D-Net

{"type_source": "SVN", "goal": "package -U source:jar", 

"url": "http://svn-public.driver.research-infrastructures.eu/driver/dnet45/modules/dnet-openaireplus-mapping-utils/trunk/", "deploy_repository": "dnet45-snapshots", "version": "4", "mail": "sandro.labruzzo@isti.cnr.it,michele.artini@isti.cnr.it, claudio.atzori@isti.cnr.it, alessia.bardi@isti.cnr.it", "deploy_repository_url": "http://maven.research-infrastructures.eu/nexus/content/repositories/dnet45-snapshots", "name": "dnet-openaireplus-mapping-utils"}

package eu.dnetlib.openaire.exporter.model;

import org.junit.Before;

import org.junit.Test;

import static org.junit.Assert.assertEquals;

public class ProjectTest {

	Project pMZOS;

	Project pFP7;

	@Before

	public void setUp() throws Exception {

		pMZOS = new Project()

				.setFunder("MZOS")

				.setJurisdiction("HR")

				.setFundingpathid("irb_hr______::MZOS")

				.setAcronym("")

				.setTitle("Project Title")

				.setCode("115-1152437-2500")

				.setStartdate("2007-01-01")

				.setEnddate("2009-01-01");

		pFP7 = new Project()

				.setFunder("EC")

				.setJurisdiction("EU")

				.setFundingpathid("ec__________::EC::FP7::SP1::NMP")

				.setAcronym("REFFIBRE")

				.setTitle("Project Title")

				.setCode("604187")

				.setStartdate("2013-11-01")

				.setEnddate("20015-01-01");

	}

	@Test

	public void testIdNamespaceMZOS(){

		String ns = pMZOS.getIdnamespace();

		assertEquals("info:eu-repo/grantAgreement/MZOS//115-1152437-2500/HR", ns);

	}

	@Test

	public void testIdNamespaceFP7(){

		String ns = pFP7.getIdnamespace();

		assertEquals("info:eu-repo/grantAgreement/EC/FP7/604187/EU", ns);

	}

}

package eu.dnetlib.openaire.exporter.model;

import java.io.IOException;

import java.io.StringReader;

import com.google.common.collect.Lists;

import org.apache.commons.io.IOUtils;

import org.springframework.core.io.ClassPathResource;

import org.springframework.core.io.Resource;

import static org.junit.Assert.assertEquals;

/**

 * Created by claudio on 22/09/16.

 */

public class ProjectDetailTest {

	private Resource projectsCsv = new ClassPathResource("/eu/dnetlib/openaire/exporter/model/projectDetails.csv");

	private Resource projectsJson = new ClassPathResource("/eu/dnetlib/openaire/exporter/model/projectDetails.json");

	//@Test

	public void testSerialisationCSV() throws IOException {

		final String csv = IOUtils.toString(projectsCsv.getInputStream());

		doTest(csv, "csv");

	}

	//@Test

	public void testSerialisationCSV2() throws IOException {

		final ProjectDetail p = new ProjectDetail()

				.setAcronym("acro")

				.setCode("01")

				.setOptional1("op1")

				.setOptional2("op2")

				.setProjectId("project_01")

				.setJsonextrainfo("extraInfo")

				.setFundingPath(Lists.newArrayList("fundingpath1", "fundingpath2"));

		doTest(p.asCSV(), "csv");

	}

	//@Test

	public void testSerialisationJSON() throws IOException {

		final String json = IOUtils.toString(projectsJson.getInputStream());

		doTest(json, "json");

	}

	private void doTest(final String data, final String format) throws IOException {

		final StringReader reader = new StringReader(data);

		for(String line : IOUtils.readLines(reader)) {

			System.out.println("line:          " + line);

			ProjectDetail p;

			String s = "";

			switch (format) {

			case "csv":

				p = ProjectDetail.fromCSV(line);

				s = p.asCSV();

				break;

			case "json":

				p = ProjectDetail.fromJson(line);

				s = p.asJson();

				break;

			default: throw new IllegalArgumentException("invalid format: " + format);

			}

			System.out.println("serialisation: " + s);

			assertEquals(s, line);

		}

	}

}

package eu.dnetlib.data.mapreduce.util;

import static org.junit.Assert.assertFalse;

import static org.junit.Assert.assertNotNull;

import java.util.List;

import org.junit.Test;

import eu.dnetlib.data.proto.KindProtos.Kind;

import eu.dnetlib.miscutils.functional.xml.IndentXmlString;

public class OafDecoderTest {

	@Test

	public void testAsXml() {

		final OafDecoder decoder = OafTest.embed(OafTest.getResult("50|id_1"), Kind.entity);

		assertNotNull(decoder);

		assertNotNull(decoder.asXml());

		System.out.println(IndentXmlString.apply(decoder.asXml()));

	}

}

package eu.dnetlib.data.mapreduce.util;

import org.junit.Test;

public class OafRowKeyDecoderTest {

	@Test

	public void test() {

		String id1 = "50|acnbad______::0a454baf9c61e63d42fb83ab549f8062";

		OafRowKeyDecoder d = OafRowKeyDecoder.decode(id1);

		System.out.println(d.getId());

	}

}

package eu.dnetlib.data.mapreduce.util;

import eu.dnetlib.data.proto.DedupProtos.Dedup.RelName;

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

import eu.dnetlib.data.proto.RelTypeProtos.RelType;

import eu.dnetlib.data.proto.RelTypeProtos.SubRelType;

import org.junit.Before;

import org.junit.Test;

import static org.junit.Assert.assertEquals;

import static org.junit.Assert.assertNotNull;

public class OafRelDecoderTest {

	private OafRel oafRel;

	@Before

	public void setUp() {

		oafRel = OafTest.getDedupRel("ID_1", "ID_2", RelType.resultResult, "isMergedIn");

	}

	@Test

	public void testSetClass() {

		OafRelDecoder d1 = OafRelDecoder.decode(oafRel);

		assertNotNull(d1);

		assertEquals("isMergedIn", d1.getRelClass());

		OafRelDecoder d2 = OafRelDecoder.decode(d1.setClassId("isMergedIn").build());

		assertEquals("isMergedIn", d2.getRelClass());

		assertEquals("isMergedIn", d2.getRelMetadata().getSemantics().getClassid());

		assertEquals("isMergedIn", d2.getRelMetadata().getSemantics().getClassname());

	}

	@Test

	public void testGetCF() {

		assertEquals("resultResult_dedup_isMergedIn", OafRelDecoder.getCFQ(RelType.resultResult, SubRelType.dedup, RelName.isMergedIn));

		assertEquals("resultResult_dedup_isMergedIn", OafRelDecoder.getCFQ(RelType.resultResult, SubRelType.dedup, "isMergedIn"));

	}

}

package eu.dnetlib.data.bulktag;

import org.apache.commons.io.IOUtils;

import org.apache.commons.lang3.StringUtils;

import org.dom4j.DocumentException;

import org.junit.Before;

import org.junit.Test;

import java.io.IOException;

import static org.junit.Assert.assertEquals;

import static org.junit.Assert.assertTrue;

/**

 * Created by miriam on 03/08/2018.

