Add command line parser

Author: yatharthranjan

build.gradle

@@ -2,7 +2,7 @@ apply plugin: 'java'
 apply plugin: 'application'
 
 group 'org.radarcns.restructurehdfs'
-version '0.3.2'
+version '0.3.3-SNAPSHOT'
 mainClassName = 'org.radarcns.RestructureAvroRecords'
 
 run {
@@ -15,6 +15,7 @@ targetCompatibility = '1.8'
 ext.avroVersion = '1.8.2'
 ext.jacksonVersion = '2.8.9'
 ext.hadoopVersion = '2.7.3'
+ext.jCommanderVersion = '1.72'
 
 repositories {
     jcenter()
@@ -27,6 +28,8 @@ dependencies {
     compile group: 'com.fasterxml.jackson.core' , name: 'jackson-databind', version: jacksonVersion
     compile group: 'com.fasterxml.jackson.dataformat' , name: 'jackson-dataformat-csv', version: jacksonVersion
 
+    compile group: 'com.beust', name: 'jcommander', version: jCommanderVersion
+
     runtime group: 'org.apache.hadoop', name: 'hadoop-hdfs', version: hadoopVersion
 
     testCompile group: 'junit', name: 'junit', version: '4.12'

src/main/java/org/radarcns/RestructureAvroRecords.java

@@ -16,6 +16,7 @@
 
 package org.radarcns;
 
+import com.beust.jcommander.JCommander;
 import com.fasterxml.jackson.databind.JsonMappingException;
 import org.apache.avro.Schema.Field;
 import org.apache.avro.file.DataFileReader;
@@ -32,6 +33,7 @@
 import org.radarcns.util.JsonAvroConverter;
 import org.radarcns.util.ProgressBar;
 import org.radarcns.util.RecordConverterFactory;
+import org.radarcns.util.commandline.CommandLineArgs;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
@@ -70,27 +72,30 @@ public class RestructureAvroRecords {
 
     private long processedFileCount;
     private long processedRecordsCount;
-    private static final boolean USE_GZIP = "gzip".equalsIgnoreCase(System.getProperty("org.radarcns.compression"));
+    private static boolean USE_GZIP;
+    private static boolean DO_DEDUPLICATE;
 
-    // Default set to false because causes loss of records from Biovotion data. https://github.com/RADAR-base/Restructure-HDFS-topic/issues/16
-    private static final boolean DO_DEDUPLICATE = "true".equalsIgnoreCase(System.getProperty("org.radarcns.deduplicate", "false"));
+    private static final CommandLineArgs commandLineArgs = new CommandLineArgs();
+    private static final JCommander parser = JCommander.newBuilder().addObject(commandLineArgs).build();
 
     public static void main(String [] args) throws Exception {
-        if (args.length != 3) {
-            System.out.println("Usage: hadoop jar restructurehdfs-all-0.2.jar <webhdfs_url> <hdfs_root_directory> <output_folder>");
-            System.exit(1);
-        }
+
+        parser.setProgramName("hadoop jar restructurehdfs-all-0.3.2.jar");
+        parser.parse(args);
+        USE_GZIP = "gzip".equalsIgnoreCase(commandLineArgs.compression);
+        DO_DEDUPLICATE = commandLineArgs.deduplicate;
 
         logger.info(new SimpleDateFormat("yyyy-MM-dd HH:mm:ss").format(new Date()));
         logger.info("Starting...");
-        logger.info("In:  " + args[0] + args[1]);
-        logger.info("Out: " + args[2]);
+        logger.info("In:  " + commandLineArgs.hdfsUri + commandLineArgs.hdfsRootDirectory);
+        logger.info("Out: " + commandLineArgs.outputDirectory);
+        logger.info("Deduplicate set to {}", DO_DEDUPLICATE);
 
         long time1 = System.currentTimeMillis();
 
