Example usage for org.apache.hadoop.fs FileSystem close

List of usage examples for org.apache.hadoop.fs FileSystem close

  1. HOME
  2. Java
  3. org.apache.hadoop
  4. org.apache.hadoop.fs.*
  5. FileSystem
  6. close

Introduction

In this page you can find the example usage for org.apache.hadoop.fs FileSystem close.

Prototype

@Override
public void close() throws IOException

Source Link

Document

Close this FileSystem instance.

Usage

From source file:eagle.dataproc.impl.storm.hdfs.UserProfileGenerationHDFSSpout.java

License:Apache License

public void copyFiles() {
    LOG.info("Inside listFiles()");
    //Configuration conf = new Configuration();
    JobConf conf = new JobConf();
    // _____________ TO TEST THAT CORRECT HADOOP JARs ARE INCLUDED __________________
    ClassLoader cl = ClassLoader.getSystemClassLoader();
    URL[] urls = ((URLClassLoader) cl).getURLs();
    if (LOG.isDebugEnabled()) {
        for (URL url : urls) {
            LOG.debug(url.getFile());//from www  .  j  a  va  2  s .  co m
        }
    }
    // _________________________________________
    String hdfsConnectionStr = configContext.getString("dataSourceConfig.hdfsConnection");
    LOG.info("HDFS connection string: " + hdfsConnectionStr);

    String hdfsPath = configContext.getString("dataSourceConfig.hdfsPath");
    LOG.info("HDFS path: " + hdfsPath);

    String copyToPath = configContext.getString("dataSourceConfig.copyToPath");
    LOG.info("copyToPath: " + copyToPath);
    String srcPathStr = new String("hdfs://" + hdfsConnectionStr + hdfsPath);
    Path srcPath = new Path(srcPathStr);
    LOG.info("listFiles called");
    LOG.info("srcPath: " + srcPath);
    try {
        FileSystem fs = srcPath.getFileSystem(conf);
        /*CompressionCodecFactory codecFactory = new CompressionCodecFactory(conf); 
        CompressionCodec codec = codecFactory.getCodec(srcPath);
        DataInputStream inputStream = new DataInputStream(codec.createInputStream(fs.open(srcPath)));
        */

        Path destPath = new Path(copyToPath);
        LOG.info("Destination path: " + destPath);
        String userListFileName = configContext.getString("dataSourceConfig.userList");
        //loggerHDFSSpout.info("userListFileName: " + userListFileName);
        List<String> userList = getUser(userListFileName);
        for (String user : userList) {
            Path finalSrcPath = new Path(srcPath.getName() + "/" + user);
            fs.copyToLocalFile(finalSrcPath, destPath);
        }
        LOG.info("Copy to local succeed");
        fs.close();

    } catch (IOException e) {
        // TODO Auto-generated catch block
        e.printStackTrace();
    }

}

From source file:eagle.service.security.hdfs.HDFSFileSystem.java

License:Apache License

/**
 * Brows the Files for the specific Path
  */*from   w  ww  . j a v  a  2s  .  co m*/
 * @param filePath
 * @return listOfFiles
 * @throws Exception 
 */
public List<FileStatus> browse(String filePath) throws Exception {
    LOG.info("HDFS File Path   :  " + filePath + "   and EndPoint  : " + hdfsEndPoint);
    FileSystem hdfsFileSystem = null;
    FileStatus[] listStatus;
    try {
        Configuration config = createConfig();
        hdfsFileSystem = getFileSystem(config);
        Path path = new Path(filePath);
        listStatus = hdfsFileSystem.listStatus(path);
    } catch (Exception ex) {
        LOG.error(" Exception when browsing files for the path " + filePath, ex.getMessage());
        throw new Exception(" Exception When browsing Files in HDFS .. Message :  " + ex.getMessage());
    } finally {
        //Close the file system
        if (hdfsFileSystem != null)
            hdfsFileSystem.close();
    }
    return Arrays.asList(listStatus);
}

From source file:eagle.service.security.hdfs.HDFSFileSystem.java

License:Apache License

/**
 * Browse Resources of passed Path and its sub directories
 * Note :: Sub directories only to determine the Sensitivity Type of Root Directories
 * @param filePath//from  w w w .j  a va 2 s . com
 * @return fileSystemResponseObj{ with SubDirectoryMap and ListOf FileStatus}
 * @throws Exception
 */
public HDFSFileSystemResponseEntity browseResources(String filePath) throws Exception {
    LOG.info("HDFS File Path   :  " + filePath + "   and EndPoint  : " + hdfsEndPoint);
    FileSystem hdfsFileSystem = null;
    FileStatus[] listStatus;
    HDFSFileSystemResponseEntity response = new HDFSFileSystemResponseEntity();
    Map<String, List<String>> subdirectoriesMap = new HashMap<String, List<String>>();
    try {
        Configuration config = createConfig();
        hdfsFileSystem = getFileSystem(config);
        Path path = new Path(filePath);
        listStatus = hdfsFileSystem.listStatus(path);
        LOG.info(" Browsing Sub Directories .... ");
        // Browse Sub- directories
        for (FileStatus fileStatus : listStatus) {
            FileStatus[] fileStatusList = null;
            if (fileStatus.isDirectory())
                fileStatusList = hdfsFileSystem.listStatus(new Path(fileStatus.getPath().toUri().getPath()));

            if (fileStatusList != null && fileStatusList.length > 0)
                subdirectoriesMap.put(fileStatus.getPath().toUri().getPath(), /*  Key would be Parent */
                        getSubDirectories(fileStatusList) /*  Value Would be Child Paths */);
        }
        response.setFileList(Arrays.asList(listStatus));
        response.setSubDirectoriesMap(subdirectoriesMap);

    } catch (Exception ex) {
        LOG.error(" Exception when browsing files for the path " + filePath, ex.getMessage());
        throw new Exception(
                " Exception When browsing Files/Directories in HDFS .. Message :  " + ex.getMessage());
    } finally {
        //Close the file system
        if (hdfsFileSystem != null)
            hdfsFileSystem.close();
    }
    return response;
}