 */

public class CommunityConfigurationFactoryTest {

    private String xml;

    @Before

    public void setUp() throws IOException, DocumentException {

        xml = IOUtils.toString(getClass().getResourceAsStream("community_configuration.xml"));

    }

    @Test

    public void parseTest() throws DocumentException {

        final CommunityConfiguration cc = CommunityConfigurationFactory.newInstance(xml);

        assertEquals(5,cc.size());

        cc.getCommunityList().forEach(c -> assertTrue(StringUtils.isNoneBlank(c.getId())));

    }

}

package eu.dnetlib.data.transform;

import java.io.IOException;

import java.io.InputStream;

import java.io.StringReader;

import java.util.List;

import java.util.Objects;

import java.util.Properties;

import java.util.concurrent.atomic.AtomicInteger;

import java.util.function.Function;

import java.util.stream.Collectors;

import java.util.stream.StreamSupport;

import com.google.protobuf.InvalidProtocolBufferException;

import com.mongodb.client.MongoCollection;

import com.mongodb.client.MongoDatabase;

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

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

import eu.dnetlib.data.transform.xml2.DatasetToProto;

import eu.dnetlib.miscutils.collections.Pair;

import eu.dnetlib.miscutils.datetime.HumanTime;

import org.apache.commons.io.IOUtils;

import org.apache.commons.lang3.time.StopWatch;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

import org.apache.commons.math.stat.descriptive.DescriptiveStatistics;

import org.bson.Document;

import org.dom4j.DocumentException;

import org.dom4j.io.SAXReader;

import org.junit.Before;

import org.junit.Test;

import org.junit.runner.RunWith;

import org.springframework.beans.factory.annotation.Autowired;

import org.springframework.test.context.ContextConfiguration;

import org.springframework.test.context.junit4.SpringJUnit4ClassRunner;

import static org.junit.Assert.assertFalse;

import static org.junit.Assert.assertNotNull;

import static org.junit.Assert.assertTrue;

@RunWith(SpringJUnit4ClassRunner.class)

@ContextConfiguration(classes = { ConfigurationTestConfig.class })

public class ParserToProtoIT {

	private static final Log log = LogFactory.getLog(ParserToProtoIT.class);

	private static final String DATACITE = "datacite";

	private static final String NARCIS = "narcis";

	private static final int BATCH_SIZE = 10000;

	private static final int LOG_FREQ = 5000;

	private static final int LIMIT = 10000;

	public static final String BODY = "body";

	private static String basePathProfiles = "/eu/dnetlib/test/profiles/TransformationRuleDSResources/TransformationRuleDSResourceType/";

	private int batchSize = BATCH_SIZE;

	private int logFreq = LOG_FREQ;

	private int limit = LIMIT;

	@Autowired

	private MongoDatabase db;

	@Autowired()

	private Properties testProperties;

	@Before

	public void setUp() {

		logFreq = Integer.valueOf(testProperties.getProperty("test.logFreq", String.valueOf(LOG_FREQ)));

		batchSize = Integer.valueOf(testProperties.getProperty("test.batchSize", String.valueOf(BATCH_SIZE)));

		limit = Integer.valueOf(testProperties.getProperty("test.limit", String.valueOf(limit)));

	}

	@Test

	public void testParseDataciteWithVTD() throws IOException {

		doTest(s -> new Pair<>(s, new DatasetToProto().apply(s)), DATACITE);

	}

	@Test

	public void testParseDataciteWithXSLT() throws IOException {

		final String xslt = IOUtils.toString(loadFromTransformationProfile("odf2hbase.xml"));

		final XsltRowTransformer transformer = XsltRowTransformerFactory.newInstance(xslt);

		doTest(rowToOaf(transformer), DATACITE);

	}

	@Test

	public void testParseNarcisWithXSLT() throws IOException {

		final String xslt = IOUtils.toString(loadFromTransformationProfile("oaf2hbase.xml"));

		final XsltRowTransformer transformer = XsltRowTransformerFactory.newInstance(xslt);

		doTest(rowToOaf(transformer), NARCIS);

	}

	//// HELPERS

	private void doTest(final Function<String, Pair<String, Oaf>> mapper, final String collectionName) {

		final MongoCollection<Document> collection = db.getCollection(collectionName);

		final long collectionSize = collection.count();

		log.info(String.format("found %s records in collection '%s'", collectionSize, collectionName));

		final AtomicInteger read = new AtomicInteger(0);

		final DescriptiveStatistics stats = new DescriptiveStatistics();

		final StopWatch recordTimer = new StopWatch();

		final StopWatch totalTimer = StopWatch.createStarted();

		StreamSupport.stream(collection.find().batchSize(batchSize).spliterator(), false)

				.limit(limit)

				.peek(d -> {

					if (read.addAndGet(1) % logFreq == 0) {

						log.info(String.format("records read so far %s", read.get()));

						//log.info(String.format("stats so far %s", stats.toString()));

					}

				})

				.map(d -> (String) d.get("body"))

				.filter(Objects::nonNull)

				.collect(Collectors.toList())   // load them in memory first

				.stream()

				.peek(s -> recordTimer.start())

				.map(mapper)

				.forEach(pair -> {

					recordTimer.stop();

					stats.addValue(recordTimer.getTime());

					recordTimer.reset();

					assertNotNull(pair);

					assertTrue(pair.getValue().hasEntity());

					try {

						final org.dom4j.Document doc = new SAXReader().read(new StringReader(pair.getKey()));

						final OafEntity entity = pair.getValue().getEntity();

						//TODO add more asserts

						assertTrue(entity.getId().contains(doc.valueOf("/*[local-name() = 'record']/*[local-name() = 'header']/*[local-name() = 'objIdentifier']/text()")));

					} catch (DocumentException e) {

						throw new IllegalArgumentException("unable to parse record " + pair.getKey(), e);

					}

				});

		totalTimer.stop();

		log.info(String.format("processed %s/%s records in %s", read.get(), collectionSize, HumanTime.exactly(totalTimer.getTime())));

		log.info(stats.toString());

	}

	private Function<String, Pair<String, Oaf>> rowToOaf(final XsltRowTransformer transformer) {

		return xml -> {

			final List<Row> rows = transformer.apply(xml);

			if (rows.isEmpty()) {

				return null;

			}

			return rows.stream()

					.filter(row -> row.getColumn(BODY) != null)

					.map(row -> row.getColumn(BODY))

					.map(c -> c.getValue())

					.map(b -> {

						try {

							return Oaf.parseFrom(b);

						} catch (InvalidProtocolBufferException e) {

							throw new IllegalStateException(e);

						}

					})

					.filter(Objects::nonNull)

					.map(oaf -> new Pair<>(xml, oaf))

					.findFirst()

					.get();

		};

	}

	private InputStream loadFromTransformationProfile(final String profilePath) {

		log.info("Loading xslt from: " + basePathProfiles + profilePath);

		InputStream profile = getClass().getResourceAsStream(basePathProfiles + profilePath);

		SAXReader saxReader = new SAXReader();

		org.dom4j.Document doc = null;

		try {

			doc = saxReader.read(profile);

		} catch (DocumentException e) {

			e.printStackTrace();

			throw new RuntimeException(e);

		}

		String xslt = doc.selectSingleNode("//SCRIPT/CODE/*[local-name()='stylesheet']").asXML();

		//log.info(xslt);

		return IOUtils.toInputStream(xslt);

	}

}

package eu.dnetlib.data.transform.xml;

import java.util.List;

import eu.dnetlib.data.transform.xml.OpenTrialsXsltFunctions.JsonProv;

import org.junit.After;

import org.junit.Before;

import org.junit.Test;

import static org.junit.Assert.assertEquals;

/**

 * OpenTrialsXsltFunctions Tester.

