source: other-projects/hathitrust/solr-extracted-features/trunk/src/main/java/org/hathitrust/PrepareForIngest.java@ 30951

package org.hathitrust;

import java.io.Serializable;
import org.apache.commons.cli.*;

import org.apache.spark.api.java.*;
import org.apache.spark.SparkConf;

public class PrepareForIngest implements Serializable
{
    private static final long serialVersionUID = 1L;

    public static final int NUM_PARTITIONS = 6; // default would appear to be 2
    
    protected String _input_dir;
    protected String _json_list_filename;
    protected String _output_dir;
    
    protected int    _verbosity;

    public PrepareForIngest(String input_dir, String json_list_filename, String output_dir, int verbosity)
    {
        _input_dir = input_dir;
        _json_list_filename = (json_list_filename != null) ? json_list_filename : input_dir;
        _output_dir = output_dir;
        _verbosity = verbosity;
    }

    public void exec()
    {   
        String spark_app_name = "HathiTrust Extract Features: Prepare for Solr Ingest";
        spark_app_name += " [" + _json_list_filename + "]";

        SparkConf conf = new SparkConf().setAppName(spark_app_name);
        JavaSparkContext jsc = new JavaSparkContext(conf);
        
        if (_verbosity >= 2) {
            System.out.println("Default Minimum Partions: " + jsc.defaultMinPartitions());
            System.out.println("Default Parallelism: " + jsc.defaultParallelism());
        }
                
        JavaRDD<String> json_list_data = jsc.textFile(_json_list_filename,NUM_PARTITIONS).cache();

        PagedJSON paged_json = new PagedJSON(_input_dir,_output_dir,_verbosity);
        JavaRDD<String> json_ids = json_list_data.flatMap(paged_json).cache();

        long num_ids = json_ids.count();
        System.out.println("");
        System.out.println("############");
        System.out.println("# Number of page ids: " + num_ids);
        System.out.println("############");
        System.out.println("");

        System.out.println("############");
        String rdd_save_file = "rdd-solr-json-page-files";
        json_ids.saveAsTextFile(rdd_save_file);
        System.out.println("# Saved RDD of Solr JSON page files, top-level, as:");
        System.out.println("#  " + rdd_save_file);
        System.out.println("############");
        System.out.println("");
        
        jsc.close();
    }

    public static void main(String[] args) {


        Options options = new Options();

        //.withType(Integer.class)
/*
        options.addOption(OptionBuilder.withLongOpt("verbosity")
                .withDescription("Set to control the level of debugging output [0=none, 1=some, 2=lots]")
                .hasArg()
                .withArgName("v")
                .isRequired(false)
                .create());
*/
        //Option num_cores_opt = new Option("n", "num-cores", true, "Number of cores to use");
        //num_cores_opt.setRequired(false);
        //options.addOption(num_cores_opt);

        Option verbosity_opt = new Option("v", "verbosity", true, 
                "Set to control the level of debugging output [0=none, 1=some, 2=lots]");
        verbosity_opt.setRequired(false);
        options.addOption(verbosity_opt);
        
        //CommandLineParser parser = new DefaultParser(); // 1.3 and above
        
        // need to work with CLI v1.2 as this is the JAR that is bundled with Hadoop/Spark 
        CommandLineParser parser = new GnuParser(); 
        HelpFormatter formatter = new HelpFormatter();
        CommandLine cmd;

        try {
            cmd = parser.parse(options, args);
        }
        catch (ParseException e) {
            System.err.println(e.getMessage());
            formatter.printHelp("RUN.bash [options] json-file-list.txt input-dir output-dir", options);
            System.exit(1);
            return;
        }

        //value = ((Integer)cmdLine.getParsedOptionValue("num-cores")).intValue();
        //value = ((Integer)cmdLine.getOptionValue("num-cores","2")).intValue();

        //cmd.hasOption("json-filelist")
        
        String verbosity_str = cmd.getOptionValue("verbosity","0");
        int verbosity = Integer.parseInt(verbosity_str);

        String[] filtered_args = cmd.getArgs();

        if (filtered_args.length != 3) {
            formatter.printHelp("RUN.bash [options] json-filelist.txt input-dir output-dir", options);
            System.exit(1);
        }
        String json_list_filename = filtered_args[0];
        String input_dir  = filtered_args[1];
        String output_dir = filtered_args[2];

        PrepareForIngest prep_for_ingest = new PrepareForIngest(input_dir,json_list_filename,output_dir,verbosity);
        prep_for_ingest.exec();

    }
}