D-Net

package eu.dnetlib.data.mapreduce.hbase.misc;

import java.io.IOException;

import java.util.Map;

import java.util.Set;

import java.util.stream.Collectors;

import java.util.stream.Stream;

import com.google.common.collect.Sets;

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

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

import eu.dnetlib.data.proto.DatasourceProtos.Datasource;

import eu.dnetlib.data.proto.FieldTypeProtos.KeyValue;

import eu.dnetlib.data.proto.OafProtos.Oaf;

import eu.dnetlib.data.proto.ResultProtos.Result.Instance;

import org.apache.commons.lang.StringUtils;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

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;

/**

 * Reads publications and datasources,

 *  for publications, if publication is related to an EC project emit(collectedFrom@id, 1);

 *  for datasources emit(id, ds)

 *

 * @author claudio

 */

public class PredatoryJournalsMapper extends TableMapper<Text, ImmutableBytesWritable> {

	/**

	 * logger.

	 */

	private static final Log log = LogFactory.getLog(PredatoryJournalsMapper.class); // NOPMD by marko on 11/24/08 5:02 PM

	private static final String PREDATORY_JOURNALS = "Predatory Journal Counters";

	private static final byte[] isProducedBy = "resultProject_outcome_isProducedBy".getBytes();

	private Text keyOut;

	private ImmutableBytesWritable valueOut;

	@Override

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

		super.setup(context);

		keyOut = new Text("");

		valueOut = new ImmutableBytesWritable();

	}

	@Override

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

		try {

			final OafRowKeyDecoder keyDecoder = OafRowKeyDecoder.decode(keyIn.copyBytes());

			final byte[] body = value.getValue(keyDecoder.getType().toString().getBytes(), DedupUtils.BODY_B);

			if (body != null) {

				final Oaf oaf = Oaf.parseFrom(body);

				if (oaf.getDataInfo().getDeletedbyinference() != false) {

					switch (keyDecoder.getType()) {

					case result:

						final Map<byte[], byte[]> projectRels = value.getFamilyMap(isProducedBy);

						if (projectRels != null) {

							long ecProjects = projectRels.keySet().stream()

									.map(String::new)

									.filter(s -> StringUtils.contains(s, "corda"))

									.count();

							final Set<String> ids = Sets.newHashSet();

							if (ecProjects > 0) {

								ids.addAll(

										Stream.concat(

												oaf.getEntity().getCollectedfromList().stream()

													.map(KeyValue::getKey),

												oaf.getEntity().getResult().getInstanceList().stream()

													.map(Instance::getHostedby)

													.map(KeyValue::getKey)

										).collect(Collectors.toList()));

								ids.forEach(dsId -> {

									emit(context, dsId, Bytes.toBytes(1));

									context.getCounter(PREDATORY_JOURNALS, "publications").increment(1);

								});

							}

						}

						break;

					case datasource:

						final Datasource ds = oaf.getEntity().getDatasource();

						final String dsType = ds.getMetadata().getDatasourcetype().getClassid();

						if (StringUtils.contains(dsType, "journal")) {

							emit(context, oaf.getEntity().getId(), body);

							context.getCounter(PREDATORY_JOURNALS, "journals").increment(1);

						}

						break;

					default:

						break;

					}

				}

			}

		} catch (final Throwable e) {

			log.error("error exporting the following record from HBase: " + value.toString(), e);

			context.getCounter("error", e.getClass().getName()).increment(1);

			throw new RuntimeException(e);

		}

	}

	private void emit(final Context context, final String key, byte[] value) {

		keyOut.set(key);

		valueOut.set(value);

		try {

			context.write(keyOut, valueOut);

		} catch (IOException | InterruptedException e) {

			throw new RuntimeException(e);

		}

	}

}