Example usage for org.apache.hadoop.mapred JobConf setNumMapTasks

List of usage examples for org.apache.hadoop.mapred JobConf setNumMapTasks

  1. HOME
  2. Java
  3. org.apache.hadoop
  4. org.apache.hadoop.mapred.*
  5. JobConf
  6. setNumMapTasks

Introduction

In this page you can find the example usage for org.apache.hadoop.mapred JobConf setNumMapTasks.

Prototype

public void setNumMapTasks(int n)

Source Link

Document

Set the number of map tasks for this job.

Usage

From source file:io.aos.t4f.hadoop.mapred.WordCountMapReduceTest.java

License:Apache License

/**
 * The main driver for word count map/reduce program.
 * Invoke this method to submit the map/reduce job.
 * @throws IOException When there is communication problems with the 
 *                     job tracker.// w w  w.  j a  v  a  2s  .  c om
 */
public int run(String[] args) throws Exception {
    JobConf conf = new JobConf(getConf(), WordCountMapReduceTest.class);
    conf.setJobName("wordcount");

    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(IntWritable.class);

    conf.setMapperClass(WordCountMapper.class);
    conf.setCombinerClass(WordCountReducer.class);
    conf.setReducerClass(WordCountReducer.class);

    List<String> other_args = new ArrayList<String>();
    for (int i = 0; i < args.length; ++i) {
        try {
            if ("-m".equals(args[i])) {
                conf.setNumMapTasks(Integer.parseInt(args[++i]));
            } else if ("-r".equals(args[i])) {
                conf.setNumReduceTasks(Integer.parseInt(args[++i]));
            } else {
                other_args.add(args[i]);
            }
        } catch (NumberFormatException except) {
            System.out.println("ERROR: Integer expected instead of " + args[i]);
            return printUsage();
        } catch (ArrayIndexOutOfBoundsException except) {
            System.out.println("ERROR: Required parameter missing from " + args[i - 1]);
            return printUsage();
        }
    }
    // Make sure there are exactly 2 parameters left.
    if (other_args.size() != 2) {
        System.out.println("ERROR: Wrong number of parameters: " + other_args.size() + " instead of 2.");
        return printUsage();
    }
    FileInputFormat.setInputPaths(conf, other_args.get(0));
    FileOutputFormat.setOutputPath(conf, new Path(other_args.get(1)));

    JobClient.runJob(conf);
    return 0;
}

From source file:io.aos.t4f.hadoop.mapred.WordCountTest.java

License:Apache License

private JobConf createJobConf() {
    JobConf conf = mrCluster.createJobConf();
    conf.setJobName("wordcount test");

    conf.setMapperClass(WordCountMapper.class);
    conf.setReducerClass(WordCountReducer.class);

    conf.setInputFormat(TextInputFormat.class);
    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(IntWritable.class);
    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(IntWritable.class);
    conf.setNumMapTasks(1);
    conf.setNumReduceTasks(1);//ww w.ja  v  a  2s.c  o  m
    FileInputFormat.setInputPaths(conf, input);
    FileOutputFormat.setOutputPath(conf, output);
    return conf;
}

From source file:ivory.core.index.MergeGlobalStatsAcrossIndexSegments.java

License:Apache License

public int runTool() throws Exception {

    JobConf conf = new JobConf(getConf(), MergeGlobalStatsAcrossIndexSegments.class);
    FileSystem fs = FileSystem.get(conf);

    String collectionName = conf.get("Ivory.CollectionName");
    String indexPaths = conf.get("Ivory.IndexPaths");
    String dataOutputPath = conf.get("Ivory.DataOutputPath");
    int dfThreshold = conf.getInt("Ivory.DfThreshold", 0);

    // first, compute size of global term space
    Path tmpPaths = new Path("/tmp/index-paths.txt");

    FSDataOutputStream out = fs.create(tmpPaths, true);
    for (String s : indexPaths.split(",")) {
        out.write(new String(s + "\n").getBytes());
    }// w  w  w  .  j  ava2s.c  o m
    out.close();

    LOG.info("Job: ComputeNumberOfTermsAcrossIndexSegments");
    conf.setJobName("ComputeNumberOfTermsAcrossIndexSegments:" + collectionName);

    FileInputFormat.addInputPath(conf, tmpPaths);

    conf.setNumMapTasks(1);
    conf.setNumReduceTasks(1);