From source file:edu.emory.bmi.medicurator.storage.HdfsStorage.java

License:Apache License

/**
 * save data from InputStream to a relative path
 * required parent directories will be created if not exists
 * @param path String path/*from  www. j  a va 2  s .c  o  m*/
 * @param in InputStream
  * @return boolean
  */
public boolean saveToPath(String path, InputStream in) {
    try {
        Configuration conf = new Configuration();
        FileSystem hdfs = FileSystem.get(new URI(Constants.HDFS_URI), conf);
        Path dst = new Path(hdfsPath + path);

        //Create a new file and write data to it.
        hdfs.mkdirs(dst.getParent());
        FSDataOutputStream out = hdfs.create(dst, true);
        byte[] b = new byte[1024];
        int numBytes = 0;
        while ((numBytes = in.read(b)) > 0) {
            out.write(b, 0, numBytes);
        }
        //Close all the file descripters
        out.close();
        hdfs.close();
        return true;
    } catch (Exception e) {
        System.out.println("[ERROR]  Exception occurs when HdfsStorage saveToPath (" + path + ")" + e);
    }
    return false;
}

From source file:edu.nyu.vida.data_polygamy.feature_identification.IndexCreation.java

License:BSD License

/**
 * @param args//from  w  ww  .ja  v  a2s.  c  o  m
 */
@SuppressWarnings({ "deprecation" })
public static void main(String[] args) throws IOException, InterruptedException, ClassNotFoundException {

    Options options = new Options();

    Option forceOption = new Option("f", "force", false,
            "force the computation of the index and events " + "even if files already exist");
    forceOption.setRequired(false);
    options.addOption(forceOption);

    Option thresholdOption = new Option("t", "use-custom-thresholds", false,
            "use custom thresholds for regular and rare events, defined in HDFS_HOME/"
                    + FrameworkUtils.thresholdDir + " file");
    thresholdOption.setRequired(false);
    options.addOption(thresholdOption);

    Option gOption = new Option("g", "group", true,
            "set group of datasets for which the indices and events" + " will be computed");
    gOption.setRequired(true);
    gOption.setArgName("GROUP");
    gOption.setArgs(Option.UNLIMITED_VALUES);
    options.addOption(gOption);

    Option machineOption = new Option("m", "machine", true, "machine identifier");
    machineOption.setRequired(true);
    machineOption.setArgName("MACHINE");
    machineOption.setArgs(1);
    options.addOption(machineOption);

    Option nodesOption = new Option("n", "nodes", true, "number of nodes");
    nodesOption.setRequired(true);
    nodesOption.setArgName("NODES");
    nodesOption.setArgs(1);
    options.addOption(nodesOption);

    Option s3Option = new Option("s3", "s3", false, "data on Amazon S3");
    s3Option.setRequired(false);
    options.addOption(s3Option);

    Option awsAccessKeyIdOption = new Option("aws_id", "aws-id", true,
            "aws access key id; " + "this is required if the execution is on aws");
    awsAccessKeyIdOption.setRequired(false);
    awsAccessKeyIdOption.setArgName("AWS-ACCESS-KEY-ID");
    awsAccessKeyIdOption.setArgs(1);
    options.addOption(awsAccessKeyIdOption);

    Option awsSecretAccessKeyOption = new Option("aws_key", "aws-id", true,
            "aws secrect access key; " + "this is required if the execution is on aws");
    awsSecretAccessKeyOption.setRequired(false);
    awsSecretAccessKeyOption.setArgName("AWS-SECRET-ACCESS-KEY");
    awsSecretAccessKeyOption.setArgs(1);
    options.addOption(awsSecretAccessKeyOption);

    Option bucketOption = new Option("b", "s3-bucket", true,
            "bucket on s3; " + "this is required if the execution is on aws");
    bucketOption.setRequired(false);
    bucketOption.setArgName("S3-BUCKET");
    bucketOption.setArgs(1);
    options.addOption(bucketOption);

    Option helpOption = new Option("h", "help", false, "display this message");
    helpOption.setRequired(false);
    options.addOption(helpOption);

    HelpFormatter formatter = new HelpFormatter();
    CommandLineParser parser = new PosixParser();
    CommandLine cmd = null;

    try {
        cmd = parser.parse(options, args);
    } catch (ParseException e) {
        formatter.printHelp("hadoop jar data-polygamy.jar "
                + "edu.nyu.vida.data_polygamy.feature_identification.IndexCreation", options, true);
        System.exit(0);
    }

    if (cmd.hasOption("h")) {
        formatter.printHelp("hadoop jar data-polygamy.jar "
                + "edu.nyu.vida.data_polygamy.feature_identification.IndexCreation", options, true);
        System.exit(0);
    }

    boolean s3 = cmd.hasOption("s3");
    String s3bucket = "";
    String awsAccessKeyId = "";
    String awsSecretAccessKey = "";

    if (s3) {
        if ((!cmd.hasOption("aws_id")) || (!cmd.hasOption("aws_key")) || (!cmd.hasOption("b"))) {
            System.out.println(
                    "Arguments 'aws_id', 'aws_key', and 'b'" + " are mandatory if execution is on AWS.");
            formatter.printHelp("hadoop jar data-polygamy.jar "
                    + "edu.nyu.vida.data_polygamy.feature_identification.IndexCreation", options, true);
            System.exit(0);
        }
        s3bucket = cmd.getOptionValue("b");
        awsAccessKeyId = cmd.getOptionValue("aws_id");
        awsSecretAccessKey = cmd.getOptionValue("aws_key");
    }

    boolean snappyCompression = false;
    boolean bzip2Compression = false;
    String machine = cmd.getOptionValue("m");
    int nbNodes = Integer.parseInt(cmd.getOptionValue("n"));

    Configuration s3conf = new Configuration();
    if (s3) {
        s3conf.set("fs.s3.awsAccessKeyId", awsAccessKeyId);
        s3conf.set("fs.s3.awsSecretAccessKey", awsSecretAccessKey);
        s3conf.set("bucket", s3bucket);
    }

    String datasetNames = "";
    String datasetIds = "";

    ArrayList<String> shortDataset = new ArrayList<String>();
    ArrayList<String> shortDatasetIndex = new ArrayList<String>();
    HashMap<String, String> datasetAgg = new HashMap<String, String>();
    HashMap<String, String> datasetId = new HashMap<String, String>();
    HashMap<String, HashMap<Integer, Double>> datasetRegThreshold = new HashMap<String, HashMap<Integer, Double>>();
    HashMap<String, HashMap<Integer, Double>> datasetRareThreshold = new HashMap<String, HashMap<Integer, Double>>();

