D-Net

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

import com.google.common.collect.ArrayListMultimap;

import com.google.common.collect.Multimap;

import com.lambdaworks.redis.RedisClusterConnection;

import com.lambdaworks.redis.RedisURI;

import com.lambdaworks.redis.cluster.ClusterClientOptions;

import com.lambdaworks.redis.cluster.RedisAdvancedClusterConnection;

import com.lambdaworks.redis.cluster.RedisClusterClient;

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

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

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.ArrayList;

import java.util.List;

import java.util.Map;

import java.util.concurrent.TimeUnit;

/**

 * 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 SourceBuildMapper extends Mapper<LongWritable, Text, Text, Text> {

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

    //address of your redis server

    private String redisHost;

    private Integer redisPort;

    private LodConfiguration lodConfiguration;

    private String lastExecutionDate;

    //   private StatefulRedisClusterConnection connection;

    private RedisClusterClient clusterClient;

    // private  RedisAdvancedClusterCommands<String, String> syncCommands;

    RedisClusterConnection<String, String> connection;

    public static enum ENTITIES_COUNTER {

        BLOCKING_KEYS,

        REDIS_RECORDS

    }

    @Override

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

        lodConfiguration = new LodConfiguration();

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

        redisHost = context.getConfiguration().get("lod.redisHost");

        redisPort = Integer.parseInt(context.getConfiguration().get("lod.redisPort"));

        // initLettuceRedis();

    }

    private void initLettuceRedis() {

        RedisURI redisUri = RedisURI.Builder.redis(redisHost).withPort(redisPort).build();

        clusterClient = RedisClusterClient.create(redisUri);

        clusterClient.setOptions(new ClusterClientOptions.Builder()

                .refreshClusterView(true)

                .refreshPeriod(1, TimeUnit.MINUTES)

                .build());

        RedisClusterConnection<String, String> connection = clusterClient.connectCluster();

        System.out.println("Connected to Redis");

    /*    connection = clusterClient.connect();

        syncCommands = connection.sync();*/

    }

    @Override

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

        try {

            //get ID - output source_recordID so we can group by id and get all props of a record

            StringBuffer value = new StringBuffer();

            Multimap<String, String> Fields = ArrayListMultimap.create();

            String[] pairs = result.toString().split(",");

            for (int i = 0; i < pairs.length; i += 2) {

                if (i + 1 < pairs.length) Fields.put(pairs[i], pairs[i + 1]);

            }

            //get ID

            String id = "source_" + (String) ((List) Fields.get("id")).get(0);

            for (Map.Entry<String, String> entry : Fields.entries()) {

                //   if (lodConfiguration.isValidField(entityType, entry.getKey().replace("/", "\\/")))

                {

                    value.append(entry.getKey() + "\t" + entry.getValue() + ",");

                }

            }

            //Write to cache

            // writeToRedis(id, value.toString());

            context.getCounter(ENTITIES_COUNTER.REDIS_RECORDS).increment(1);

            ArrayList<String> blockingKeys = Blocking.tokenBlocking(Fields);

            for (String blockingKey : blockingKeys) {

                //Write BlockingKey, RecordID to output

                context.write(new Text(blockingKey), new Text(id));

                context.getCounter(ENTITIES_COUNTER.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);

        }

    }

    private void writeToRedis(String key, String value) throws Exception {

        try {

            //syncCommands.set(key, value);

            connection.set(key, value)

            ;            //jedisCluster.set(key, value);

            log.info("Writing out to redis " + key + " , " + value);

        } catch (Exception e) {

            log.error("Error writing entity to Redis", e);

            throw new RuntimeException("Error writing to Redis  ", e);

        }

    }

    @Override

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

      //  connection.close();

        //clusterClient.shutdown();

        super.cleanup(context);

    }

    public Logger getLog() {

        return log;

    }

    public void setLog(Logger log) {

        this.log = log;

    }

    public String getRedisHost() {

        return redisHost;

    }

    public void setRedisHost(String redisHost) {

        this.redisHost = redisHost;

    }

    public Integer getRedisPort() {

        return redisPort;

    }

    public void setRedisPort(Integer redisPort) {

        this.redisPort = redisPort;

    }

    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;

    }

}