@Override
public int run(String[] args) throws Exception {
final int ret = parseArgs(args);
if (ret < 0) {
return ret;
}
Job job = Job.getInstance(getConf());
job.setJarByClass(GreeDiFirst.class);
job.setJobName(String.format("Coverage-GreeDiFirst[%s %s]", partitionCount, selectCount));
job.getConfiguration().setInt(Fields.PARTITION_COUNT.get(), partitionCount);
job.getConfiguration().setInt(Fields.SELECT_COUNT.get(), selectCount);
job.setNumReduceTasks(partitionCount);
SetupHelper.getInstance().setSequenceInput(job, inputPath).setTextOutput(job, outputPath);
job.setMapOutputKeyClass(IntWritable.class);
job.setMapOutputValueClass(DocumentWithVectorWritable.class);
job.setOutputKeyClass(NullWritable.class);
job.setOutputValueClass(IntWritable.class);
job.setMapperClass(Map.class);
job.setReducerClass(GreeDiReducer.class);
// Delete the output directory if it exists already.
FileSystem.get(getConf()).delete(new Path(outputPath), true);
job.waitForCompletion(true);
return 0;
}
@Override
public void map(Text key, VectorWritable value, Context context)
throws IOException, InterruptedException {
final int partitionCount =
context
.getConfiguration()
.getInt(Fields.PARTITION_COUNT.get(), Defaults.PARTITION_COUNT.get());
final int partition = Integer.parseInt(key.toString()) % partitionCount;
final IntWritable outKey = new IntWritable(partition);
final DocumentWithVectorWritable outValue = new DocumentWithVectorWritable(key, value);
context.write(outKey, outValue);
}
@SuppressWarnings("static-access")
private int parseArgs(String[] args) {
Options options = new Options();
options.addOption(
OptionBuilder.withArgName("path")
.hasArg()
.withDescription("Tfidf vectors")
.create(Fields.INPUT.get()));
options.addOption(
OptionBuilder.withArgName("path")
.hasArg()
.withDescription("Selected articles")
.create(Fields.OUTPUT.get()));
options.addOption(
OptionBuilder.withArgName("integer")
.hasArg()
.withDescription("Partition count")
.create(Fields.PARTITION_COUNT.get()));
options.addOption(
OptionBuilder.withArgName("integer")
.hasArg()
.withDescription("Select count")
.create(Fields.SELECT_COUNT.get()));
CommandLine cmdline;
CommandLineParser parser = new GnuParser();
try {
cmdline = parser.parse(options, args);
} catch (ParseException exp) {
logger.fatal("Error parsing command line: " + exp.getMessage());
return -1;
}
if (!cmdline.hasOption(Fields.INPUT.get()) || !cmdline.hasOption(Fields.OUTPUT.get())) {
HelpFormatter formatter = new HelpFormatter();
formatter.printHelp(this.getClass().getName(), options);
ToolRunner.printGenericCommandUsage(System.out);
return -1;
}
inputPath = cmdline.getOptionValue(Fields.INPUT.get());
outputPath = cmdline.getOptionValue(Fields.OUTPUT.get());
partitionCount = Defaults.PARTITION_COUNT.get();
if (cmdline.hasOption(Fields.PARTITION_COUNT.get())) {
partitionCount = Integer.parseInt(cmdline.getOptionValue(Fields.PARTITION_COUNT.get()));
if (partitionCount <= 0) {
System.err.println("Error: \"" + partitionCount + "\" has to be positive!");
return -1;
}
}
selectCount = Integer.parseInt(cmdline.getOptionValue(Fields.SELECT_COUNT.get()));
logger.info("Tool name: " + this.getClass().getName());
logger.info(" - input: " + inputPath);
logger.info(" - output: " + outputPath);
logger.info(" - partitions: " + partitionCount);
logger.info(" - select: " + selectCount);
return 0;
}