    Path path = null;
    FileSystem fs = FileSystem.get(new Configuration());
    BufferedReader br;

    boolean removeExistingFiles = cmd.hasOption("f");
    boolean isThresholdUserDefined = cmd.hasOption("t");

    for (String dataset : cmd.getOptionValues("g")) {

        // getting aggregates
        String[] aggregate = FrameworkUtils.searchAggregates(dataset, s3conf, s3);
        if (aggregate.length == 0) {
            System.out.println("No aggregates found for " + dataset + ".");
            continue;
        }

        // getting aggregates header
        String aggregatesHeaderFileName = FrameworkUtils.searchAggregatesHeader(dataset, s3conf, s3);
        if (aggregatesHeaderFileName == null) {
            System.out.println("No aggregate header for " + dataset);
            continue;
        }

        String aggregatesHeader = s3bucket + FrameworkUtils.preProcessingDir + "/" + aggregatesHeaderFileName;

        shortDataset.add(dataset);
        datasetId.put(dataset, null);

        if (s3) {
            path = new Path(aggregatesHeader);
            fs = FileSystem.get(path.toUri(), s3conf);
        } else {
            path = new Path(fs.getHomeDirectory() + "/" + aggregatesHeader);
        }

        br = new BufferedReader(new InputStreamReader(fs.open(path)));
        datasetAgg.put(dataset, br.readLine().split("\t")[1]);
        br.close();
        if (s3)
            fs.close();
    }

    if (shortDataset.size() == 0) {
        System.out.println("No datasets to process.");
        System.exit(0);
    }

    // getting dataset id

    if (s3) {
        path = new Path(s3bucket + FrameworkUtils.datasetsIndexDir);
        fs = FileSystem.get(path.toUri(), s3conf);
    } else {
        path = new Path(fs.getHomeDirectory() + "/" + FrameworkUtils.datasetsIndexDir);
    }
    br = new BufferedReader(new InputStreamReader(fs.open(path)));
    String line = br.readLine();
    while (line != null) {
        String[] dt = line.split("\t");
        if (datasetId.containsKey(dt[0])) {
            datasetId.put(dt[0], dt[1]);
            datasetNames += dt[0] + ",";
            datasetIds += dt[1] + ",";
        }
        line = br.readLine();
    }
    br.close();

    datasetNames = datasetNames.substring(0, datasetNames.length() - 1);
    datasetIds = datasetIds.substring(0, datasetIds.length() - 1);
    Iterator<String> it = shortDataset.iterator();
    while (it.hasNext()) {
        String dataset = it.next();
        if (datasetId.get(dataset) == null) {
            System.out.println("No dataset id for " + dataset);
            System.exit(0);
        }
    }

    // getting user defined thresholds

    if (isThresholdUserDefined) {
        if (s3) {
            path = new Path(s3bucket + FrameworkUtils.thresholdDir);
            fs = FileSystem.get(path.toUri(), s3conf);
        } else {
            path = new Path(fs.getHomeDirectory() + "/" + FrameworkUtils.thresholdDir);
        }
        br = new BufferedReader(new InputStreamReader(fs.open(path)));
        line = br.readLine();
        while (line != null) {
            // getting dataset name
            String dataset = line.trim();
            HashMap<Integer, Double> regThresholds = new HashMap<Integer, Double>();
            HashMap<Integer, Double> rareThresholds = new HashMap<Integer, Double>();
            line = br.readLine();
            while ((line != null) && (line.split("\t").length > 1)) {
                // getting attribute ids and thresholds
                String[] keyVals = line.trim().split("\t");
                int att = Integer.parseInt(keyVals[0].trim());
                regThresholds.put(att, Double.parseDouble(keyVals[1].trim()));
                rareThresholds.put(att, Double.parseDouble(keyVals[2].trim()));
                line = br.readLine();
            }
            datasetRegThreshold.put(dataset, regThresholds);
            datasetRareThreshold.put(dataset, rareThresholds);
        }
        br.close();
    }
    if (s3)
        fs.close();

    // datasets that will use existing merge tree
    ArrayList<String> useMergeTree = new ArrayList<String>();

    // creating index for each spatio-temporal resolution

    FrameworkUtils.createDir(s3bucket + FrameworkUtils.indexDir, s3conf, s3);

    HashSet<String> input = new HashSet<String>();

    for (String dataset : shortDataset) {

        String indexCreationOutputFileName = s3bucket + FrameworkUtils.indexDir + "/" + dataset + "/";
        String mergeTreeFileName = s3bucket + FrameworkUtils.mergeTreeDir + "/" + dataset + "/";

        if (removeExistingFiles) {
            FrameworkUtils.removeFile(indexCreationOutputFileName, s3conf, s3);
            FrameworkUtils.removeFile(mergeTreeFileName, s3conf, s3);
            FrameworkUtils.createDir(mergeTreeFileName, s3conf, s3);
        } else if (datasetRegThreshold.containsKey(dataset)) {
            FrameworkUtils.removeFile(indexCreationOutputFileName, s3conf, s3);
            if (FrameworkUtils.fileExists(mergeTreeFileName, s3conf, s3)) {
                useMergeTree.add(dataset);
            }
        }

        if (!FrameworkUtils.fileExists(indexCreationOutputFileName, s3conf, s3)) {
            input.add(s3bucket + FrameworkUtils.aggregatesDir + "/" + dataset);
            shortDatasetIndex.add(dataset);
        }

    }

    if (input.isEmpty()) {
        System.out.println("All the input datasets have indices.");
        System.out.println("Use -f in the beginning of the command line to force the computation.");
        System.exit(0);
    }

    String aggregateDatasets = "";
    it = input.iterator();
    while (it.hasNext()) {
        aggregateDatasets += it.next() + ",";
    }

    Job icJob = null;
    Configuration icConf = new Configuration();
    Machine machineConf = new Machine(machine, nbNodes);

    String jobName = "index";
    String indexOutputDir = s3bucket + FrameworkUtils.indexDir + "/tmp/";