    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.setInputFormat(NLineInputFormat.class);
    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(PairOfIntLong.class);
    conf.setOutputFormat(NullOutputFormat.class);

    conf.setMapperClass(MyMapper.class);
    conf.setReducerClass(IdentityReducer.class);

    long startTime = System.currentTimeMillis();
    RunningJob job = JobClient.runJob(conf);
    LOG.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");

    Counters counters = job.getCounters();

    long totalNumTerms = counters.findCounter("org.apache.hadoop.mapred.Task$Counter", 6, "REDUCE_INPUT_GROUPS")
            .getCounter();

    LOG.info("total number of terms in global dictionary = " + totalNumTerms);

    // now build the dictionary
    fs.delete(new Path(dataOutputPath), true);

    conf = new JobConf(getConf(), MergeGlobalStatsAcrossIndexSegments.class);

    LOG.info("Job: MergeGlobalStatsAcrossIndexSegments");
    conf.setJobName("MergeGlobalStatsAcrossIndexSegments:" + collectionName);

    FileInputFormat.addInputPath(conf, tmpPaths);

    conf.setNumMapTasks(1);
    conf.setNumReduceTasks(1);

    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.setInputFormat(NLineInputFormat.class);
    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(PairOfIntLong.class);
    conf.setOutputFormat(NullOutputFormat.class);

    conf.setMapperClass(MyMapper.class);
    conf.setReducerClass(MyReducer.class);

    conf.setLong("Ivory.IndexNumberOfTerms", (int) totalNumTerms);

    startTime = System.currentTimeMillis();
    job = JobClient.runJob(conf);
    LOG.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");

    // compute some # docs, collection length, avg doc length
    long collectionLength = 0;
    int docCount = 0;
    for (String index : indexPaths.split(",")) {
        LOG.info("reading stats for " + index);

        RetrievalEnvironment env = new RetrievalEnvironment(index, fs);

        long l = env.readCollectionLength();
        int n = env.readCollectionDocumentCount();

        LOG.info(" - CollectionLength: " + l);
        LOG.info(" - CollectionDocumentCount: " + n);

        collectionLength += l;
        docCount += n;
    }

    float avgdl = (float) collectionLength / docCount;

    LOG.info("all index segments: ");
    LOG.info(" - CollectionLength: " + collectionLength);
    LOG.info(" - CollectionDocumentCount: " + docCount);
    LOG.info(" - AverageDocumentLenght: " + avgdl);

    RetrievalEnvironment env = new RetrievalEnvironment(dataOutputPath, fs);

    env.writeCollectionAverageDocumentLength(avgdl);
    env.writeCollectionLength(collectionLength);
    env.writeCollectionDocumentCount(docCount);

    return 0;
}

From source file:ivory.core.preprocess.BuildTargetLangWeightedIntDocVectors.java

License:Apache License

@SuppressWarnings("deprecation")
public int runTool() throws Exception {
    //      sLogger.setLevel(Level.DEBUG);

    sLogger.info("PowerTool: GetTargetLangWeightedIntDocVectors");

    JobConf conf = new JobConf(BuildTargetLangWeightedIntDocVectors.class);
    FileSystem fs = FileSystem.get(conf);

    String indexPath = getConf().get("Ivory.IndexPath");

    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);

    String outputPath = env.getWeightedIntDocVectorsDirectory();
    int mapTasks = getConf().getInt("Ivory.NumMapTasks", 0);
    int minSplitSize = getConf().getInt("Ivory.MinSplitSize", 0);
    String collectionName = getConf().get("Ivory.CollectionName");

    sLogger.info("Characteristics of the collection:");
    sLogger.info(" - CollectionName: " + collectionName);
    sLogger.info("Characteristics of the job:");
    sLogger.info(" - NumMapTasks: " + mapTasks);
    sLogger.info(" - MinSplitSize: " + minSplitSize);

    String vocabFile = getConf().get("Ivory.FinalVocab");
    DistributedCache.addCacheFile(new URI(vocabFile), conf);