 *

 */

public class OpenTrialsXsltFunctionsTest {

	private String jsonProv = "[{\"url\" : \"http://apps.who.int/trialsearch/Trial3.aspx?trialid=NCT00378508\", \"sourceID\" : \"77eb42c5-0ec7-4e31-963a-5736b66f2d26\", \"sourceName\" : \"ictrp\"},{\"url\" : \"https://www.clinicaltrials.gov/ct2/show/NCT00378508?lup_e=02%2F04%2F2016&lup_s=01%2F01%2F2001&rank=175075&resultsxml=true\", \"sourceID\" : \"b389497c-0833-432b-a09b-930526b7b4d4\", \"sourceName\" : \"nct\"}]";

	private String jsonProvWithNull = "[{\"url\" : \"\", \"sourceID\" : \"77eb42c5-0ec7-4e31-963a-5736b66f2d26\", \"sourceName\" : \"ictrp\"},{\"url\" : \"https://www.clinicaltrials.gov/ct2/show/NCT00378508?lup_e=02%2F04%2F2016&lup_s=01%2F01%2F2001&rank=175075&resultsxml=true\", \"sourceID\" : \"b389497c-0833-432b-a09b-930526b7b4d4\", \"sourceName\" : \"nct\"}]";

	private String jidentifiers = "{112683,NCT00920439}";

	private String jsonRecord = "[{\"source_id\" : \"nct\", \"source_url\" : \"https://clinicaltrials.gov/ct2/show/NCT02321059\", \"is_primary\" : true},{\"source_id\" : \"ictrp\", \"source_url\" : \"http://apps.who.int/trialsearch/Trial3.aspx?trialid=NCT02321059\", \"is_primary\" : false}]";

	private String jsonRecordNull = "[{\"source_id\" : \"nct\", \"source_url\" : \"https://clinicaltrials.gov/ct2/show/NCT02321059\"},{\"source_id\" : \"ictrp\", \"source_url\" : \"http://apps.who.int/trialsearch/Trial3.aspx?trialid=NCT02321059\", \"is_primary\" : false}]";

	private String jsonRecordVoid = "[{\"source_id\" : \"\", \"source_url\" : \"\", \"is_primary\" : \"\"}]";

	private String jsonRecondPrimary = "[{\"source_id\" : \"nct\", \"source_url\" : \"https://clinicaltrials.gov/ct2/show/NCT02321059\", \"is_primary\" : false},{\"source_id\" : \"ictrp\", \"source_url\" : \"http://apps.who.int/trialsearch/Trial3.aspx?trialid=NCT02321059\", \"is_primary\" : false}]";

	private String jsonPeopleVoid ="[{\"person_name\" : null, \"person_id\" : null, \"person_role\" : null}]";

	private String jsonPeopleOne = "[{\"person_name\" : \"Henk Verheul, M.D., PhD\", \"person_id\" : \"116438e9-f8b1-46e5-a1f8-20f851cab73c\", \"person_role\" : \"principal_investigator\"}]";

	private String jsonPeopleMore = "[{\"person_name\" : \"Henk Verheul, M.D., PhD\", \"person_id\" : \"116438e9-f8b1-46e5-a1f8-20f851cab73c\", \"person_role\" : \"principal_investigator\"},{\"person_name\" : \"Miriam Pippolippo Baglioni, PhD\", \"person_id\" : \"fake\", \"person_role\" : \"principal_investigator\"}]";

	private String jsonOrganizationVoid = "[{\"organization_name\" : null, \"organization_id\" : null, \"organization_role\" : null}]";

	private String jsonOrganizationOne = "[{\"organization_name\" : \"Södertälje sjukhus AB\", \"organization_id\" : \"15f0d004-b82b-408c-8605-38a57352468d\", \"organization_role\" : \"sponsor\"}]";

	private String jsonOrganizationMore = "[{\"organization_name\" : \"Södertälje sjukhus AB\", \"organization_id\" : \"15f0d004-b82b-408c-8605-38a57352468d\", \"organization_role\" : \"sponsor\"},{\"organization_name\" : \"Miriam Baglioni AB\", \"organization_id\" : \"fake\", \"organization_role\" : \"primary_sponsor\"}]";

	private String jsonLocationVoid = "[{\"location_name\" : null}]";

	private String jsonLocationOne = "[{\"location_name\" : \"China\"}]";

	private String jsonLocationMore = "[{\"location_name\" : \"China\"},{\"location_name\" : \"North Korea\"}]";

	@Before

	public void before() throws Exception {

	}

	@After

	public void after() throws Exception {

	}

	/**

	 * Method: getProvs(String jsonProvList)

	 */

	@Test

	public void testGetProvs() throws Exception {

		List<JsonProv> list = OpenTrialsXsltFunctions.getProvs(jsonProv);

		assertEquals(2, list.size());

	}

	/**

	 * Method: getMainIdentifierURL(String jsonProvList)

	 */

	@Test

	public void testGetMainIdentifierURL() throws Exception {

		String url = OpenTrialsXsltFunctions.getMainIdentifierURL(jsonProv);

		assertEquals( "http://apps.who.int/trialsearch/Trial3.aspx?trialid=NCT00378508", url );

		url = OpenTrialsXsltFunctions.getMainIdentifierURL(jsonProvWithNull);

		assertEquals("https://www.clinicaltrials.gov/ct2/show/NCT00378508?lup_e=02%2F04%2F2016&lup_s=01%2F01%2F2001&rank=175075&resultsxml=true", url);

	}

	@Test

	public void testGetPrimaryRecordUrl(){

		String url = OpenTrialsXsltFunctions.getPrimaryRecordUrl(jsonRecord);

		assertEquals("https://clinicaltrials.gov/ct2/show/NCT02321059", url);

	}

	@Test

	public void testGetPrimaryRecordID(){

		String id = OpenTrialsXsltFunctions.getPrimaryRecordIdentifier(jsonRecord);

		assertEquals("nct", id);

	}

	@Test

	public void testGetPrimaryRecordUrlNull(){

		String url = OpenTrialsXsltFunctions.getPrimaryRecordUrl(jsonRecordNull);

		assertEquals("https://clinicaltrials.gov/ct2/show/NCT02321059", url);

	}

	@Test

	public void testGetPrimaryRecordUrlVoid(){

		String url = OpenTrialsXsltFunctions.getPrimaryRecordUrl(jsonRecordVoid);

		assertEquals("", url);

	}

	@Test

	public void testGetPrimaryRecordUrlNoPrimary(){

		String url = OpenTrialsXsltFunctions.getPrimaryRecordUrl(jsonRecondPrimary);

		assertEquals("https://clinicaltrials.gov/ct2/show/NCT02321059", url);

	}

	@Test

	public void testGetPrimaryRecordIDNoPrimary(){

		String id = OpenTrialsXsltFunctions.getPrimaryRecordIdentifier(jsonRecondPrimary);

		assertEquals("nct", id);

	}

	@Test

	public void testGetPrincipalInvestigatorsVoid(){

		String url = OpenTrialsXsltFunctions.getPrincipalInvestigators(jsonPeopleVoid);

		assertEquals("",url);

	}

	@Test

	public void testGetPrincipalInvestigatorsOne(){

		String url = OpenTrialsXsltFunctions.getPrincipalInvestigators(jsonPeopleOne);

		assertEquals("Verheul, Henk", url);

	}

	@Test

	public void testGetPrincipalInvestigatorsMore(){

		String url = OpenTrialsXsltFunctions.getPrincipalInvestigators(jsonPeopleMore);

		assertEquals("Verheul, Henk@@Baglioni, Miriam Pippolippo", url);

	}

	@Test

	public void testgGetTrialOrganizationsVoid(){

		String url = OpenTrialsXsltFunctions.getTrialOrganizations(jsonOrganizationVoid);

		assertEquals("",url);

	}

	@Test

	public void testgGetTrialOrganizationsOne(){

		String url = OpenTrialsXsltFunctions.getTrialOrganizations(jsonOrganizationOne);

		assertEquals("Södertälje sjukhus AB@sponsor", url);

	}

	@Test

	public void testgGetTrialOrganizationsMore(){

		String url = OpenTrialsXsltFunctions.getTrialOrganizations(jsonOrganizationMore);

		assertEquals("Södertälje sjukhus AB@sponsor@@Miriam Baglioni AB@sponsor", url);

	}

	@Test

	public void testgGetTrialLocationsVoid(){

		String url = OpenTrialsXsltFunctions.getTrialLocations(jsonLocationVoid);

		assertEquals("",url);