    FrameworkUtils.removeFile(indexOutputDir, s3conf, s3);

    icConf.set("dataset-name", datasetNames);
    icConf.set("dataset-id", datasetIds);

    if (!useMergeTree.isEmpty()) {
        String useMergeTreeStr = "";
        for (String dt : useMergeTree) {
            useMergeTreeStr += dt + ",";
        }
        icConf.set("use-merge-tree", useMergeTreeStr.substring(0, useMergeTreeStr.length() - 1));
    }

    for (int i = 0; i < shortDataset.size(); i++) {
        String dataset = shortDataset.get(i);
        String id = datasetId.get(dataset);
        icConf.set("dataset-" + id + "-aggregates", datasetAgg.get(dataset));
        if (datasetRegThreshold.containsKey(dataset)) {
            HashMap<Integer, Double> regThresholds = datasetRegThreshold.get(dataset);
            String thresholds = "";
            for (int att : regThresholds.keySet()) {
                thresholds += String.valueOf(att) + "-" + String.valueOf(regThresholds.get(att)) + ",";
            }
            icConf.set("regular-" + id, thresholds.substring(0, thresholds.length() - 1));
        }

        if (datasetRareThreshold.containsKey(dataset)) {
            HashMap<Integer, Double> rareThresholds = datasetRareThreshold.get(dataset);
            String thresholds = "";
            for (int att : rareThresholds.keySet()) {
                thresholds += String.valueOf(att) + "-" + String.valueOf(rareThresholds.get(att)) + ",";
            }
            icConf.set("rare-" + id, thresholds.substring(0, thresholds.length() - 1));
        }
    }

    icConf.set("mapreduce.tasktracker.map.tasks.maximum", String.valueOf(machineConf.getMaximumTasks()));
    icConf.set("mapreduce.tasktracker.reduce.tasks.maximum", String.valueOf(machineConf.getMaximumTasks()));
    icConf.set("mapreduce.jobtracker.maxtasks.perjob", "-1");
    icConf.set("mapreduce.reduce.shuffle.parallelcopies", "20");
    icConf.set("mapreduce.input.fileinputformat.split.minsize", "0");
    icConf.set("mapreduce.task.io.sort.mb", "200");
    icConf.set("mapreduce.task.io.sort.factor", "100");
    //icConf.set("mapreduce.task.timeout", "1800000");
    machineConf.setMachineConfiguration(icConf);

    if (s3) {
        machineConf.setMachineConfiguration(icConf);
        icConf.set("fs.s3.awsAccessKeyId", awsAccessKeyId);
        icConf.set("fs.s3.awsSecretAccessKey", awsSecretAccessKey);
        icConf.set("bucket", s3bucket);
    }

    if (snappyCompression) {
        icConf.set("mapreduce.map.output.compress", "true");
        icConf.set("mapreduce.map.output.compress.codec", "org.apache.hadoop.io.compress.SnappyCodec");
        //icConf.set("mapreduce.output.fileoutputformat.compress.codec", "org.apache.hadoop.io.compress.SnappyCodec");
    }
    if (bzip2Compression) {
        icConf.set("mapreduce.map.output.compress", "true");
        icConf.set("mapreduce.map.output.compress.codec", "org.apache.hadoop.io.compress.BZip2Codec");
        //icConf.set("mapreduce.output.fileoutputformat.compress.codec", "org.apache.hadoop.io.compress.BZip2Codec");
    }

    icJob = new Job(icConf);
    icJob.setJobName(jobName);

    icJob.setMapOutputKeyClass(AttributeResolutionWritable.class);
    icJob.setMapOutputValueClass(SpatioTemporalFloatWritable.class);
    icJob.setOutputKeyClass(AttributeResolutionWritable.class);
    icJob.setOutputValueClass(TopologyTimeSeriesWritable.class);
    //icJob.setOutputKeyClass(Text.class);
    //icJob.setOutputValueClass(Text.class);

    icJob.setMapperClass(IndexCreationMapper.class);
    icJob.setReducerClass(IndexCreationReducer.class);
    icJob.setNumReduceTasks(machineConf.getNumberReduces());

    icJob.setInputFormatClass(SequenceFileInputFormat.class);
    //icJob.setOutputFormatClass(SequenceFileOutputFormat.class);
    LazyOutputFormat.setOutputFormatClass(icJob, SequenceFileOutputFormat.class);
    //LazyOutputFormat.setOutputFormatClass(icJob, TextOutputFormat.class);
    SequenceFileOutputFormat.setCompressOutput(icJob, true);
    SequenceFileOutputFormat.setOutputCompressionType(icJob, CompressionType.BLOCK);

    FileInputFormat.setInputDirRecursive(icJob, true);
    FileInputFormat.setInputPaths(icJob, aggregateDatasets.substring(0, aggregateDatasets.length() - 1));
    FileOutputFormat.setOutputPath(icJob, new Path(indexOutputDir));

    icJob.setJarByClass(IndexCreation.class);

    long start = System.currentTimeMillis();
    icJob.submit();
    icJob.waitForCompletion(true);
    System.out.println(jobName + "\t" + (System.currentTimeMillis() - start));

    // moving files to right place
    for (String dataset : shortDatasetIndex) {
        String from = s3bucket + FrameworkUtils.indexDir + "/tmp/" + dataset + "/";
        String to = s3bucket + FrameworkUtils.indexDir + "/" + dataset + "/";
        FrameworkUtils.renameFile(from, to, s3conf, s3);
    }

}

From source file:edu.nyu.vida.data_polygamy.pre_processing.PreProcessingMapper.java

License:BSD License

@Override
public void setup(Context context) throws IOException, InterruptedException {

    conf = context.getConfiguration();//from  ww  w  .j  av a  2  s .co  m

    String bucket = conf.get("bucket", "");
    if (bucket.equals(""))
        s3 = false;

    // defaults
    Path defaults = new Path(conf.get("defaults", ""));

