("Invalid dictionary format");
      }
     
      List<Queue<Pair<String,Double>>> topWords = topWordsForTopics(input, config, wordList, numWords);

      File output = null;
      if (cmdLine.hasOption(outOpt)) {
        output = new File(cmdLine.getValue(outOpt).toString());
        if (!output.exists() && !output.mkdirs()) {
          throw new IOException("Could not create directory: " + output);
        }
      }
      printTopWords(topWords, output);
View Full Code Here
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
        .create();
   
    Parser parser = new Parser();
    parser.setGroup(group);
    try {
      CommandLine cmdLine = parser.parse(args);
      if (cmdLine.hasOption(helpOpt)) {
        CommandLineUtil.printHelp(group);
        return;
      }
     
      String inputPath = (String) cmdLine.getValue(dirInputPathOpt);
      String outputPath = (String) cmdLine.getValue(dirOutputPathOpt);
      String catFile = (String) cmdLine.getValue(categoriesOpt);
      Class<? extends Analyzer> analyzerClass = WikipediaAnalyzer.class;
      if (cmdLine.hasOption(analyzerOpt)) {
        String className = cmdLine.getValue(analyzerOpt).toString();
        analyzerClass = Class.forName(className).asSubclass(Analyzer.class);
        // try instantiating it, b/c there isn't any point in setting it if
        // you can't instantiate it
        ClassUtils.instantiateAs(analyzerClass, Analyzer.class);
      }
      runJob(inputPath, outputPath, catFile, cmdLine.hasOption(exactMatchOpt),
        analyzerClass);
    } catch (OptionException e) {
      log.error("Exception", e);
      CommandLineUtil.printHelp(group);
    } catch (ClassNotFoundException e) {
View Full Code Here
189
190
191
192
193
194
195
196
197
198
199
    Parser parser = new Parser();
    parser.setHelpOption(help);
    parser.setHelpTrigger("--help");
    parser.setGroup(normalArgs);
    parser.setHelpFormatter(new HelpFormatter(" ", "", " ", 130));
    CommandLine cmdLine = parser.parseAndHelp(args);

    if (cmdLine == null) {
      return false;
    }
View Full Code Here
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
   
    Parser parser = new Parser();
    parser.setGroup(group);
    parser.setHelpOption(helpOpt);
    try {
      CommandLine cmdLine = parser.parse(args);
      if (cmdLine.hasOption(helpOpt)) {
        CommandLineUtil.printHelp(group);
        return;
      }
     
      String inputPath = (String) cmdLine.getValue(dirInputPathOpt);
      String outputPath = (String) cmdLine.getValue(dirOutputPathOpt);
     
      String catFile = "";
      if (cmdLine.hasOption(categoriesOpt)) {
        catFile = (String) cmdLine.getValue(categoriesOpt);
      }
     
      boolean all = false;
      if (cmdLine.hasOption(allOpt)) {
        all = true;
      }
      runJob(inputPath, outputPath, catFile, cmdLine.hasOption(exactMatchOpt), all);
    } catch (OptionException e) {
      log.error("Exception", e);
      CommandLineUtil.printHelp(group);
    } catch (InterruptedException e) {
      log.error("Exception", e);
View Full Code Here
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
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
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
        .withOption(weightOpt).withOption(minDFOpt).create();

    try {
      Parser parser = new Parser();
      parser.setGroup(group);
      CommandLine cmdLine = parser.parse(args);

      if (cmdLine.hasOption(helpOpt)) {
       
        CommandLineUtil.printHelp(group);
        return;
      }
      // Springify all this
      if (cmdLine.hasOption(inputOpt)) { // Lucene case
        File file = new File(cmdLine.getValue(inputOpt).toString());
        if (!file.isDirectory()) {
          throw new IllegalArgumentException("Lucene directory: " + file.getAbsolutePath()
              +  " does not exist or is not a directory");
        }

        long maxDocs = Long.MAX_VALUE;
        if (cmdLine.hasOption(maxOpt)) {
          maxDocs = Long.parseLong(cmdLine.getValue(maxOpt).toString());
        }
        if (maxDocs < 0) {
          throw new IllegalArgumentException("maxDocs must be >= 0");
        }

        Directory dir = FSDirectory.open(file);
        IndexReader reader = IndexReader.open(dir, true);