	}

	@Test

	public void testgGetTrialLocationsOne(){

		String url = OpenTrialsXsltFunctions.getTrialLocations(jsonLocationOne);

		assertEquals("China", url);

	}

	@Test

	public void testgGetTrialLocationsMore(){

		String url = OpenTrialsXsltFunctions.getTrialLocations(jsonLocationMore);

		assertEquals("China@@North Korea", url);

	}

	@Test

	public void testGetNotPrimaryRecordUrlPrimary(){

		String url = OpenTrialsXsltFunctions.getNotPrimaryRecordUrl(jsonRecondPrimary);

		assertEquals("http://apps.who.int/trialsearch/Trial3.aspx?trialid=NCT02321059", url);

	}

	@Test

	public void testGetNotPrimaryRecordUrlVoid(){

		String url = OpenTrialsXsltFunctions.getNotPrimaryRecordUrl(jsonRecordVoid);

		assertEquals("", url);

	}

	@Test

	public void testGetNotPrimaryRecordUrl(){

		String url = OpenTrialsXsltFunctions.getNotPrimaryRecordUrl(jsonRecord);

		assertEquals("http://apps.who.int/trialsearch/Trial3.aspx?trialid=NCT02321059", url);

	}

}

package eu.dnetlib.data.transform.xml;

import org.junit.After;

import org.junit.Before;

import org.junit.Test;

import static org.junit.Assert.assertEquals;

import static org.junit.Assert.assertNotNull;

/**

 *

 * Created by miriam on 04/05/2017.

 */

public class FWFXsltFunctionsTest {

	private String namesurname ="Gerhard SOMMER";

	private String noSurname = "Gerhard";

	private String noName = "SOMMER";

	private String twoNames = "Gerhard Pippo SOMMER";

	private String twoSurname = "Gerhard PIPPO SOMMER";

	private String nonamesurname = "";

	private String organization ="Universität Linz - Institut für Computational Perception; Universität für Musik und darstellende Kunst Graz - Institut 1: Komposition, Musiktheorie, Musikgeschichte und Dirigieren; Universität Mozarteum Salzburg - Institut für Musikalische Rezeptions- und Interpretationsgeschichte; Anton Bruckner Privatuniversität - Institut für Theorie und Geschichte der Musik; Eliette und Herbert von Karajan Institut - Eliette und Herbert von Karajan Institut";

	@Before

	public void before() throws Exception {

	}

	@After

	public void after() throws Exception {

	}

	@Test

	public void testGetNamesNameNoNameSurname() throws Exception {

		String ret = FWFXsltFunctions.getName(nonamesurname,true);

		assertEquals("",ret );

	}

	@Test

	public void testGetNamesSurnameNoNameSurname() throws Exception {

		String ret = FWFXsltFunctions.getName(nonamesurname,false);

		assertEquals("",ret );

	}

	@Test

	public void testGetNamesNameTwoSurname() throws Exception {

		String ret = FWFXsltFunctions.getName(twoSurname,true);

		assertEquals("Gerhard",ret );

	}

	@Test

	public void testGetNamesSurnameTwoSurname() throws Exception {

		String ret = FWFXsltFunctions.getName(twoSurname,false);

		assertEquals("PIPPO SOMMER",ret );

	}

	@Test

	public void testGetNamesNameTwoNames() throws Exception {

		String ret = FWFXsltFunctions.getName(twoNames,true);

		assertEquals("Gerhard Pippo",ret );

	}

	@Test

	public void testGetNamesSurnameTwoNames() throws Exception {

		String ret = FWFXsltFunctions.getName(twoNames,false);

		assertEquals("SOMMER",ret );

	}

	/**

	 * Method: getProvs(String jsonProvList)

	 */

	@Test

	public void testGetNamesName() throws Exception {

		String ret = FWFXsltFunctions.getName(namesurname,true);

		assertEquals("Gerhard",ret );

	}

	@Test

	public void testGetNamesSurname() throws Exception {

		String ret = FWFXsltFunctions.getName(namesurname,false);

		assertEquals("SOMMER",ret );

	}

	@Test

	public void testGetNamesNameNoSurname() throws Exception {

		String ret = FWFXsltFunctions.getName(noSurname,true);

		assertEquals("Gerhard",ret );

	}

	@Test

	public void testGetNamesSurnameNoSurname() throws Exception {

		String ret = FWFXsltFunctions.getName(noSurname,false);

		assertEquals("",ret );

	}

	@Test

	public void testGetNamesNameNoName() throws Exception {

		String ret = FWFXsltFunctions.getName(noName,true);

		assertEquals("",ret );

	}

	@Test

	public void testGetNamesSurnameNoName() throws Exception {

		String ret = FWFXsltFunctions.getName(noName,false);

		assertEquals("SOMMER",ret );

	}

	@Test

	public void TestGetMd5()throws Exception{

		String md5 = FWFXsltFunctions.getMd5(organization);

		System.out.println(md5);

		assertNotNull(md5);

	}

}

package eu.dnetlib.data.transform.xml.vtd;

import java.io.IOException;

import java.util.function.Function;

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

import eu.dnetlib.data.transform.xml2.DatasetToProto;

import eu.dnetlib.data.transform.xml2.PublicationToProto;

import org.apache.commons.io.IOUtils;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

import org.junit.Test;

import static org.junit.Assert.assertNotNull;

public class VtdParserToProtoTest {

	private static final Log log = LogFactory.getLog(VtdParserToProtoTest.class);

	@Test

	public void testParsePublication() throws IOException {

		doTest("/eu/dnetlib/data/transform/publication.xml", new PublicationToProto());

	}

	@Test

	public void testParseDataset() throws IOException {

		doTest("/eu/dnetlib/data/transform/dataset.xml", new DatasetToProto());

	}

	@Test

	public void testParseDataset2() throws IOException {

		doTest("/eu/dnetlib/data/transform/dataset2.xml", new DatasetToProto());

	}

	private void doTest(final String filePath, Function<String, Oaf> f) throws IOException {

		final String xml = IOUtils.toString(getClass().getResourceAsStream(filePath));

		assertNotNull(xml);

		final Oaf oaf = f.apply(xml);

		assertNotNull(oaf);

		log.info(oaf.toString());

	}

}

package eu.dnetlib.data.transform.xml.vtd;

import java.io.IOException;

import java.io.InputStream;

import java.util.Properties;

import com.mongodb.MongoClient;

import com.mongodb.client.MongoDatabase;

import org.springframework.context.annotation.Bean;

import org.springframework.context.annotation.Configuration;

import org.springframework.core.io.ClassPathResource;

@Configuration

public class ConfigurationTestConfig {

	@Bean

	public MongoDatabase db() throws IOException {

		final Properties p = new Properties();

		final ClassPathResource cp = new ClassPathResource("/eu/dnetlib/data/transform/xml/vtd/test.properties");

		try (final InputStream stream = cp.getInputStream()) {

			p.load(stream);

		}

		final MongoClient mongo = new MongoClient(

				p.getProperty("mongodb.host"),

				Integer.valueOf(p.getProperty("mongodb.port")));

		return mongo.getDatabase(p.getProperty("mongodb.dbname"));

	}

}

package eu.dnetlib.data.transform.xml.vtd;

import com.mongodb.client.MongoCollection;

import com.mongodb.client.MongoDatabase;

import eu.dnetlib.data.transform.xml2.DatasetToProto;

import org.apache.commons.lang3.time.StopWatch;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

import org.apache.commons.math.stat.descriptive.DescriptiveStatistics;

import org.bson.Document;

import org.junit.Test;

import org.junit.runner.RunWith;

import org.springframework.beans.factory.annotation.Autowired;

import org.springframework.test.context.ContextConfiguration;

import org.springframework.test.context.junit4.SpringJUnit4ClassRunner;

import java.io.IOException;

import java.util.Objects;

import java.util.concurrent.atomic.AtomicInteger;

import java.util.stream.StreamSupport;

import static org.junit.Assert.assertNotNull;

import static org.junit.Assert.assertTrue;