    FileSystem fs = null;
    BufferedReader br = null;

    if (!s3)
        fs = FileSystem.get(new Configuration());
    else
        fs = FileSystem.get(defaults.toUri(), conf);
    br = new BufferedReader(new InputStreamReader(fs.open(defaults)));
    paramDefaults = br.readLine().split(",");
    br.close();
    if (s3)
        fs.close();

    temporalResolution = utils.temporalResolution(conf.get("temporal-resolution"));
    spatialResolution = utils.spatialResolution(conf.get("spatial-resolution"));
    currentSpatialResolution = utils.spatialResolution(conf.get("current-spatial-resolution"));
    gridResolution = (conf.get("grid-resolution", "").equals("")) ? 0
            : Integer.parseInt(conf.get("grid-resolution", ""));
    sizeSpatioTemp = Integer.parseInt(conf.get("size-spatio-temporal", "0"));

    // positions
    String[] temporalArray = (conf.get("temporal-pos", "").equals("")) ? new String[0]
            : conf.get("temporal-pos", "").split(",");
    String[] spatialPosArray = (conf.get("spatial-pos", "").equals("")) ? new String[0]
            : conf.get("spatial-pos", "").split(",");
    String[] xPositionsArray = (conf.get("xPositions", "").equals("")) ? new String[0]
            : conf.get("xPositions", "").split(",");
    String[] yPositionsArray = (conf.get("yPositions", "").equals("")) ? new String[0]
            : conf.get("yPositions", "").split(",");

    temporalPos = FrameworkUtils.getIntArray(temporalArray);
    spatialPos = FrameworkUtils.getIntArray(spatialPosArray);
    xPositions = FrameworkUtils.getIntArray(xPositionsArray);
    yPositions = FrameworkUtils.getIntArray(yPositionsArray);

    // reading header
    Path header = new Path(conf.get("header", ""));

    if (s3)
        fs = FileSystem.get(header.toUri(), conf);
    br = new BufferedReader(new InputStreamReader(fs.open(header)));
    parameterNames = br.readLine().split(",", -1);
    br.close();
    if (s3)
        fs.close();

    /**
     * Spatial Resolution
     */

    switch (currentSpatialResolution) {

    case FrameworkUtils.POINTS:
        spatialTranslation = SpatialResolutionUtils.pointsResolution(spatialResolution, gridResolution,
                xPositions, yPositions, conf);
        break;
    case FrameworkUtils.NBHD:
        spatialTranslation = SpatialResolutionUtils.nbhdResolution(spatialResolution, spatialPos);
        break;
    case FrameworkUtils.ZIP:
        spatialTranslation = SpatialResolutionUtils.zipResolution(spatialResolution, spatialPos, true, conf);
        break;
    case FrameworkUtils.GRID:
        spatialTranslation = SpatialResolutionUtils.gridResolution(spatialResolution, spatialPos);
        break;
    case FrameworkUtils.CITY:
        spatialTranslation = new ToCity(spatialPos);
        break;
    default:
        System.out.println("Something is wrong...");
        System.exit(-1);
        break;
    }

}

From source file:edu.nyu.vida.data_polygamy.pre_processing.PreProcessingMapper.java

License:BSD License

@Override
public void cleanup(Context context) throws IOException {

    if (records > 0) {
        conf = context.getConfiguration();
        Path headerFile = null;/*from  w ww .  ja  v  a  2  s  .  co m*/
        FileSystem fs = null;

        if (s3) {
            headerFile = new Path(conf.get("aggregates", ""));
            fs = FileSystem.get(headerFile.toUri(), conf);
        } else {
            fs = FileSystem.get(new Configuration());
            headerFile = new Path(
                    fs.getHomeDirectory() + "/" + context.getConfiguration().get("aggregates", ""));
        }

        // we cannot have multiple mappers writing the same file
        if (fs.exists(headerFile)) {
            if (s3)
                fs.close();
            return;
        }
        FSDataOutputStream fsDataOutputStream = fs.create(headerFile);

        String output = "";

        for (int i = 0; i < keyNames.length; i++) {
            if (keyNames[i] != null)
                output += keyNames[i] + ",";
        }

        output = output.substring(0, output.length() - 1) + "\t";

        Iterator<Integer> it = aggregatesIndex.iterator();
        while (it.hasNext()) {
            Integer index = it.next();
            output += aggregates.get(index) + ",";
        }

        BufferedWriter bw = new BufferedWriter(new OutputStreamWriter(fsDataOutputStream));
        bw.write(output.substring(0, output.length() - 1) + "\n");
        bw.write(String.valueOf(nbParameters));
        bw.close();
        if (s3)
            fs.close();
    }
}

From source file:edu.nyu.vida.data_polygamy.relationship_computation.Relationship.java

License:BSD License

public static void addDatasets(String[] groupCmd, ArrayList<String> group, ArrayList<String> datasets,
        HashMap<String, String> datasetAgg, Path path, FileSystem fs, Configuration s3conf, boolean s3,
        String s3bucket) throws IOException {
    for (String dataset : groupCmd) {

        // getting indices

        String[] index = FrameworkUtils.searchIndex(dataset, s3conf, s3);
        if (index.length == 0) {
            System.out.println("No indices found for " + dataset + ".");
            continue;
        }/* w  w w. ja  va 2 s .co  m*/

        // getting aggregate headers

        String aggregatesHeaderFileName1 = FrameworkUtils.searchAggregatesHeader(dataset, s3conf, s3);
        if (aggregatesHeaderFileName1 == null) {
            System.out.println("No aggregate header for " + dataset);
            continue;
        }
        String aggregatesHeader1 = s3bucket + FrameworkUtils.preProcessingDir + "/" + aggregatesHeaderFileName1;

        if (s3) {
            path = new Path(aggregatesHeader1);
            fs = FileSystem.get(path.toUri(), s3conf);
        } else {
            path = new Path(fs.getHomeDirectory() + "/" + aggregatesHeader1);
        }
        BufferedReader br = new BufferedReader(new InputStreamReader(fs.open(path)));
        datasetAgg.put(dataset, br.readLine().split("\t")[1]);
        br.close();
        if (s3)
            fs.close();

        datasets.add(dataset);
        group.add(dataset);
    }
}

From source file:edu.nyu.vida.data_polygamy.relationship_computation.Relationship.java

License:BSD License

/**
 * @param args//from ww  w. java  2 s.  com
 * @throws ParseException 
 */
@SuppressWarnings({ "deprecation" })
public static void main(String[] args) throws IOException, InterruptedException, ClassNotFoundException {

