D-Net

package eu.dnetlib.data.mapreduce.hbase.lodExport.build;

import eu.dnetlib.data.mapreduce.hbase.lodExport.utils.configuration.LodConfiguration;

import eu.dnetlib.data.mapreduce.hbase.lodExport.utils.caching.RedisUtils;

import eu.dnetlib.data.mapreduce.hbase.lodExport.utils.blocking.Blocking;

import org.apache.hadoop.io.LongWritable;

import org.apache.hadoop.io.Text;

import org.apache.hadoop.mapreduce.Mapper;

import org.apache.log4j.Logger;

import java.io.IOException;

import java.util.HashMap;

import java.util.List;

import java.util.Map;

/**

 * Mapper Class that reads HBASE contents and prepares them for the StatsDB

 * export

 */

/*

-> Parse LOD dump files

        Process lod input files and divide by entity type (both source and target)

        Transform to id, array of [ properties] format

        Store to HDFS

        For -> Multiple outputs and inputs

        Multiple inputs: all source and target datasets and their corresponding mappings

        M/O: separate output files for each dataset: mark records so that they are written to the correct one

*/

public class SourceBuildNoCacheMapper extends Mapper<LongWritable, Text, Text, Text> {

    private Logger log = Logger.getLogger(this.getClass());

    private LodConfiguration lodConfiguration;

    private String lastExecutionDate;

    private String stopWords;

    private Map<String, Integer> stopWordsMap = new HashMap();

    private static final String LINE_DELIM = "\t.\t";

    private static final String FIELD_DELIM = "\t";

    public static enum SOURCE_BUILD_COUNTERS {

        BLOCKING_KEYS,

        REDIS_RECORDS

    }

    @Override

    protected void setup(Context context) throws IOException, InterruptedException {

        try {

            lodConfiguration = new LodConfiguration();

            lodConfiguration.load(context.getConfiguration().get("lod.sourceMappings"));

            stopWords = context.getConfiguration().get("lod.stopwords");

            for (String stopword : stopWords.split(",")) {

                stopWordsMap.put(stopword, 0);

            }

        } catch (Exception ex) {

            log.error("An error occured during Mapper Setup " + ex.toString(), ex);

            System.out.println(ex.getCause().toString());

        }

    }

    @Override

    protected void map(final LongWritable keyIn, final Text result, final Context context) throws IOException {

        try {

            StringBuilder id = new StringBuilder();

            String[] triples = result.toString().split(LINE_DELIM);

            for (String triple : triples) {

                String[] fields = triple.split(FIELD_DELIM);

                if (fields.length == 3) {

                    if (id.length() < 1) {

                        id.append("source_").append(fields[0]);

                    }

                    String property = fields[1];

                    String value = fields[2];

                    List<String> blockingKeys = Blocking.tokenBlocking(value, stopWordsMap);

                    for (String blockingKey : blockingKeys) {

                        //Write BlockingKey, RecordID to output

                        context.write(new Text(blockingKey), new Text(id + FIELD_DELIM + result.toString()));

                        context.getCounter(SOURCE_BUILD_COUNTERS.BLOCKING_KEYS).increment(1);

                    }

                }

            }

        } catch (Exception e) {

            log.error("Error writing entity to M/R output", e);

            log.error("result error    " + result.toString());

            throw new RuntimeException(e);

        }

    }

    @Override

    protected void cleanup(Context context) throws IOException, InterruptedException {

        super.cleanup(context);

        log.info("Cleaning up mapper...");

    }

    public LodConfiguration getLodConfiguration() {

        return lodConfiguration;

    }

    public void setLodConfiguration(LodConfiguration lodConfiguration) {

        this.lodConfiguration = lodConfiguration;

    }

    public String getLastExecutionDate() {

        return lastExecutionDate;

    }

    public void setLastExecutionDate(String lastExecutionDate) {

        this.lastExecutionDate = lastExecutionDate;

    }

}