@RunWith(SpringJUnit4ClassRunner.class)

@ContextConfiguration(classes = { ConfigurationTestConfig.class })

public class VtdParserToProtoIT {

	private static final Log log = LogFactory.getLog(VtdParserToProtoIT.class);

	public static final String COLLECTION_NAME = "datacite";

	private static final int BATCH_SIZE = 10000;

	public static final int LOG_FREQ = 5000;

	@Autowired

	private MongoDatabase db;

	@Test

	public void testParseDatacite() throws IOException {

		final MongoCollection<Document> collection = db.getCollection(COLLECTION_NAME);

		final long collectionSize = collection.count();

		log.info(String.format("found %s records in collection '%s'", collectionSize, COLLECTION_NAME));

		final AtomicInteger read = new AtomicInteger(0);

		final DescriptiveStatistics stats = new DescriptiveStatistics();

		final StopWatch timer = new StopWatch();

		final DatasetToProto mapper = new DatasetToProto();

		StreamSupport.stream(collection.find().batchSize(BATCH_SIZE).spliterator(), false)

				.peek(d -> {

					if (read.addAndGet(1) % LOG_FREQ == 0) {

						log.info(String.format("records read so far %s", read.get()));

						log.info(String.format("stats so far %s", stats.toString()));

					}

				})

				.map(d -> (String) d.get("body"))

				.filter(Objects::nonNull)

				.peek(s -> timer.start())

				.map(mapper)

				.forEach(oaf -> {

					assertNotNull(oaf);

					assertTrue(oaf.hasEntity());

					timer.stop();

					stats.addValue(timer.getTime());

					timer.reset();

				});

		log.info(String.format("processed %s/%s records", read.get(), collectionSize));

	}

}

package eu.dnetlib.data.transform.xml.vtd;

import java.io.InputStream;

import java.util.List;

import com.ximpleware.AutoPilot;

import com.ximpleware.VTDGen;

import com.ximpleware.VTDNav;

import eu.dnetlib.data.transform.xml2.Node;

import eu.dnetlib.data.transform.xml2.VtdUtilityParser;

import org.apache.commons.io.IOUtils;

import org.junit.Assert;

import org.junit.Test;

public class VtdUtilityParserTest {

	@Test

	public void testUtils1() {

		String xpath = VtdUtilityParser.xpath("a", "b", "c");

		Assert.assertTrue("/*[local-name()='a']/*[local-name()='b']/*[local-name()='c']".equals(xpath));

	}

    @Test

    public void testPartser() throws Exception {

        final InputStream resource = this.getClass().getResourceAsStream("/eu/dnetlib/data/transform/publication.xml");

        final String record =IOUtils.toString(resource);

        final VTDGen vg = VtdUtilityParser.parseXml(record);

        final VTDNav vn = vg.getNav();

        final AutoPilot ap = new AutoPilot(vn);

        List<Node> nodes = VtdUtilityParser.getNodes(ap, vn, "//*[local-name()='referenceaa']");

        nodes.forEach(n -> Assert.assertTrue(n.getAttributes().keySet().size()>0));

        System.out.println(VtdUtilityParser.countNodes(ap, vn, "count(//*[local-name()='CobjIdentifier'])"));

    }

}

package eu.dnetlib.data.transform.xml2;

import java.io.IOException;

import java.util.function.Function;

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

import org.apache.commons.io.IOUtils;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

import org.junit.Test;

import static org.junit.Assert.assertNotNull;

public class VtdParserToProtoTest {

	private static final Log log = LogFactory.getLog(VtdParserToProtoTest.class);

	@Test

	public void testParsePublication() throws IOException {

		doTest("/eu/dnetlib/data/transform/publication.xml", new PublicationToProto());

	}

	@Test

	public void testParseDataset() throws IOException {

		doTest("/eu/dnetlib/data/transform/dataset.xml", new DatasetToProto());

	}

	@Test

		public void testParseDataset2() throws IOException {

		doTest("/eu/dnetlib/data/transform/dataset2.xml", new DatasetToProto());

	}

	private void doTest(final String filePath, Function<String, Oaf> f) throws IOException {

		final String xml = IOUtils.toString(getClass().getResourceAsStream(filePath));

		assertNotNull(xml);

		final Oaf oaf = f.apply(xml);

		assertNotNull(oaf);

		log.info(oaf.toString());

	}

}

package eu.dnetlib.data.transform.xml2;

import java.io.InputStream;

import java.util.List;

import com.ximpleware.AutoPilot;

import com.ximpleware.VTDGen;

import com.ximpleware.VTDNav;

import eu.dnetlib.data.transform.xml2.Node;

import eu.dnetlib.data.transform.xml2.VtdUtilityParser;

import org.apache.commons.io.IOUtils;

import org.junit.Assert;

import org.junit.Test;

import static eu.dnetlib.data.transform.xml2.VtdUtilityParser.parseXml;

public class VtdUtilityParserTest {

	@Test

	public void testUtils1() {

		String xpath = VtdUtilityParser.xpath("a", "b", "c");

		Assert.assertTrue("/*[local-name()='a']/*[local-name()='b']/*[local-name()='c']".equals(xpath));

	}

    @Test

    public void testPartser() throws Exception {

        final InputStream resource = this.getClass().getResourceAsStream("/eu/dnetlib/data/transform/publication.xml");

        final String record =IOUtils.toString(resource);

        final VTDGen vg = parseXml(record);

        final VTDNav vn = vg.getNav();

        final AutoPilot ap = new AutoPilot(vn);

        List<Node> nodes = VtdUtilityParser.getNodes(ap, vn, "//*[local-name()='referenceaa']");

        nodes.forEach(n -> Assert.assertTrue(n.getAttributes().keySet().size()>0));

        System.out.println(VtdUtilityParser.countNodes(ap, vn, "count(//*[local-name()='CobjIdentifier'])"));

    }

}

package eu.dnetlib.data.transform;

import static org.junit.Assert.assertFalse;

import static org.junit.Assert.assertNotNull;

import java.io.IOException;

import java.io.StringWriter;

import org.apache.commons.codec.binary.Base64;

import org.apache.commons.io.IOUtils;

import org.apache.commons.logging.Log;

import org.apache.commons.logging.LogFactory;

import org.apache.solr.common.SolrInputDocument;

import org.apache.solr.common.SolrInputField;

import org.dom4j.DocumentException;

import org.junit.Before;

import org.junit.Test;

import com.google.protobuf.InvalidProtocolBufferException;

import com.googlecode.protobuf.format.JsonFormat;

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

import eu.dnetlib.data.proto.KindProtos.Kind;

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

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

import eu.dnetlib.functionality.index.solr.feed.InputDocumentFactory;

public class SolrProtoMapperTest {

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

	private String fields;

	@Before

	public void setUp() throws IOException {

		final StringWriter sw = new StringWriter();

		IOUtils.copy(getClass().getResourceAsStream("fields.xml"), sw);

		fields = sw.toString();

		assertNotNull(fields);

		assertFalse(fields.isEmpty());

		log.info(fields);