    Path inputPath = new Path(PwsimEnvironment.getFileNameWithPars(indexPath, "TermDocs"));
    Path weightedVectorsPath = new Path(outputPath);

    if (fs.exists(weightedVectorsPath)) {
        sLogger.info("Output path already exists!");
        return -1;
    }// w  ww.  ja va  2 s .  com
    conf.setJobName("GetWeightedIntDocVectors:" + collectionName);
    conf.setNumMapTasks(mapTasks);
    conf.setNumReduceTasks(0);
    conf.setInt("mapred.min.split.size", minSplitSize);
    conf.set("mapred.child.java.opts", "-Xmx2048m");
    conf.setBoolean("Ivory.Normalize", getConf().getBoolean("Ivory.Normalize", false));
    FileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, weightedVectorsPath);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(WeightedIntDocVector.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);
    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(WeightedIntDocVector.class);

    conf.setMapperClass(MyMapper.class);

    long startTime = System.currentTimeMillis();

    RunningJob rj = JobClient.runJob(conf);
    sLogger.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");
    Counters counters = rj.getCounters();

    long numOfDocs = (long) counters.findCounter(Docs.Total).getCounter();

    return (int) numOfDocs;
}

From source file:ivory.core.preprocess.BuildWeightedIntDocVectors.java

License:Apache License

@SuppressWarnings("deprecation")
public int runTool() throws Exception {
    sLogger.setLevel(Level.WARN);

    sLogger.info("PowerTool: GetWeightedIntDocVectors");

    // create a new JobConf, inheriting from the configuration of this
    // PowerTool/*from   w  ww .  j a v  a2s  .  co m*/
    JobConf conf = new JobConf(getConf(), BuildWeightedIntDocVectors.class);
    FileSystem fs = FileSystem.get(conf);

    String indexPath = conf.get("Ivory.IndexPath");
    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);
    String outputPath = env.getWeightedIntDocVectorsDirectory();
    int mapTasks = conf.getInt("Ivory.NumMapTasks", 0);
    int minSplitSize = conf.getInt("Ivory.MinSplitSize", 0);
    String collectionName = conf.get("Ivory.CollectionName");

    sLogger.info("Characteristics of the collection:");
    sLogger.info(" - CollectionName: " + collectionName);
    sLogger.info("Characteristics of the job:");
    sLogger.info(" - NumMapTasks: " + mapTasks);
    sLogger.info(" - MinSplitSize: " + minSplitSize);

    String dfByIntFilePath = env.getDfByIntData();
    String cfByIntFilePath = env.getCfByIntData();

    /* add df table to cache */
    if (!fs.exists(new Path(dfByIntFilePath))) {
        throw new RuntimeException("Error, df data file " + dfByIntFilePath + "doesn't exist!");
    }
    DistributedCache.addCacheFile(new URI(dfByIntFilePath), conf);

    /* add cf table to cache */
    if (!fs.exists(new Path(cfByIntFilePath))) {
        throw new RuntimeException("Error, cf data file " + cfByIntFilePath + "doesn't exist!");
    }
    DistributedCache.addCacheFile(new URI(cfByIntFilePath), conf);

    /* add dl table to cache */
    Path docLengthFile = env.getDoclengthsData();
    if (!fs.exists(docLengthFile)) {
        throw new RuntimeException("Error, doc-length data file " + docLengthFile + "doesn't exist!");
    }
    DistributedCache.addCacheFile(docLengthFile.toUri(), conf);

    Path inputPath = new Path(env.getIntDocVectorsDirectory());
    Path weightedVectorsPath = new Path(outputPath);

    if (fs.exists(weightedVectorsPath)) {
        sLogger.info("Output path already exists!");
        return 0;
    }

    //fs.delete(weightedVectirsPath, true);

    conf.setJobName("GetWeightedIntDocVectors:" + collectionName);
    conf.setNumMapTasks(mapTasks);
    conf.setNumReduceTasks(0);
    conf.setInt("mapred.min.split.size", minSplitSize);
    conf.set("mapred.child.java.opts", "-Xmx2048m");

    FileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, weightedVectorsPath);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(WeightedIntDocVector.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);
    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(WeightedIntDocVector.class);

    conf.setMapperClass(MyMapper.class);
    //conf.setInt("mapred.task.timeout",3600000);

    long startTime = System.currentTimeMillis();

    RunningJob job = JobClient.runJob(conf);
    sLogger.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");

    return 0;
}

From source file:ivory.core.preprocess.BuildWeightedTermDocVectors.java

License:Apache License

@SuppressWarnings("deprecation")
public int runTool() throws Exception {
    sLogger.info("PowerTool: GetWeightedTermDocVectors");

