Example usage for org.apache.hadoop.mapred FileInputFormat setInputPaths

List of usage examples for org.apache.hadoop.mapred FileInputFormat setInputPaths

  1. HOME
  2. Java
  3. org.apache.hadoop
  4. org.apache.hadoop.mapred.*
  5. FileInputFormat
  6. setInputPaths

Introduction

In this page you can find the example usage for org.apache.hadoop.mapred FileInputFormat setInputPaths.

Prototype

public static void setInputPaths(JobConf conf, Path... inputPaths)

Source Link

Document

Set the array of Path s as the list of inputs for the map-reduce job.

Usage

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;
    }//from   w  w w . j a v  a 2 s  . c om
    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// ww w  .  j av a2 s  .c  om
    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  a va2  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 ww w  .j a v  a 2 s. 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));
    }/*ww  w . jav a2s .c om*/

    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//from   w w  w  . j  av a2s  .  c  o 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;/*from  w ww .  j  av a2  s  . c om*/
    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;
}

From source file:ivory.preprocess.BuildTermDocVectors.java

License:Apache License

@SuppressWarnings("unchecked")
public int runTool() throws Exception {
    // create a new JobConf, inheriting from the configuration of this
    // PowerTool/*from  w w w .ja  va2 s. c  o  m*/
    JobConf conf = new JobConf(getConf(), BuildTermDocVectors.class);
    FileSystem fs = FileSystem.get(conf);

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

    String collectionName = conf.get("Ivory.CollectionName");
    String collectionPath = conf.get("Ivory.CollectionPath");
    String inputFormat = conf.get("Ivory.InputFormat");
    String tokenizer = conf.get("Ivory.Tokenizer");
    String mappingClass = conf.get("Ivory.DocnoMappingClass");

    sLogger.info("PowerTool: BuildTermDocVectors");
    sLogger.info(" - CollectionName: " + collectionName);
    sLogger.info(" - CollectionPath: " + collectionPath);
    sLogger.info(" - InputputFormat: " + inputFormat);
    sLogger.info(" - Tokenizer: " + tokenizer);
    sLogger.info(" - DocnoMappingClass: " + mappingClass);
    sLogger.info(" - NumMapTasks: " + mapTasks);
    sLogger.info(" - NumReduceTasks: " + 0);

    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);
    Path mappingFile = env.getDocnoMappingData();

    if (!fs.exists(mappingFile)) {
        sLogger.error("Error, docno mapping data file " + mappingFile + "doesn't exist!");
        return 0;
    }

    DistributedCache.addCacheFile(mappingFile.toUri(), conf);

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

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

    if (collectionPath.indexOf(",") == -1) {
        FileInputFormat.setInputPaths(conf, new Path(collectionPath));
        sLogger.info("Adding input path " + collectionPath);
    } else {
        String[] paths = collectionPath.split(",");
        for (String p : paths) {
            FileInputFormat.addInputPath(conf, new Path(p));
            sLogger.info("Adding input path " + p);
        }
    }

    Path outputPath = new Path(env.getTermDocVectorsDirectory());
    if (fs.exists(outputPath)) {
        sLogger.info("TermDocVectors already exist: Skipping!");
    } else {
        env.writeCollectionName(collectionName);
        env.writeCollectionPath(collectionPath);
        env.writeInputFormat(inputFormat);
        env.writeDocnoMappingClass(mappingClass);
        env.writeTokenizerClass(tokenizer);

        conf.set("mapred.child.java.opts", "-Xmx2048m");
        conf.setInt("mapred.task.timeout", 60000000);

        FileOutputFormat.setOutputPath(conf, outputPath);

        conf.setInputFormat((Class<? extends InputFormat>) Class.forName(inputFormat));
        conf.setOutputFormat(SequenceFileOutputFormat.class);
        SequenceFileOutputFormat.setOutputCompressionType(conf, SequenceFile.CompressionType.RECORD);

        conf.setMapOutputKeyClass(IntWritable.class);
        conf.setMapOutputValueClass(LazyTermDocVector.class);
        conf.setOutputKeyClass(IntWritable.class);
        conf.setOutputValueClass(LazyTermDocVector.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");

        Counters counters = job.getCounters();