	}

	@Test

	public void testProto2SolrDocument() throws DocumentException, InvalidProtocolBufferException {

		final SolrProtoMapper mapper = new SolrProtoMapper(fields);

		assertNotNull(mapper);

		final OafEntity.Builder entity = OafTest.getResultBuilder("01");

		entity.addChildren(OafTest.getResultBuilder("01_children"));

		final Oaf oaf = OafTest.embed(entity.build(), Kind.entity).getOaf();

		assertNotNull(oaf.getEntity().getChildrenList());

		assertFalse(oaf.getEntity().getChildrenList().isEmpty());

		log.info("byte[] size: " + oaf.toByteArray().length);

		log.info("json size:   " + JsonFormat.printToString(oaf).length());

		log.info("base64 size: " + Base64.encodeBase64String(oaf.toByteArray()).length());

		final byte[] decodeBase64 = Base64.decodeBase64(Base64.encodeBase64String(oaf.toByteArray()));

		log.info("decoded: " + JsonFormat.printToString(Oaf.parseFrom(decodeBase64)));

		final SolrInputDocument doc = mapper.map(oaf, InputDocumentFactory.getParsedDateField("2015-02-15"), "asd", "action-set");

		assertNotNull(doc);

		for (final SolrInputField f : doc.values()) {

			log.info(f);

		}

	}

}

package eu.dnetlib.data.transform;

import java.io.IOException;

import java.io.InputStream;

import java.util.Properties;

import com.mongodb.MongoClient;

import com.mongodb.client.MongoDatabase;

import org.springframework.context.annotation.Bean;

import org.springframework.context.annotation.Configuration;

import org.springframework.core.io.ClassPathResource;

@Configuration

public class ConfigurationTestConfig {

	@Bean

	public MongoDatabase db() throws IOException {

		final Properties p = testProperties();

		final MongoClient mongo = new MongoClient(

				p.getProperty("mongodb.host"),

				Integer.valueOf(p.getProperty("mongodb.port")));

		return mongo.getDatabase(p.getProperty("mongodb.dbname"));

	}

	@Bean

	public Properties testProperties() throws IOException {

		final Properties p = new Properties();

		final ClassPathResource cp = new ClassPathResource("test.properties");

		try (final InputStream stream = cp.getInputStream()) {

			p.load(stream);

		}

		return p;

	}

}

{"projectId":"corda_______::200130","acronym":"SUPERCOMPETITORS","code":"200130","optional1":"ERC-SG","optional2":"http://erc.europa.eu","jsonextrainfo":"{}","fundingPath":["\u003cfundingtree\u003e\u003cfunder\u003e\u003cid\u003eec__________::EC\u003c/id\u003e\u003cshortname\u003eEC\u003c/shortname\u003e\u003cname\u003eEuropean Commission\u003c/name\u003e\u003cjurisdiction\u003eEU\u003c/jurisdiction\u003e\u003c/funder\u003e\u003cfunding_level_2\u003e\u003cid\u003eec__________::EC::FP7::SP2::ERC\u003c/id\u003e\u003cdescription\u003eERC\u003c/description\u003e\u003cname\u003eERC\u003c/name\u003e\u003cclass\u003eec:program\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_1\u003e\u003cid\u003eec__________::EC::FP7::SP2\u003c/id\u003e\u003cdescription\u003eSP2-Ideas\u003c/description\u003e\u003cname\u003eSP2\u003c/name\u003e\u003cclass\u003eec:specificprogram\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_0\u003e\u003cid\u003eec__________::EC::FP7\u003c/id\u003e\u003cdescription\u003eSEVENTH FRAMEWORK PROGRAMME\u003c/description\u003e\u003cname\u003eFP7\u003c/name\u003e\u003cparent/\u003e\u003cclass\u003eec:frameworkprogram\u003c/class\u003e\u003c/funding_level_0\u003e\u003c/parent\u003e\u003c/funding_level_1\u003e\u003c/parent\u003e\u003c/funding_level_2\u003e\u003c/fundingtree\u003e"]}

{"projectId":"corda_______::200141","acronym":"QUESPACE","code":"200141","optional1":"ERC-SG","optional2":"http://erc.europa.eu","jsonextrainfo":"{}","fundingPath":["\u003cfundingtree\u003e\u003cfunder\u003e\u003cid\u003eec__________::EC\u003c/id\u003e\u003cshortname\u003eEC\u003c/shortname\u003e\u003cname\u003eEuropean Commission\u003c/name\u003e\u003cjurisdiction\u003eEU\u003c/jurisdiction\u003e\u003c/funder\u003e\u003cfunding_level_2\u003e\u003cid\u003eec__________::EC::FP7::SP2::ERC\u003c/id\u003e\u003cdescription\u003eERC\u003c/description\u003e\u003cname\u003eERC\u003c/name\u003e\u003cclass\u003eec:program\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_1\u003e\u003cid\u003eec__________::EC::FP7::SP2\u003c/id\u003e\u003cdescription\u003eSP2-Ideas\u003c/description\u003e\u003cname\u003eSP2\u003c/name\u003e\u003cclass\u003eec:specificprogram\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_0\u003e\u003cid\u003eec__________::EC::FP7\u003c/id\u003e\u003cdescription\u003eSEVENTH FRAMEWORK PROGRAMME\u003c/description\u003e\u003cname\u003eFP7\u003c/name\u003e\u003cparent/\u003e\u003cclass\u003eec:frameworkprogram\u003c/class\u003e\u003c/funding_level_0\u003e\u003c/parent\u003e\u003c/funding_level_1\u003e\u003c/parent\u003e\u003c/funding_level_2\u003e\u003c/fundingtree\u003e"]}

{"projectId":"corda_______::200165","acronym":"PROTEOMICS OF CHERNO","code":"200165","optional1":"MC-IRG","optional2":"http://cordis.europa.eu/fp7/people.htm","jsonextrainfo":"{}","fundingPath":["\u003cfundingtree\u003e\u003cfunder\u003e\u003cid\u003eec__________::EC\u003c/id\u003e\u003cshortname\u003eEC\u003c/shortname\u003e\u003cname\u003eEuropean Commission\u003c/name\u003e\u003cjurisdiction\u003eEU\u003c/jurisdiction\u003e\u003c/funder\u003e\u003cfunding_level_2\u003e\u003cid\u003eec__________::EC::FP7::SP3::PEOPLE\u003c/id\u003e\u003cdescription\u003eMarie-Curie Actions\u003c/description\u003e\u003cname\u003ePEOPLE\u003c/name\u003e\u003cclass\u003eec:program\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_1\u003e\u003cid\u003eec__________::EC::FP7::SP3\u003c/id\u003e\u003cdescription\u003eSP3-People\u003c/description\u003e\u003cname\u003eSP3\u003c/name\u003e\u003cclass\u003eec:specificprogram\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_0\u003e\u003cid\u003eec__________::EC::FP7\u003c/id\u003e\u003cdescription\u003eSEVENTH FRAMEWORK PROGRAMME\u003c/description\u003e\u003cname\u003eFP7\u003c/name\u003e\u003cparent/\u003e\u003cclass\u003eec:frameworkprogram\u003c/class\u003e\u003c/funding_level_0\u003e\u003c/parent\u003e\u003c/funding_level_1\u003e\u003c/parent\u003e\u003c/funding_level_2\u003e\u003c/fundingtree\u003e"]}

nih_________::3R01GM073898-02S1,,3R01GM073898-02S1,23188 $,,"{""orgname"":""UNIVERSITY OF CALIFORNIA SAN DIEGO"", ""activity"":""R01"", ""administeringic"":""GM"", ""serialnumber"":""73898"", ""coreprojectnum"":""R01GM073898""}","[""\u003cfundingtree\u003e\n      \u003cfunder\u003e\n         \u003cid\u003enih_________::NIH\u003c/id\u003e\n         \u003cshortname\u003eNIH\u003c/shortname\u003e\n         \u003cname\u003eNational Institutes of Health\u003c/name\u003e\n         \u003cjurisdiction\u003eUS\u003c/jurisdiction\u003e\n      \u003c/funder\u003e\n      \u003cfunding_level_0\u003e\n         \u003cid\u003enih_________::NIH::NATIONAL_INSTITUTE_OF_GENERAL_MEDICAL_SCIENCES\u003c/id\u003e\n         \u003cname\u003eNATIONAL INSTITUTE OF GENERAL MEDICAL SCIENCES\u003c/name\u003e\n         \u003cdescription\u003eNATIONAL INSTITUTE OF GENERAL MEDICAL SCIENCES\u003c/description\u003e\n         \u003cparent/\u003e\n         \u003cclass\u003enih:fundingStream\u003c/class\u003e\n      \u003c/funding_level_0\u003e\n   \u003c/fundingtree\u003e""]"