    JobConf conf = new JobConf(BuildWeightedTermDocVectors.class);
    FileSystem fs = FileSystem.get(conf);

    String indexPath = getConf().get("Ivory.IndexPath");
    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);
    String outputPath = env.getWeightedTermDocVectorsDirectory();
    int mapTasks = getConf().getInt("Ivory.NumMapTasks", 0);
    int minSplitSize = getConf().getInt("Ivory.MinSplitSize", 0);
    String collectionName = getConf().get("Ivory.CollectionName");

    String termsFilePath = env.getIndexTermsData();
    String termsIdsFilePath = env.getIndexTermIdsData();
    String termIdMappingFilePath = env.getIndexTermIdMappingData();
    String dfByTermFilePath = env.getDfByTermData();

    Path inputPath = new Path(env.getTermDocVectorsDirectory());
    Path weightedVectorsPath = new Path(outputPath);

    if (fs.exists(weightedVectorsPath)) {
        //fs.delete(weightedVectorsPath, true);
        sLogger.info("Output path already exists!");
        return 0;
    }//from   w w  w. j  av a 2  s.c o m

    /* add terms file to cache */
    if (!fs.exists(new Path(termsFilePath)) || !fs.exists(new Path(termsIdsFilePath))
            || !fs.exists(new Path(termIdMappingFilePath))) {
        throw new RuntimeException("Error, terms file " + termsFilePath + "/" + termsIdsFilePath + "/"
                + termIdMappingFilePath + "doesn't exist!");
    }
    DistributedCache.addCacheFile(new URI(termsFilePath), conf);
    DistributedCache.addCacheFile(new URI(termsIdsFilePath), conf);
    DistributedCache.addCacheFile(new URI(termIdMappingFilePath), conf);

    /* add df table to cache */
    if (!fs.exists(new Path(dfByTermFilePath))) {
        throw new RuntimeException("Error, df data file " + dfByTermFilePath + "doesn't exist!");
    }
    DistributedCache.addCacheFile(new URI(dfByTermFilePath), conf);

    /* add dl table to cache */
    Path docLengthFile = env.getDoclengthsData();
    if (!fs.exists(docLengthFile)) {
        throw new RuntimeException("Error, doc-length data file " + docLengthFile + "doesn't exist!");
    }
    DistributedCache.addCacheFile(docLengthFile.toUri(), conf);

    conf.setMapperClass(MyMapper.class);
    //conf.setInt("mapred.task.timeout",3600000);
    conf.setJobName("GetWeightedTermDocVectors:" + collectionName);
    conf.setNumMapTasks(mapTasks);
    conf.setNumReduceTasks(0);
    conf.setInt("mapred.min.split.size", minSplitSize);
    conf.set("mapred.child.java.opts", "-Xmx2048m");
    conf.setInt("Ivory.MinNumTerms", getConf().getInt("Ivory.MinNumTerms", Integer.MAX_VALUE));
    conf.setBoolean("Ivory.Normalize", getConf().getBoolean("Ivory.Normalize", false));
    if (getConf().get("Ivory.ShortDocLengths") != null) {
        conf.set("Ivory.ShortDocLengths", getConf().get("Ivory.ShortDocLengths"));
    }
    conf.set("Ivory.ScoringModel", getConf().get("Ivory.ScoringModel"));

    FileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, weightedVectorsPath);
    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(HMapSFW.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);
    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(HMapSFW.class);

    sLogger.info("Running job: " + conf.getJobName());

    long startTime = System.currentTimeMillis();
    RunningJob job = JobClient.runJob(conf);
    sLogger.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");

    return 0;
}

From source file:ivory.index.BuildIntPostingsForwardIndex.java

License:Apache License

public int runTool() throws Exception {
    JobConf conf = new JobConf(getConf(), BuildIntPostingsForwardIndex.class);
    FileSystem fs = FileSystem.get(conf);

    int mapTasks = conf.getInt("Ivory.NumMapTasks", 0);
    int minSplitSize = conf.getInt("Ivory.MinSplitSize", 0);
    String indexPath = conf.get("Ivory.IndexPath");

    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);
    String collectionName = env.readCollectionName();

    sLogger.info("Tool: BuildIntPostingsForwardIndex");
    sLogger.info(" - IndexPath: " + indexPath);
    sLogger.info(" - CollectionName: " + collectionName);

    conf.setJobName("BuildIntPostingsForwardIndex:" + collectionName);