    Options options = new Options();

    Option forceOption = new Option("f", "force", false,
            "force the computation of the relationship " + "even if files already exist");
    forceOption.setRequired(false);
    options.addOption(forceOption);

    Option scoreOption = new Option("sc", "score", true, "set threhsold for relationship score");
    scoreOption.setRequired(false);
    scoreOption.setArgName("SCORE THRESHOLD");
    options.addOption(scoreOption);

    Option strengthOption = new Option("st", "strength", true, "set threhsold for relationship strength");
    strengthOption.setRequired(false);
    strengthOption.setArgName("STRENGTH THRESHOLD");
    options.addOption(strengthOption);

    Option completeRandomizationOption = new Option("c", "complete-randomization", false,
            "use complete randomization when performing significance tests");
    completeRandomizationOption.setRequired(false);
    options.addOption(completeRandomizationOption);

    Option idOption = new Option("id", "ids", false, "output id instead of names for datasets and attributes");
    idOption.setRequired(false);
    options.addOption(idOption);

    Option g1Option = new Option("g1", "first-group", true, "set first group of datasets");
    g1Option.setRequired(true);
    g1Option.setArgName("FIRST GROUP");
    g1Option.setArgs(Option.UNLIMITED_VALUES);
    options.addOption(g1Option);

    Option g2Option = new Option("g2", "second-group", true, "set second group of datasets");
    g2Option.setRequired(false);
    g2Option.setArgName("SECOND GROUP");
    g2Option.setArgs(Option.UNLIMITED_VALUES);
    options.addOption(g2Option);

    Option machineOption = new Option("m", "machine", true, "machine identifier");
    machineOption.setRequired(true);
    machineOption.setArgName("MACHINE");
    machineOption.setArgs(1);
    options.addOption(machineOption);

    Option nodesOption = new Option("n", "nodes", true, "number of nodes");
    nodesOption.setRequired(true);
    nodesOption.setArgName("NODES");
    nodesOption.setArgs(1);
    options.addOption(nodesOption);

    Option s3Option = new Option("s3", "s3", false, "data on Amazon S3");
    s3Option.setRequired(false);
    options.addOption(s3Option);

    Option awsAccessKeyIdOption = new Option("aws_id", "aws-id", true,
            "aws access key id; " + "this is required if the execution is on aws");
    awsAccessKeyIdOption.setRequired(false);
    awsAccessKeyIdOption.setArgName("AWS-ACCESS-KEY-ID");
    awsAccessKeyIdOption.setArgs(1);
    options.addOption(awsAccessKeyIdOption);

    Option awsSecretAccessKeyOption = new Option("aws_key", "aws-id", true,
            "aws secrect access key; " + "this is required if the execution is on aws");
    awsSecretAccessKeyOption.setRequired(false);
    awsSecretAccessKeyOption.setArgName("AWS-SECRET-ACCESS-KEY");
    awsSecretAccessKeyOption.setArgs(1);
    options.addOption(awsSecretAccessKeyOption);

    Option bucketOption = new Option("b", "s3-bucket", true,
            "bucket on s3; " + "this is required if the execution is on aws");
    bucketOption.setRequired(false);
    bucketOption.setArgName("S3-BUCKET");
    bucketOption.setArgs(1);
    options.addOption(bucketOption);

    Option helpOption = new Option("h", "help", false, "display this message");
    helpOption.setRequired(false);
    options.addOption(helpOption);

    Option removeOption = new Option("r", "remove-not-significant", false,
            "remove relationships that are not" + "significant from the final output");
    removeOption.setRequired(false);
    options.addOption(removeOption);

    HelpFormatter formatter = new HelpFormatter();
    CommandLineParser parser = new PosixParser();
    CommandLine cmd = null;

    try {
        cmd = parser.parse(options, args);
    } catch (ParseException e) {
        formatter.printHelp("hadoop jar data-polygamy.jar "
                + "edu.nyu.vida.data_polygamy.relationship_computation.Relationship", options, true);
        System.exit(0);
    }

    if (cmd.hasOption("h")) {
        formatter.printHelp("hadoop jar data-polygamy.jar "
                + "edu.nyu.vida.data_polygamy.relationship_computation.Relationship", options, true);
        System.exit(0);
    }

    boolean s3 = cmd.hasOption("s3");
    String s3bucket = "";
    String awsAccessKeyId = "";
    String awsSecretAccessKey = "";

    if (s3) {
        if ((!cmd.hasOption("aws_id")) || (!cmd.hasOption("aws_key")) || (!cmd.hasOption("b"))) {
            System.out.println(
                    "Arguments 'aws_id', 'aws_key', and 'b'" + " are mandatory if execution is on AWS.");
            formatter.printHelp(
                    "hadoop jar data-polygamy.jar "
                            + "edu.nyu.vida.data_polygamy.relationship_computation.Relationship",
                    options, true);
            System.exit(0);
        }
        s3bucket = cmd.getOptionValue("b");
        awsAccessKeyId = cmd.getOptionValue("aws_id");
        awsSecretAccessKey = cmd.getOptionValue("aws_key");
    }

    boolean snappyCompression = false;
    boolean bzip2Compression = false;
    String machine = cmd.getOptionValue("m");
    int nbNodes = Integer.parseInt(cmd.getOptionValue("n"));

    Configuration s3conf = new Configuration();
    if (s3) {
        s3conf.set("fs.s3.awsAccessKeyId", awsAccessKeyId);
        s3conf.set("fs.s3.awsSecretAccessKey", awsSecretAccessKey);
        s3conf.set("bucket", s3bucket);
    }

    Path path = null;
    FileSystem fs = FileSystem.get(new Configuration());

    ArrayList<String> shortDataset = new ArrayList<String>();
    ArrayList<String> firstGroup = new ArrayList<String>();
    ArrayList<String> secondGroup = new ArrayList<String>();
    HashMap<String, String> datasetAgg = new HashMap<String, String>();

    boolean removeNotSignificant = cmd.hasOption("r");
    boolean removeExistingFiles = cmd.hasOption("f");
    boolean completeRandomization = cmd.hasOption("c");
    boolean hasScoreThreshold = cmd.hasOption("sc");
    boolean hasStrengthThreshold = cmd.hasOption("st");
    boolean outputIds = cmd.hasOption("id");
    String scoreThreshold = hasScoreThreshold ? cmd.getOptionValue("sc") : "";
    String strengthThreshold = hasStrengthThreshold ? cmd.getOptionValue("st") : "";

