D-Net

import static org.junit.Assert.*;

	@Test

	public void testXpath() throws DocumentException {

		final String value = "CONICYT";

		String ftree = "<fundingtree><funder><id>conicytf____::CONICYT</id><shortname>"+value+"</shortname><name>ComisiónNacionaldeInvestigaciónCientíficayTecnológica</name><jurisdiction>CL</jurisdiction></funder><funding_level_1><id>conicytf____::CONICYT::FONDECYT::REGULAR</id><description>Fondecytstream,REGULAR</description><name>Fondecytstream,REGULAR</name><class>conicyt:fondecytfundings</class><parent><funding_level_0><id>conicytf____::CONICYT::FONDECYT</id><name>FONDECYT</name><description>Fondecytfundings</description><parent/><class>conicyt:fondecytfundings</class></funding_level_0></parent></funding_level_1></fundingtree>";

		final Document doc = new SAXReader().read(new StringReader(ftree));

		assertEquals(value, doc.valueOf("//fundingtree/funder/shortname/text()"));

	}

	public void testParseOpenAPCrecord() throws Exception {

		doTest(loadFromTransformationProfile("oaf2hbase.xml"), load("recordOpenAPC.xml"));

	}

	@Test

<?xml version="1.0" encoding="UTF-8"?>

<oai:record xmlns:datacite="http://datacite.org/schema/kernel-4"

            xmlns:date="http://exslt.org/dates-and-times"

            xmlns:dc="http://purl.org/dc/elements/1.1/"

            xmlns:dr="http://www.driver-repository.eu/namespace/dr"

            xmlns:dri="http://www.driver-repository.eu/namespace/dri"

            xmlns:oaf="http://namespace.openaire.eu/oaf" xmlns:oai="http://www.openarchives.org/OAI/2.0/">

    <oai:header>

        <dri:objIdentifier>openapc_____::000023f9cb6e3a247c764daec4273cbc</dri:objIdentifier>

        <dri:recordIdentifier>10.1155/2015/439379</dri:recordIdentifier>

        <dri:dateOfCollection>2019-12-03T09:34:54.664+01:00</dri:dateOfCollection>

        <oaf:datasourceprefix>openapc_____</oaf:datasourceprefix>

        <dr:dateOfTransformation>2019-12-05T11:21:45.648+01:00</dr:dateOfTransformation>

    </oai:header>

    <metadata xmlns="http://namespace.openaire.eu/">

        <oaf:identifier identifierType="doi">10.1155/2015/439379</oaf:identifier>

        <oaf:identifier identifierType="pmcid">PMC4354964</oaf:identifier>

        <oaf:identifier identifierType="pmid">25811027</oaf:identifier>

        <oaf:processingchargeamount currency="EUR">1721.47</oaf:processingchargeamount>

        <oaf:journal issn="2314-6133">BioMed Research International</oaf:journal>

        <dc:license>http://creativecommons.org/licenses/by/3.0/</dc:license>

        <dr:CobjCategory type="publication">0004</dr:CobjCategory>

        <oaf:accessrights>OPEN</oaf:accessrights>

        <datacite:rights rightsURI="http://purl.org/coar/access_right/c_abf2">open access</datacite:rights>

        <oaf:hostedBy id="openaire____::openapc_initiative" name="OpenAPC Initiative"/>

        <oaf:collectedFrom id="openaire____::openapc_initiative" name="OpenAPC Initiative"/>

    </metadata>

    <oaf:about xmlns:oai="http://wwww.openarchives.org/OAI/2.0/">

        <oaf:datainfo>

            <oaf:inferred>false</oaf:inferred>

            <oaf:deletedbyinference>false</oaf:deletedbyinference>

            <oaf:trust>0.9</oaf:trust>

            <oaf:inferenceprovenance/>

            <oaf:provenanceaction classid="sysimport:crosswalk:datasetarchive"

                                  classname="sysimport:crosswalk:datasetarchive"

                                  schemeid="dnet:provenanceActions" schemename="dnet:provenanceActions"/>

        </oaf:datainfo>

    </oaf:about>

</oai:record>

import java.util.stream.Stream;

    public static final int MAX_RELATIONS = 10000;

        HashSet<String> valid_relation = Stream.of(value.raw())

                .map(kv -> kv.split())

                .filter(s -> sem_rel.equals(new String(s.getFamily())))

                .map(s -> asOaf(s.getValue()))

                .filter(Objects::nonNull)

                .filter(o -> isValid(o))

                .filter(o -> !o.getDataInfo().getDeletedbyinference())

                .map(o -> o.getRel().getTarget())

                .limit(MAX_RELATIONS)

                .collect(Collectors.toCollection(HashSet::new));

import org.apache.hadoop.io.NullWritable;

import org.apache.hadoop.mapreduce.lib.output.LazyOutputFormat;

import java.util.stream.Stream;

		keyOut = new Text("");

			Stream.of(value.raw())

					.filter(kv -> {

						final String q = Bytes.toString(kv.getQualifier());

						boolean skip = q.startsWith("update") || q.equals(DedupUtils.BODY_S);

						if (skip) {

							context.getCounter("export", String.format("skipped %s", StringUtils.substring(q, 0, 6))).increment(1);

						}

						return !skip;

					})

					.filter(kv -> !"".equals(Bytes.toString(kv.getValue())))

					.map(kv -> kv.getValue())

					.forEach(v -> {

						try {

							emit(context, OafProtos.Oaf.parseFrom(v));

						} catch (IOException | InterruptedException e) {

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

							throw new RuntimeException(e);

						}

					});

import java.util.stream.Stream;

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

import eu.dnetlib.data.proto.OafProtos;

import org.apache.commons.lang.StringUtils;

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

			Stream.of(value.raw())

					.filter(kv -> !"".equals(Bytes.toString(kv.getValue())))

					.forEach(kv -> {

						try {

							emit(rowKey, kv.getFamily(), kv.getQualifier(), kv.getValue(), context);

						} catch (IOException | InterruptedException e) {

							throw new RuntimeException(e);

						}

					});

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

        rel.setCollectedfrom(r.getCollectedfromCount() > 0 ?