corda_______::100202,RECOMP,100202,JTI-CP-ARTEMIS,http://cordis.europa.eu/fp7/home_en.html,{},"[""\u003cfundingtree\u003e\u003cfunder\u003e\u003cid\u003eec__________::EC\u003c/id\u003e\u003cshortname\u003eEC\u003c/shortname\u003e\u003cname\u003eEuropean Commission\u003c/name\u003e\u003cjurisdiction\u003eEU\u003c/jurisdiction\u003e\u003c/funder\u003e\u003cfunding_level_2\u003e\u003cid\u003eec__________::EC::FP7::SP1::SP1-JTI\u003c/id\u003e\u003cdescription\u003eJoint Technology Initiatives (Annex IV-SP1)\u003c/description\u003e\u003cname\u003eSP1-JTI\u003c/name\u003e\u003cclass\u003eec:program\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_1\u003e\u003cid\u003eec__________::EC::FP7::SP1\u003c/id\u003e\u003cdescription\u003eSP1-Cooperation\u003c/description\u003e\u003cname\u003eSP1\u003c/name\u003e\u003cclass\u003eec:specificprogram\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_0\u003e\u003cid\u003eec__________::EC::FP7\u003c/id\u003e\u003cdescription\u003eSEVENTH FRAMEWORK PROGRAMME\u003c/description\u003e\u003cname\u003eFP7\u003c/name\u003e\u003cparent/\u003e\u003cclass\u003eec:frameworkprogram\u003c/class\u003e\u003c/funding_level_0\u003e\u003c/parent\u003e\u003c/funding_level_1\u003e\u003c/parent\u003e\u003c/funding_level_2\u003e\u003c/fundingtree\u003e""]"

corda__h2020::633080,MACC-III,633080,SPACE,SPACE,{},"[""\u003cfundingtree\u003e\u003cfunder\u003e\u003cid\u003eec__________::EC\u003c/id\u003e\u003cshortname\u003eEC\u003c/shortname\u003e\u003cname\u003eEuropean Commission\u003c/name\u003e\u003cjurisdiction\u003eEU\u003c/jurisdiction\u003e\u003c/funder\u003e\u003cfunding_level_1\u003e\u003cid\u003eec__________::EC::H2020::CSA\u003c/id\u003e\u003cdescription\u003eCoordination and support action\u003c/description\u003e\u003cname\u003eCSA\u003c/name\u003e\u003cclass\u003eec:h2020toas\u003c/class\u003e\u003cparent\u003e\u003cfunding_level_0\u003e\u003cid\u003eec__________::EC::H2020\u003c/id\u003e\u003cname\u003eH2020\u003c/name\u003e\u003cdescription\u003eHorizon 2020 Framework Programme\u003c/description\u003e\u003cparent/\u003e\u003cclass\u003eec:h2020fundings\u003c/class\u003e\u003c/funding_level_0\u003e\u003c/parent\u003e\u003c/funding_level_1\u003e\u003c/fundingtree\u003e""]"

nsf_________::0000096,,0000096,,,{},"[""\u003cfundingtree\u003e\u003cfunder\u003e\u003cid\u003ensf_________::NSF\u003c/id\u003e\u003cshortname\u003eNSF\u003c/shortname\u003e\u003cname\u003eNational Science Foundation\u003c/name\u003e\u003cjurisdiction\u003eUS\u003c/jurisdiction\u003e\u003c/funder\u003e\u003cfunding_level_1\u003e\u003cid\u003ensf_________::NSF::OD::OD/OIA\u003c/id\u003e\u003cdescription\u003eOffice of Integrative Activities\u003c/description\u003e\u003cname\u003eOffice of Integrative Activities\u003c/name\u003e\u003cparent\u003e\u003cfunding_level_0\u003e\u003cid\u003ensf_________::NSF::OD\u003c/id\u003e\u003cdescription\u003eOffice of the Director\u003c/description\u003e\u003cname\u003eOffice of the Director\u003c/name\u003e\u003cparent/\u003e\u003cclass\u003ensf:fundingStream\u003c/class\u003e\u003c/funding_level_0\u003e\u003c/parent\u003e\u003c/funding_level_1\u003e\u003c/fundingtree\u003e""]"

fct_________::100107,PTDC/SAU-ESA/100107/2008,100107,,,{},"[""\u003cfundingtree\u003e\u003cfunder\u003e\u003cid\u003efct_________::FCT\u003c/id\u003e\u003cshortname\u003eFCT\u003c/shortname\u003e\u003cname\u003eFundação para a Ciência e a Tecnologia, I.P.\u003c/name\u003e\u003cjurisdiction\u003ePT\u003c/jurisdiction\u003e\u003c/funder\u003e\u003cfunding_level_0\u003e\u003cid\u003efct_________::FCT::5876-PPCDTI\u003c/id\u003e\u003cdescription\u003e5876-PPCDTI\u003c/description\u003e\u003cname\u003e5876-PPCDTI\u003c/name\u003e\u003cparent/\u003e\u003cclass\u003efct:program\u003c/class\u003e\u003c/funding_level_0\u003e\u003c/fundingtree\u003e""]"

<communities>

    <community id="fet-fp7">

        <oacommunity/>

        <subjects/>

        <datasources/>

        <zenodocommunities/>

    </community>

    <community id="fet-h2020">

        <subjects/>

        <datasources/>

        <zenodocommunities/>

    </community>

    <community id="oa-pg">

        <subjects/>

        <datasources/>

        <zenodocommunities/>

    </community>

    <community id="ee">

        <subjects>

            <subject>SDG13 - Climate action</subject>

            <subject>SDG8 - Decent work and economic growth</subject>

            <subject>SDG15 - Life on land</subject>

            <subject>SDG2 - Zero hunger</subject>

            <subject>SDG17 - Partnerships for the goals</subject>

            <subject>SDG10 - Reduced inequalities</subject>

            <subject>SDG5 - Gender equality</subject>

            <subject>SDG12 - Responsible consumption and production</subject>

            <subject>SDG14 - Life below water</subject>

            <subject>SDG6 - Clean water and sanitation</subject>

            <subject>SDG11 - Sustainable cities and communities</subject>

            <subject>SDG1 - No poverty</subject>

            <subject>SDG3 - Good health and well being</subject>

            <subject>SDG7 - Affordable and clean energy</subject>

            <subject>SDG4 - Quality education</subject>

            <subject>SDG9 - Industry innovation and infrastructure</subject>

            <subject>SDG16 - Peace justice and strong institutions</subject>

        </subjects>

        <datasources/>

        <zenodocommunities>

            <zenodocommunity>

                <zenodoid>123</zenodoid>

                <selcriteria>800A</selcriteria>

            </zenodocommunity>

        </zenodocommunities>

    </community>

    <community id="dh-ch">

        <subjects/>

        <datasources/>

        <zenodocommunities/>

    </community>

    <community id="fam">

        <subjects/>

        <datasources/>

        <zenodocommunities/>

    </community>

    <community id="ni">

        <subjects>

            <subject>brain mapping</subject>

            <subject>brain imaging</subject>

            <subject>electroencephalography</subject>

            <subject>arterial spin labelling</subject>

            <subject>brain fingerprinting</subject>

            <subject>brain</subject>

            <subject>neuroimaging</subject>

            <subject>Multimodal Brain Image Analysis</subject>

            <subject>fMRI</subject>

            <subject>neuroinformatics</subject>

            <subject>fetal brain</subject>

            <subject>brain ultrasonic imaging</subject>

            <subject>topographic brain mapping</subject>

            <subject>diffusion tensor imaging</subject>

            <subject>computerized knowledge assessment</subject>

            <subject>connectome mapping</subject>

            <subject>brain magnetic resonance imaging</subject>

            <subject>brain abnormalities</subject>

        </subjects>

        <datasources>

            <datasource>

                <openaireId>re3data_____::5b9bf9171d92df854cf3c520692e9122</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>doajarticles::c7d3de67dc77af72f6747157441252ec</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>re3data_____::8515794670370f49c1d176c399c714f5</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>doajarticles::d640648c84b10d425f96f11c3de468f3</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>doajarticles::0c0e74daa5d95504eade9c81ebbd5b8a</openaireId>