        Weight weight;
        if (cmdLine.hasOption(weightOpt)) {
          String wString = cmdLine.getValue(weightOpt).toString();
          if ("tf".equalsIgnoreCase(wString)) {
            weight = new TF();
          } else if ("tfidf".equalsIgnoreCase(wString)) {
            weight = new TFIDF();
          } else {
            throw new OptionException(weightOpt);
          }
        } else {
          weight = new TFIDF();
        }

        String field = cmdLine.getValue(fieldOpt).toString();

        int minDf = 1;
        if (cmdLine.hasOption(minDFOpt)) {
          minDf = Integer.parseInt(cmdLine.getValue(minDFOpt).toString());
        }

        int maxDFPercent = 99;
        if (cmdLine.hasOption(maxDFPercentOpt)) {
          maxDFPercent = Integer.parseInt(cmdLine.getValue(maxDFPercentOpt).toString());
        }

        TermInfo termInfo = new CachedTermInfo(reader, field, minDf, maxDFPercent);
        VectorMapper mapper = new TFDFMapper(reader, weight, termInfo);

        double norm = LuceneIterable.NO_NORMALIZING;
        if (cmdLine.hasOption(powerOpt)) {
          String power = cmdLine.getValue(powerOpt).toString();
          if ("INF".equals(power)) {
            norm = Double.POSITIVE_INFINITY;
          } else {
            norm = Double.parseDouble(power);
          }
        }

        String idField = null;
        if (cmdLine.hasOption(idFieldOpt)) {
          idField = cmdLine.getValue(idFieldOpt).toString();
        }

        LuceneIterable iterable;
        if (norm == LuceneIterable.NO_NORMALIZING) {
          iterable = new LuceneIterable(reader, idField, field, mapper, LuceneIterable.NO_NORMALIZING);
        } else {
          iterable = new LuceneIterable(reader, idField, field, mapper, norm);
        }

        String outFile = cmdLine.getValue(outputOpt).toString();
        log.info("Output File: {}", outFile);

        VectorWriter vectorWriter;
        if (cmdLine.hasOption(outWriterOpt)) {
          String outWriter = cmdLine.getValue(outWriterOpt).toString();
          if ("file".equals(outWriter)) {
            BufferedWriter writer = new BufferedWriter(new FileWriter(outFile));
            vectorWriter = new JWriterVectorWriter(writer);
          } else {
            vectorWriter = getSeqFileWriter(outFile);
          }
        } else {
          vectorWriter = getSeqFileWriter(outFile);
        }

        long numDocs = vectorWriter.write(iterable, maxDocs);
        vectorWriter.close();
        log.info("Wrote: {} vectors", numDocs);

        String delimiter = cmdLine.hasOption(delimiterOpt) ? cmdLine.getValue(delimiterOpt).toString() : "\t";
       
        File dictOutFile = new File(cmdLine.getValue(dictOutOpt).toString());
        log.info("Dictionary Output file: {}", dictOutFile);
        BufferedWriter writer = new BufferedWriter(new OutputStreamWriter(
            new FileOutputStream(dictOutFile), Charset.forName("UTF8")));
        JWriterTermInfoWriter tiWriter = new JWriterTermInfoWriter(writer, delimiter, field);
        tiWriter.write(termInfo);
View Full Code Here
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
            printKeyOpt).withOption(sizeOpt).create();

    try {
      Parser parser = new Parser();
      parser.setGroup(group);
      CommandLine cmdLine = parser.parse(args);

      if (cmdLine.hasOption(helpOpt)) {

        printHelp(group);
        return;
      }

      if (cmdLine.hasOption(seqOpt)) {
        Path path = new Path(cmdLine.getValue(seqOpt).toString());
        //System.out.println("Input Path: " + path); interferes with output?
        Configuration conf = new Configuration();

        FileSystem fs = FileSystem.get(path.toUri(), conf);

        String dictionaryType = "text";
        if (cmdLine.hasOption(dictTypeOpt)) {
          dictionaryType = cmdLine.getValue(dictTypeOpt).toString();
        }

        String[] dictionary = null;
        if (cmdLine.hasOption(dictOpt)) {
          if ("text".equals(dictionaryType)) {
            dictionary = VectorHelper.loadTermDictionary(new File(cmdLine.getValue(dictOpt).toString()));
          } else if ("sequencefile".equals(dictionaryType)) {
            dictionary = VectorHelper.loadTermDictionary(conf, fs, cmdLine.getValue(dictOpt).toString());
          } else {
            throw new OptionException(dictTypeOpt);
          }
        }
        boolean useJSON = cmdLine.hasOption(centroidJSonOpt);
        boolean sizeOnly = cmdLine.hasOption(sizeOpt);
        SequenceFile.Reader reader = new SequenceFile.Reader(fs, path, conf);