    Path inputPath = new Path(env.getPostingsDirectory());
    FileInputFormat.setInputPaths(conf, inputPath);

    Path postingsIndexPath = new Path(env.getPostingsIndexData());

    if (fs.exists(postingsIndexPath)) {
        sLogger.info("Postings forward index path already exists!");
        return 0;
    }//from   w  w  w  .  j  a v a2s  .c  om
    conf.setNumMapTasks(mapTasks);
    conf.setNumReduceTasks(1);

    conf.setInt("mapred.min.split.size", minSplitSize);
    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(Text.class);
    conf.setOutputKeyClass(Text.class);
    conf.setOutputValueClass(Text.class);
    conf.setOutputFormat(NullOutputFormat.class);

    conf.setMapRunnerClass(MyMapRunner.class);
    conf.setReducerClass(MyReducer.class);

    JobClient.runJob(conf);

    return 0;
}

From source file:ivory.index.BuildIPInvertedIndexDocSorted.java

License:Apache License

@SuppressWarnings("unused")
public int runTool() throws Exception {
    JobConf conf = new JobConf(getConf(), BuildIPInvertedIndexDocSorted.class);
    FileSystem fs = FileSystem.get(conf);

    String indexPath = conf.get("Ivory.IndexPath");
    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);

    String collectionName = env.readCollectionName();

    int mapTasks = conf.getInt("Ivory.NumMapTasks", 0);
    int reduceTasks = conf.getInt("Ivory.NumReduceTasks", 0);
    int minSplitSize = conf.getInt("Ivory.MinSplitSize", 0);
    int collectionDocCnt = env.readCollectionDocumentCount();

    LOG.info("PowerTool: BuildIPInvertedIndexDocSorted");
    LOG.info(" - IndexPath: " + indexPath);
    LOG.info(" - CollectionName: " + collectionName);
    LOG.info(" - CollectionDocumentCount: " + collectionDocCnt);
    LOG.info(" - NumMapTasks: " + mapTasks);
    LOG.info(" - NumReduceTasks: " + reduceTasks);
    LOG.info(" - MinSplitSize: " + minSplitSize);

    if (!fs.exists(new Path(indexPath))) {
        fs.mkdirs(new Path(indexPath));
    }/*from  ww w  .ja va  2  s  .c o m*/

    Path inputPath = new Path(env.getIntDocVectorsDirectory());
    Path postingsPath = new Path(env.getPostingsDirectory());

    if (fs.exists(postingsPath)) {
        LOG.info("Postings already exist: no indexing will be performed.");
        return 0;
    }

    conf.setJobName("BuildIPInvertedIndex:" + collectionName);

    conf.setNumMapTasks(mapTasks);
    conf.setNumReduceTasks(reduceTasks);

    conf.setInt("Ivory.CollectionDocumentCount", collectionDocCnt);

    conf.setInt("mapred.min.split.size", minSplitSize);
    conf.set("mapred.child.java.opts", "-Xmx2048m");

    FileInputFormat.setInputPaths(conf, inputPath);
    FileOutputFormat.setOutputPath(conf, postingsPath);

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);

    conf.setMapOutputKeyClass(PairOfInts.class);
    conf.setMapOutputValueClass(TermPositions.class);
    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(PostingsListDocSortedPositional.class);

    conf.setMapperClass(MyMapper.class);
    conf.setReducerClass(MyReducer.class);
    conf.setPartitionerClass(MyPartitioner.class);

    long startTime = System.currentTimeMillis();
    RunningJob job = JobClient.runJob(conf);
    LOG.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");

    env.writePostingsType("ivory.data.PostingsListDocSortedPositional");

    return 0;
}

From source file:ivory.preprocess.BuildIntDocVectors.java

License:Apache License

@SuppressWarnings("unused")
public int runTool() throws Exception {
    // create a new JobConf, inheriting from the configuration of this
    // PowerTool/*w  ww  .  ja  v a 2  s.co m*/
    JobConf conf = new JobConf(getConf(), BuildIntDocVectors.class);
    FileSystem fs = FileSystem.get(conf);

    String indexPath = conf.get("Ivory.IndexPath");
    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);

    int mapTasks = conf.getInt("Ivory.NumMapTasks", 0);