                <selcriteria/>

            </datasource>

        </datasources>

        <zenodocommunities/>

    </community>

    <community id="mes">

        <subjects>

            <subject>marine</subject>

            <subject>ocean</subject>

            <subject>fish</subject>

            <subject>aqua</subject>

            <subject>sea</subject>

        </subjects>

        <datasources>

            <datasource>

                <openaireId>re3data_____::9633d1e8c4309c833c2c442abeb0cfeb</openaireId>

                <selcriteria/>

            </datasource>

        </datasources>

        <zenodocommunities/>

    </community>

    <community id="aginfra">

        <subjects>

            <subject>animal production and health</subject>

            <subject>fisheries and aquaculture</subject>

            <subject>food safety and human nutrition</subject>

            <subject>information management</subject>

            <subject>food technology</subject>

            <subject>agri-food education and extension</subject>

            <subject>natural resources and environment</subject>

            <subject>food system</subject>

            <subject>engineering technology and Research</subject>

            <subject>agriculture</subject>

            <subject>food safety risk assessment</subject>

            <subject>food security</subject>

            <subject>farming practices and systems</subject>

            <subject>plant production and protection</subject>

            <subject>agri-food economics and policy</subject>

            <subject>food distribution</subject>

            <subject>forestry</subject>

        </subjects>

        <datasources>

            <datasource>

                <openaireId>opendoar____::1a551829d50f1400b0dab21fdd969c04</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>opendoar____::49af6c4e558a7569d80eee2e035e2bd7</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>opendoar____::0266e33d3f546cb5436a10798e657d97</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>opendoar____::fd4c2dc64ccb8496e6f1f94c85f30d06</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>opendoar____::41bfd20a38bb1b0bec75acf0845530a7</openaireId>

                <selcriteria/>

            </datasource>

            <datasource>

                <openaireId>opendoar____::87ae6fb631f7c8a627e8e28785d9992d</openaireId>

                <selcriteria/>

            </datasource>

        </datasources>

        <zenodocommunities/>

    </community>

    <community id="clarin">

        <oacommunity>oac_clarin</oacommunity>

        <subjects/>

        <datasources>

            <datasource>

                <openaireId>re3data_____::a507cdacc5bbcc08761c92185dee5cab</openaireId>

                <selcriteria/>

            </datasource>

        </datasources>

        <zenodocommunities/>

    </community>

</communities>

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

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

        xmlns:oaf="http://namespace.openaire.eu/oaf" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">

	<oai:header xmlns="http://namespace.openaire.eu/"

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

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

	            xmlns:oai="http://www.openarchives.org/OAI/2.0/" xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance">

		<dri:objIdentifier>r37b0ad08687::00066f8025c9f9c99eaf3bba41e0ee0b</dri:objIdentifier>

		<dri:recordIdentifier>10.5281/zenodo.11877</dri:recordIdentifier>

		<dri:dateOfCollection>2018-04-30T02:17:56.67Z</dri:dateOfCollection>

		<oaf:datasourceprefix>r37b0ad08687</oaf:datasourceprefix>

		<identifier xmlns="http://www.openarchives.org/OAI/2.0/">oai:zenodo.org:11877</identifier>

		<datestamp xmlns="http://www.openarchives.org/OAI/2.0/">2017-05-30T02:33:14Z</datestamp>

		<setSpec xmlns="http://www.openarchives.org/OAI/2.0/">user-dighl</setSpec>

		<setSpec xmlns="http://www.openarchives.org/OAI/2.0/">openaire_data</setSpec>

		<setSpec xmlns="http://www.openarchives.org/OAI/2.0/">user-zenodo</setSpec>

		<dr:dateOfTransformation>2018-04-30T02:21:21.245Z</dr:dateOfTransformation>

	</oai:header>

	<metadata>

		<resource xmlns="http://datacite.org/schema/kernel-3"

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

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

		          xmlns:oai="http://www.openarchives.org/OAI/2.0/"

		          xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance" xsi:schemaLocation="http://datacite.org/schema/kernel-3 http://schema.datacite.org/meta/kernel-3/metadata.xsd">

			<identifier identifierType="DOI">10.5281/zenodo.11877</identifier>

			<alternateIdentifiers xmlns="http://www.openarchives.org/OAI/2.0/">

				<alternateIdentifier alternateIdentifierType="URL">http://dx.doi.org/10.5281/zenodo.11877</alternateIdentifier>

			</alternateIdentifiers>

			<creators>

				<creator>

					<creatorName>Johann-Mattis List</creatorName>

				</creator>

			</creators>

			<titles>

				<title>Supplementary Material For "Sequence Comparison In Historical Linguistics"</title>

			</titles>

			<publisher>Zenodo</publisher>

			<publicationYear>2014</publicationYear>

			<subjects>

				<subject>historical linguistics</subject>

				<subject>phonetic alignment</subject>

				<subject>cognate detection</subject>

				<subject>computational linguistics</subject>

			</subjects>

			<dates>

				<date dateType="Issued">2014-09-26</date>

			</dates>

			<resourceType resourceTypeGeneral="Dataset"/>

			<relatedIdentifiers>

				<relatedIdentifier relatedIdentifierType="URL" relationType="IsSupplementTo">https://github.com/SequenceComparison/SupplementaryMaterial/tree/v1.0</relatedIdentifier>

				<relatedIdentifier relatedIdentifierType="DOI" relationType="IsPartOf">10.1038/nmeth.1502</relatedIdentifier>

			</relatedIdentifiers>

			<rightsList>

				<rights rightsURI="https://creativecommons.org/licenses/by/4.0/">Creative Commons Attribution 4.0</rights>

				<rights rightsURI="info:eu-repo/semantics/openAccess">Open Access</rights>

			</rightsList>

			<descriptions>

				<description descriptionType="Abstract"><p>This is the official version of the supplementary material which was used as the basis for the study on &quot;Sequence Comparison in Historical Linguistics&quot; (List, Dusseldorf, Dusseldorf University Press).</p></description>

			</descriptions>

		</resource>

		<dr:CobjCategory>0021</dr:CobjCategory>

		<oaf:dateAccepted>2014-01-01</oaf:dateAccepted>

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

		<oaf:language>und</oaf:language>

		<oaf:projectid>corda_______::609823</oaf:projectid>

		<oaf:hostedBy id="re3data_____::r3d100010468" name="Zenodo"/>

		<oaf:collectedFrom id="re3data_____::r3d100010468" name="Zenodo"/>

	</metadata>

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

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

	       xmlns:oai="http://www.openarchives.org/OAI/2.0/" xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance">

		<provenance xmlns="http://www.openarchives.org/OAI/2.0/provenance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/provenance http://www.openarchives.org/OAI/2.0/provenance.xsd">

			<originDescription altered="true" harvestDate="2018-04-30T02:17:56.67Z">

				<baseURL>https://zenodo.org/oai2d</baseURL>

				<identifier>oai:zenodo.org:11877</identifier>

				<datestamp>2017-05-30T02:33:14Z</datestamp>

				<metadataNamespace/>