    // all datasets
    ArrayList<String> all_datasets = new ArrayList<String>();
    if (s3) {
        path = new Path(s3bucket + FrameworkUtils.datasetsIndexDir);
        fs = FileSystem.get(path.toUri(), s3conf);
    } else {
        path = new Path(fs.getHomeDirectory() + "/" + FrameworkUtils.datasetsIndexDir);
    }
    BufferedReader br = new BufferedReader(new InputStreamReader(fs.open(path)));
    String line = br.readLine();
    while (line != null) {
        all_datasets.add(line.split("\t")[0]);
        line = br.readLine();
    }
    br.close();
    if (s3)
        fs.close();
    String[] all_datasets_array = new String[all_datasets.size()];
    all_datasets.toArray(all_datasets_array);

    String[] firstGroupCmd = cmd.getOptionValues("g1");
    String[] secondGroupCmd = cmd.hasOption("g2") ? cmd.getOptionValues("g2") : all_datasets_array;
    addDatasets(firstGroupCmd, firstGroup, shortDataset, datasetAgg, path, fs, s3conf, s3, s3bucket);
    addDatasets(secondGroupCmd, secondGroup, shortDataset, datasetAgg, path, fs, s3conf, s3, s3bucket);

    if (shortDataset.size() == 0) {
        System.out.println("No datasets to process.");
        System.exit(0);
    }

    if (firstGroup.isEmpty()) {
        System.out.println("No indices from datasets in G1.");
        System.exit(0);
    }

    if (secondGroup.isEmpty()) {
        System.out.println("No indices from datasets in G2.");
        System.exit(0);
    }

    // getting dataset ids

    String datasetNames = "";
    String datasetIds = "";
    HashMap<String, String> datasetId = new HashMap<String, String>();
    Iterator<String> it = shortDataset.iterator();
    while (it.hasNext()) {
        datasetId.put(it.next(), null);
    }

    if (s3) {
        path = new Path(s3bucket + FrameworkUtils.datasetsIndexDir);
        fs = FileSystem.get(path.toUri(), s3conf);
    } else {
        path = new Path(fs.getHomeDirectory() + "/" + FrameworkUtils.datasetsIndexDir);
    }
    br = new BufferedReader(new InputStreamReader(fs.open(path)));
    line = br.readLine();
    while (line != null) {
        String[] dt = line.split("\t");
        all_datasets.add(dt[0]);
        if (datasetId.containsKey(dt[0])) {
            datasetId.put(dt[0], dt[1]);
            datasetNames += dt[0] + ",";
            datasetIds += dt[1] + ",";
        }
        line = br.readLine();
    }
    br.close();
    if (s3)
        fs.close();

    datasetNames = datasetNames.substring(0, datasetNames.length() - 1);
    datasetIds = datasetIds.substring(0, datasetIds.length() - 1);
    it = shortDataset.iterator();
    while (it.hasNext()) {
        String dataset = it.next();
        if (datasetId.get(dataset) == null) {
            System.out.println("No dataset id for " + dataset);
            System.exit(0);
        }
    }

    String firstGroupStr = "";
    String secondGroupStr = "";
    for (String dataset : firstGroup) {
        firstGroupStr += datasetId.get(dataset) + ",";
    }
    for (String dataset : secondGroup) {
        secondGroupStr += datasetId.get(dataset) + ",";
    }
    firstGroupStr = firstGroupStr.substring(0, firstGroupStr.length() - 1);
    secondGroupStr = secondGroupStr.substring(0, secondGroupStr.length() - 1);

    String relationshipsDir = "";
    if (outputIds) {
        relationshipsDir = FrameworkUtils.relationshipsIdsDir;
    } else {
        relationshipsDir = FrameworkUtils.relationshipsDir;
    }

    FrameworkUtils.createDir(s3bucket + relationshipsDir, s3conf, s3);

    String random = completeRandomization ? "complete" : "restricted";

    String indexInputDirs = "";
    String noRelationship = "";

    HashSet<String> dirs = new HashSet<String>();

    String dataset1;
    String dataset2;
    String datasetId1;
    String datasetId2;
    for (int i = 0; i < firstGroup.size(); i++) {
        for (int j = 0; j < secondGroup.size(); j++) {

            if (Integer.parseInt(datasetId.get(firstGroup.get(i))) < Integer
                    .parseInt(datasetId.get(secondGroup.get(j)))) {
                dataset1 = firstGroup.get(i);
                dataset2 = secondGroup.get(j);
            } else {
                dataset1 = secondGroup.get(j);
                dataset2 = firstGroup.get(i);
            }

            datasetId1 = datasetId.get(dataset1);
            datasetId2 = datasetId.get(dataset2);

            if (dataset1.equals(dataset2))
                continue;
            String correlationOutputFileName = s3bucket + relationshipsDir + "/" + dataset1 + "-" + dataset2
                    + "/";

            if (removeExistingFiles) {
                FrameworkUtils.removeFile(correlationOutputFileName, s3conf, s3);
            }
            if (!FrameworkUtils.fileExists(correlationOutputFileName, s3conf, s3)) {
                dirs.add(s3bucket + FrameworkUtils.indexDir + "/" + dataset1);
                dirs.add(s3bucket + FrameworkUtils.indexDir + "/" + dataset2);
            } else {
                noRelationship += datasetId1 + "-" + datasetId2 + ",";
            }
        }
    }

    if (dirs.isEmpty()) {
        System.out.println("All the relationships were already computed.");
        System.out.println("Use -f in the beginning of the command line to force the computation.");
        System.exit(0);
    }

    for (String dir : dirs) {
        indexInputDirs += dir + ",";
    }

    Configuration conf = new Configuration();
    Machine machineConf = new Machine(machine, nbNodes);

    String jobName = "relationship" + "-" + random;
    String relationshipOutputDir = s3bucket + relationshipsDir + "/tmp/";