        // write out number of postings
        int collectionDocCount = (int) counters.findCounter(Docs.Total).getCounter();
        env.writeCollectionDocumentCount(collectionDocCount);
    }

    if (fs.exists(env.getDoclengthsData())) {
        sLogger.info("DocLength data exists: Skipping!");
        return 0;
    }

    int collectionDocCount = env.readCollectionDocumentCount();
    long startTime = System.currentTimeMillis();
    writeDoclengthsData(collectionDocCount);
    sLogger.info("Job Finished in " + (System.currentTimeMillis() - startTime) / 1000.0 + " seconds");
    return 0;
}

From source file:ivory.preprocess.BuildTermDocVectorsForwardIndex.java

License:Apache License

public int runTool() throws Exception {
    JobConf conf = new JobConf(getConf(), BuildTermDocVectorsForwardIndex.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("Tool: BuildTermDocVectorsIndex");
    sLogger.info(" - IndexPath: " + indexPath);
    sLogger.info(" - CollectionName: " + collectionName);
    sLogger.info(" - NumMapTasks: " + mapTasks);

    if (!fs.exists(new Path(env.getTermDocVectorsDirectory()))) {
        sLogger.info("Error: TermDocVectors don't exist!");
        return 0;
    }// www.  j  a va 2s . c  om

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

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

    Path inputPath = new Path(env.getTermDocVectorsDirectory());
    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;
}

From source file:ivory.preprocess.BuildTermIdMap.java

License:Apache License

@SuppressWarnings("unused")
public int runTool() throws Exception {
    // create a new JobConf, inheriting from the configuration of this
    // PowerTool/*  w  w w  . j ava 2s. com*/
    JobConf conf = new JobConf(getConf(), BuildTermIdMap.class);
    FileSystem fs = FileSystem.get(conf);

    String indexPath = conf.get("Ivory.IndexPath");
    String collectionName = conf.get("Ivory.CollectionName");

    int mapTasks = conf.getInt("Ivory.NumMapTasks", 0);
    int reduceTasks = 1;
    int minSplitSize = conf.getInt("Ivory.MinSplitSize", 0);

    sLogger.info("PowerTool: BuildTermIdMap");
    sLogger.info(" - CollectionName: " + collectionName);
    sLogger.info(" - IndexPath: " + indexPath);
    sLogger.info(" - NumMapTasks: " + mapTasks);
    sLogger.info(" - NumReduceTasks: " + reduceTasks);

    RetrievalEnvironment env = new RetrievalEnvironment(indexPath, fs);

    if (!fs.exists(new Path(indexPath))) {
        sLogger.error("index path doesn't existing: skipping!");
        return 0;
    }

    Path termsFilePath = new Path(env.getIndexTermsData());
    Path termIDsFilePath = new Path(env.getIndexTermIdsData());
    Path idToTermFilePath = new Path(env.getIndexTermIdMappingData());
    Path dfByTermFilePath = new Path(env.getDfByTermData());
    Path cfByTermFilePath = new Path(env.getCfByTermData());
    Path dfByIntFilePath = new Path(env.getDfByIntData());
    Path cfByIntFilePath = new Path(env.getCfByIntData());

    if (fs.exists(termsFilePath) || fs.exists(termIDsFilePath) || fs.exists(idToTermFilePath)
            || fs.exists(dfByTermFilePath) || fs.exists(cfByTermFilePath) || fs.exists(dfByIntFilePath)
            || fs.exists(cfByIntFilePath)) {
        sLogger.info("term and term id data exist: skipping!");
        return 0;
    }

    Path tmpPath = new Path(env.getTempDirectory());
    fs.delete(tmpPath, true);

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

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

    conf.setInt("Ivory.CollectionTermCount", (int) env.readCollectionTermCount());
    conf.setInt("mapred.min.split.size", minSplitSize);
    conf.set("mapred.child.java.opts", "-Xmx2048m");

    FileInputFormat.setInputPaths(conf, new Path(env.getTermDfCfDirectory()));
    FileOutputFormat.setOutputPath(conf, tmpPath);

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

    conf.setMapOutputKeyClass(Text.class);
    conf.setMapOutputValueClass(PairOfIntLong.class);
    conf.setOutputKeyClass(Text.class);

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

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

    fs.delete(tmpPath, true);

    return 0;
}