    String collectionName = env.readCollectionName();

    sLogger.info("PowerTool: BuildIntDocVectors");
    sLogger.info(" - IndexPath: " + indexPath);
    sLogger.info(" - CollectionName: " + collectionName);
    sLogger.info(" - NumMapTasks: " + mapTasks);
    sLogger.info("This is new!");
    String termsFile = env.getIndexTermsData();
    String termIDsFile = env.getIndexTermIdsData();
    String idToTermFile = env.getIndexTermIdMappingData();

    Path termsFilePath = new Path(termsFile);
    Path termIDsFilePath = new Path(termIDsFile);

    if (!fs.exists(termsFilePath) || !fs.exists(termIDsFilePath)) {
        sLogger.error("Error, terms files don't exist!");
        return 0;
    }

    Path outputPath = new Path(env.getIntDocVectorsDirectory());
    if (fs.exists(outputPath)) {
        sLogger.info("IntDocVectors already exist: skipping!");
        return 0;
    }

    DistributedCache.addCacheFile(new URI(termsFile), conf);
    DistributedCache.addCacheFile(new URI(termIDsFile), conf);
    DistributedCache.addCacheFile(new URI(idToTermFile), conf);

    conf.setJobName("BuildIntDocVectors:" + collectionName);

    conf.setNumMapTasks(mapTasks);
    conf.setNumReduceTasks(0);

    FileInputFormat.setInputPaths(conf, env.getTermDocVectorsDirectory());
    FileOutputFormat.setOutputPath(conf, outputPath);

    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setOutputFormat(SequenceFileOutputFormat.class);
    SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.RECORD);

    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(LazyIntDocVector.class);
    conf.setOutputKeyClass(IntWritable.class);
    conf.setOutputValueClass(LazyIntDocVector.class);

    conf.setMapperClass(MyMapper.class);

    long startTime = System.currentTimeMillis();
    RunningJob job = JobClient.runJob(conf);
    sLogger.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");

    return 0;
}

From source file:ivory.preprocess.BuildIntDocVectorsForwardIndex.java

License:Apache License

public int runTool() throws Exception {
    JobConf conf = new JobConf(getConf(), BuildIntDocVectorsForwardIndex.class);
    FileSystem fs = FileSystem.get(conf);

    String indexPath = conf.get("Ivory.IndexPath");
    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);

    int mapTasks = conf.getInt("Ivory.NumMapTasks", 0);
    String collectionName = env.readCollectionName();
    boolean buildWeighted = conf.getBoolean("Ivory.BuildWeighted", false);

    sLogger.info("Tool: BuildIntDocVectorsIndex");
    sLogger.info(" - IndexPath: " + indexPath);
    sLogger.info(" - CollectionName: " + collectionName);
    sLogger.info(" - BuildWeighted: " + buildWeighted);
    sLogger.info(" - NumMapTasks: " + mapTasks);

    String intDocVectorsPath;// w  w w . j av a  2 s.co  m
    String forwardIndexPath;
    if (buildWeighted) {
        intDocVectorsPath = env.getWeightedIntDocVectorsDirectory();
        forwardIndexPath = env.getWeightedIntDocVectorsForwardIndex();
    } else {
        intDocVectorsPath = env.getIntDocVectorsDirectory();
        forwardIndexPath = env.getIntDocVectorsForwardIndex();
    }

    if (!fs.exists(new Path(intDocVectorsPath))) {
        sLogger.info("Error: IntDocVectors don't exist!");
        return 0;
    }

    if (fs.exists(new Path(forwardIndexPath))) {
        sLogger.info("IntDocVectorIndex already exists: skipping!");
        return 0;
    }

    conf.setJobName("BuildIntDocVectorsForwardIndex:" + collectionName);

    Path inputPath = new Path(intDocVectorsPath);
    FileInputFormat.setInputPaths(conf, inputPath);

    conf.setNumMapTasks(mapTasks);
    conf.setNumReduceTasks(1);

    conf.set("mapred.child.java.opts", "-Xmx2048m");

    conf.setInputFormat(SequenceFileInputFormat.class);
    conf.setMapOutputKeyClass(IntWritable.class);
    conf.setMapOutputValueClass(Text.class);
    conf.setOutputFormat(NullOutputFormat.class);

    conf.setMapRunnerClass(MyMapRunner.class);
    conf.setReducerClass(MyReducer.class);

    JobClient.runJob(conf);

    return 0;
}