-        RestructureAvroRecords restr = new RestructureAvroRecords(args[0], args[2]);
+        RestructureAvroRecords restr = new RestructureAvroRecords(commandLineArgs.hdfsUri, commandLineArgs.outputDirectory);
         try {
-            restr.start(args[1]);
+            restr.start(commandLineArgs.hdfsRootDirectory);
         } catch (IOException ex) {
             logger.error("Processing failed", ex);
         }

src/main/java/org/radarcns/util/commandline/BooleanValidator.java

@@ -0,0 +1,16 @@
+package org.radarcns.util.commandline;
+
+import com.beust.jcommander.IParameterValidator;
+import com.beust.jcommander.ParameterException;
+
+public class BooleanValidator implements IParameterValidator{
+
+    @Override
+    public void validate(String name, String value) throws ParameterException {
+        if(!value.equalsIgnoreCase("true") && !value.equalsIgnoreCase("false")) {
+            ParameterException exc = new ParameterException("Parameter " + name + " can only be true or false (found " + value +")");
+            exc.usage();
+            throw exc;
+        }
+    }
+}

src/main/java/org/radarcns/util/commandline/CommandLineArgs.java

@@ -0,0 +1,24 @@
+package org.radarcns.util.commandline;
+
+import com.beust.jcommander.Parameter;
+
+public class CommandLineArgs {
+    @Parameter(names = { "-c", "--compression" }, description = "Compression to use when converting the files. Default is none. Gzip is available.")
+    public String compression = "none";
+
+    // Default set to false because causes loss of records from Biovotion data. https://github.com/RADAR-base/Restructure-HDFS-topic/issues/16
+    @Parameter(names = { "-d", "--deduplicate" }, description = "Boolean to define if to use deduplication or not.", validateWith = BooleanValidator.class)
+    public Boolean deduplicate = false;
+
+    @Parameter(names = { "-u", "--hdfs-uri" }, description = "The HDFS uri to connect to. Eg - 'hdfs://<HOST>:<RPC_PORT>/<PATH>'.", required = true, validateWith = { HdfsUriValidator.class, PathValidator.class })
+    public String hdfsUri;
+
+    @Parameter(names = { "-i", "--hdfs-root-directory" }, description = "The input HDFS root directory from which files are to be read. Eg - '/topicAndroidNew'", required = true, validateWith = PathValidator.class)
+    public String hdfsRootDirectory;
+
+    @Parameter(names = { "-o", "--output-directory"}, description = "The output folder where the files are to be extracted.", required = true, validateWith = PathValidator.class)
+    public String outputDirectory;
+
+    @Parameter(names = { "-h", "--help"}, help = true)
+    public boolean help;
+}

src/main/java/org/radarcns/util/commandline/HdfsUriValidator.java

@@ -0,0 +1,16 @@
+package org.radarcns.util.commandline;
+
+
+import com.beust.jcommander.ParameterException;
+import com.beust.jcommander.IParameterValidator;
+
+public class HdfsUriValidator implements IParameterValidator{
+    @Override
+    public void validate(String name, String value) throws ParameterException {
+        if (! value.matches("((hdfs)|(webhdfs)):(/?/?)[^\\s]+")) {
+            ParameterException exc = new ParameterException("Parameter " + name + " should be a valid HDFS or WebHDFS URI. Eg - hdfs://<HOST>:<RPC_PORT>/<PATH>. (found " + value +")");
+            exc.usage();
+            throw exc;
+        }
+    }
+}

src/main/java/org/radarcns/util/commandline/PathValidator.java

@@ -0,0 +1,15 @@
+package org.radarcns.util.commandline;
+
+import com.beust.jcommander.ParameterException;
+import com.beust.jcommander.IParameterValidator;
+
+public class PathValidator implements IParameterValidator{
+    @Override
+    public void validate(String name, String value) throws ParameterException {
+        if (value == null || value.isEmpty()) {
+            ParameterException exc =  new ParameterException("Parameter " + name + " should be supplied. It cannot be empty or null. (found " + value +")");
+            exc.usage();
+            throw exc;
+        }
+    }
+}