D-Net

package eu.dnetlib.data.mapreduce.hbase.propagation.communitythroughorganization;

import com.google.gson.Gson;

import eu.dnetlib.data.mapreduce.hbase.propagation.Value;

import eu.dnetlib.data.mapreduce.util.OafRowKeyDecoder;

import eu.dnetlib.data.proto.OafProtos;

import eu.dnetlib.data.proto.TypeProtos;

import org.apache.commons.lang3.StringUtils;

import org.apache.hadoop.hbase.client.Result;

import org.apache.hadoop.hbase.io.ImmutableBytesWritable;

import org.apache.hadoop.hbase.mapreduce.TableMapper;

import org.apache.hadoop.hbase.util.Bytes;

import org.apache.hadoop.io.Text;

import java.io.IOException;

import java.util.Set;

import static eu.dnetlib.data.mapreduce.hbase.propagation.PropagationConstants.*;

import static eu.dnetlib.data.mapreduce.hbase.propagation.Utils.getEntity;

import static eu.dnetlib.data.mapreduce.hbase.propagation.Utils.getRelationTarget;

public class PropagationCommunityThroughOrganizationMapper extends TableMapper<ImmutableBytesWritable, Text> {

    private Text valueOut;

    private ImmutableBytesWritable keyOut;

    private OrganizationMap organizationMap;

    //seleziono il tipo della entry:

    //Result:

    //se non e' deleted by inference ed ha organizzazioni a cui e' associato,

    // //emetto id della relazione ed id del risultato per ogni organizzazione a cui e' associato

    //ORGANIZATION:

    //se non e' deleted by inference e non e' deduplicata emetto l'id della organizzazione

    //se non e' deleted by inference ed e' deduplicata, emetto id della organizzazione ed id del deduplicato e lista delle organizzazioni mergiate

    @Override

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

        super.setup(context);

        valueOut = new Text();

        keyOut = new ImmutableBytesWritable();

        organizationMap = new Gson().fromJson(context.getConfiguration().get("organization.community.map"), OrganizationMap.class);

        System.out.println("got organization map: " + new Gson().toJson(organizationMap)) ;

    }

    @Override

    protected void map(final ImmutableBytesWritable keyIn, final Result value, final Context context) throws IOException, InterruptedException {

        final TypeProtos.Type type = OafRowKeyDecoder.decode(keyIn.copyBytes()).getType();

        final OafProtos.OafEntity entity = getEntity(value, type);//getEntity already verified that it is not delByInference

        if (entity != null) {

            switch (type) {

                case organization:

                    DedupedList communityList = getCommunityList(Bytes.toString(keyIn.get()),

                            getRelationTarget(value, DEDUP_RELATION_ORGANIZATION + REL_DEDUP_REPRESENTATIVE_RESULT, context, COUNTER_PROPAGATION), context);

                    if (communityList.size() > 0){

                        valueOut.set(Value.newInstance(

                                new Gson().toJson(

                                        communityList, //search for organization it merges

                                        DedupedList.class),

                                ORGANIZATION_COMMUNITY_TRUST,

                                Type.fromorganization).toJson());

                        context.write(keyIn, valueOut);

                        context.getCounter(COUNTER_PROPAGATION, "emit for organization ").increment(1);

                    }else{

                        context.getCounter(COUNTER_PROPAGATION, "community list size = 0 ").increment(1);

                    }

                    break;

                case result:

                    Set<String> result_organization = getRelationTarget(value, RELATION_ORGANIZATION + REL_RESULT_ORGANIZATION, context, COUNTER_PROPAGATION);

                    for(String org: result_organization)

                        emit(org, Bytes.toString(keyIn.get()), context);

                    break;

            }

        }

    }

    private DedupedList getCommunityList(String organizationId, Set<String> relationTarget, Context context) {

        DedupedList communityList = new DedupedList();

        relationTarget.stream().forEach(org -> communityList.addAll(organizationMap.get(StringUtils.substringAfter(org, "|"))));

        communityList.addAll(organizationMap.get(StringUtils.substringAfter(organizationId,"|")));

        communityList.stream().forEach(c->context.getCounter(COUNTER_PROPAGATION,"found organization for " + c).increment(1));

        return communityList;

    }

    private void emit(String org, String resId, Context context) throws IOException, InterruptedException {

        keyOut.set(Bytes.toBytes(org));

        valueOut.set(Value.newInstance(resId, ORGANIZATION_COMMUNITY_TRUST, Type.fromresult).toJson());

        context.write(keyOut,valueOut);

        context.getCounter(COUNTER_PROPAGATION, "emit for result").increment(1);

    }

}