    FrameworkUtils.removeFile(relationshipOutputDir, s3conf, s3);

    for (int i = 0; i < shortDataset.size(); i++) {
        conf.set("dataset-" + datasetId.get(shortDataset.get(i)) + "-agg", datasetAgg.get(shortDataset.get(i)));
    }
    for (int i = 0; i < shortDataset.size(); i++) {
        conf.set("dataset-" + datasetId.get(shortDataset.get(i)) + "-agg-size",
                Integer.toString(datasetAgg.get(shortDataset.get(i)).split(",").length));
    }
    conf.set("dataset-keys", datasetIds);
    conf.set("dataset-names", datasetNames);
    conf.set("first-group", firstGroupStr);
    conf.set("second-group", secondGroupStr);
    conf.set("complete-random", String.valueOf(completeRandomization));
    conf.set("output-ids", String.valueOf(outputIds));
    conf.set("complete-random-str", random);
    conf.set("main-dataset-id", datasetId.get(shortDataset.get(0)));
    conf.set("remove-not-significant", String.valueOf(removeNotSignificant));
    if (noRelationship.length() > 0) {
        conf.set("no-relationship", noRelationship.substring(0, noRelationship.length() - 1));
    }
    if (hasScoreThreshold) {
        conf.set("score-threshold", scoreThreshold);
    }
    if (hasStrengthThreshold) {
        conf.set("strength-threshold", strengthThreshold);
    }

    conf.set("mapreduce.tasktracker.map.tasks.maximum", String.valueOf(machineConf.getMaximumTasks()));
    conf.set("mapreduce.tasktracker.reduce.tasks.maximum", String.valueOf(machineConf.getMaximumTasks()));
    conf.set("mapreduce.jobtracker.maxtasks.perjob", "-1");
    conf.set("mapreduce.reduce.shuffle.parallelcopies", "20");
    conf.set("mapreduce.input.fileinputformat.split.minsize", "0");
    conf.set("mapreduce.task.io.sort.mb", "200");
    conf.set("mapreduce.task.io.sort.factor", "100");
    conf.set("mapreduce.task.timeout", "2400000");

    if (s3) {
        machineConf.setMachineConfiguration(conf);
        conf.set("fs.s3.awsAccessKeyId", awsAccessKeyId);
        conf.set("fs.s3.awsSecretAccessKey", awsSecretAccessKey);
        conf.set("bucket", s3bucket);
    }

    if (snappyCompression) {
        conf.set("mapreduce.map.output.compress", "true");
        conf.set("mapreduce.map.output.compress.codec", "org.apache.hadoop.io.compress.SnappyCodec");
        //conf.set("mapreduce.output.fileoutputformat.compress.codec", "org.apache.hadoop.io.compress.SnappyCodec");
    }
    if (bzip2Compression) {
        conf.set("mapreduce.map.output.compress", "true");
        conf.set("mapreduce.map.output.compress.codec", "org.apache.hadoop.io.compress.BZip2Codec");
        //conf.set("mapreduce.output.fileoutputformat.compress.codec", "org.apache.hadoop.io.compress.BZip2Codec");
    }

    Job job = new Job(conf);
    job.setJobName(jobName);

    job.setMapOutputKeyClass(PairAttributeWritable.class);
    job.setMapOutputValueClass(TopologyTimeSeriesWritable.class);
    job.setOutputKeyClass(Text.class);
    job.setOutputValueClass(Text.class);

    job.setMapperClass(CorrelationMapper.class);
    job.setReducerClass(CorrelationReducer.class);
    job.setNumReduceTasks(machineConf.getNumberReduces());

    job.setInputFormatClass(SequenceFileInputFormat.class);
    //job.setOutputFormatClass(TextOutputFormat.class);
    LazyOutputFormat.setOutputFormatClass(job, TextOutputFormat.class);

    FileInputFormat.setInputDirRecursive(job, true);
    FileInputFormat.setInputPaths(job, indexInputDirs.substring(0, indexInputDirs.length() - 1));
    FileOutputFormat.setOutputPath(job, new Path(relationshipOutputDir));

    job.setJarByClass(Relationship.class);

    long start = System.currentTimeMillis();
    job.submit();
    job.waitForCompletion(true);
    System.out.println(jobName + "\t" + (System.currentTimeMillis() - start));

    // moving files to right place
    for (int i = 0; i < firstGroup.size(); i++) {
        for (int j = 0; j < secondGroup.size(); j++) {

            if (Integer.parseInt(datasetId.get(firstGroup.get(i))) < Integer
                    .parseInt(datasetId.get(secondGroup.get(j)))) {
                dataset1 = firstGroup.get(i);
                dataset2 = secondGroup.get(j);
            } else {
                dataset1 = secondGroup.get(j);
                dataset2 = firstGroup.get(i);
            }

            if (dataset1.equals(dataset2))
                continue;

            String from = s3bucket + relationshipsDir + "/tmp/" + dataset1 + "-" + dataset2 + "/";
            String to = s3bucket + relationshipsDir + "/" + dataset1 + "-" + dataset2 + "/";
            FrameworkUtils.renameFile(from, to, s3conf, s3);
        }
    }
}

From source file:edu.nyu.vida.data_polygamy.resolution.PointsToRegion.java

License:BSD License

public PointsToRegion(int[] xPositions, int[] yPositions, String region, int gridResolution,
        Configuration conf) {//w  w w  .  j  a  v a 2  s . c om

    String bucket = conf.get("bucket", "");

    if (region.equals("nbhd"))
        data = bucket + "neighborhood";
    else if (region.equals("grid"))
        data = bucket + "gneighborhood-" + gridResolution;
    else if (region.equals("zip")) {
        zipcode = true;
        data = bucket + "zipcode";
    } else {
        System.out.println("Invalid region.");
        System.exit(-1);
    }

    this.xPositions = xPositions;
    this.yPositions = yPositions;

    try {
        if (bucket.equals("")) {
            FileSystem fs = FileSystem.get(new Configuration());
            readData(fs.open(new Path(data)));
        } else {
            Path dataPath = new Path(data);
            FileSystem fs = FileSystem.get(dataPath.toUri(), conf);
            readData(fs.open(dataPath));
            fs.close();
        }
    } catch (IOException e) {
        e.printStackTrace();
    }
}