        Writable keyWritable = reader.getKeyClass().asSubclass(Writable.class).newInstance();
        Writable valueWritable = reader.getValueClass().asSubclass(Writable.class).newInstance();
        boolean transposeKeyValue = cmdLine.hasOption(vectorAsKeyOpt);
        try {
          Writer writer = cmdLine.hasOption(outputOpt)
                  ? new FileWriter(cmdLine.getValue(outputOpt).toString())
                  : new OutputStreamWriter(System.out);
          try {
            boolean printKey = cmdLine.hasOption(printKeyOpt);
            long i = 0;
            while (reader.next(keyWritable, valueWritable)) {
              if (printKey) {
                Writable notTheVectorWritable = transposeKeyValue ? valueWritable : keyWritable;
                writer.write(notTheVectorWritable.toString());
View Full Code Here
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
      .withOption(substringOpt).withOption(countOpt).withOption(helpOpt).create();
   
    try {
      Parser parser = new Parser();
      parser.setGroup(group);
      CommandLine cmdLine = parser.parse(args);
     
      if (cmdLine.hasOption(helpOpt)) {
       
        printHelp(group);
        return;
      }
     
      if (cmdLine.hasOption(seqOpt)) {
        Path path = new Path(cmdLine.getValue(seqOpt).toString());
        Configuration conf = new Configuration();
        FileSystem fs = FileSystem.get(path.toUri(), conf);
        SequenceFile.Reader reader = new SequenceFile.Reader(fs, path, conf);
       
        Writer writer;
        if (cmdLine.hasOption(outputOpt)) {
          writer = new FileWriter(cmdLine.getValue(outputOpt).toString());
        } else {
          writer = new OutputStreamWriter(System.out);
        }
        try {
          writer.append("Input Path: ").append(String.valueOf(path)).append('\n');

          int sub = Integer.MAX_VALUE;
          if (cmdLine.hasOption(substringOpt)) {
            sub = Integer.parseInt(cmdLine.getValue(substringOpt).toString());
          }
          boolean countOnly = cmdLine.hasOption(countOpt);
          Writable key = reader.getKeyClass().asSubclass(Writable.class).newInstance();
          Writable value = reader.getValueClass().asSubclass(Writable.class).newInstance();
          writer.append("Key class: ").append(String.valueOf(reader.getKeyClass()));
          writer.append(" Value Class: ").append(String.valueOf(value.getClass())).append('\n');
          writer.flush();
View Full Code Here
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
        .withOption(outputOpt).withOption(analyzeOpt).withOption(mrOpt).withOption(helpOpt).create();

    try {
      Parser parser = new Parser();
      parser.setGroup(group);
      CommandLine cmdLine = parser.parse(args);

      if (cmdLine.hasOption("help")) {
        CommandLineUtil.printHelp(group);
        return -1;
      }

      String dataName = cmdLine.getValue(inputOpt).toString();
      String datasetName = cmdLine.getValue(datasetOpt).toString();
      String modelName = cmdLine.getValue(modelOpt).toString();
      String outputName = (cmdLine.hasOption(outputOpt)) ? cmdLine.getValue(outputOpt).toString() : null;
      analyze = cmdLine.hasOption(analyzeOpt);
      useMapreduce = cmdLine.hasOption(mrOpt);

      log.debug("inout     : {}", dataName);
      log.debug("dataset   : {}", datasetName);
      log.debug("model     : {}", modelName);
      log.debug("output    : {}", outputName);
View Full Code Here
TOP

Related Classes of SevenZip.LzmaAlone$CommandLine

  • com.digitalpebble.behemoth.mahout.SparseVectorsFromBehemoth
  • com.digitalpebble.behemoth.tika.TikaDriver
  • org.apache.commons.cli2.commandline.DefaultingCommandLine
  • org.apache.commons.cli2.commandline.Parser
  • org.apache.hadoop.mapred.pipes.Submitter
  • org.apache.hadoop.streaming.StreamJob
  • org.apache.jackrabbit.vault.util.console.AbstractApplication
  • org.apache.jackrabbit.vault.util.console.ExecutionContext
  • org.apache.jackrabbit.vault.util.console.Test
  • org.apache.mahout.benchmark.VectorBenchmarks
    • 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.