D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 13 Projects 2 Releases 2 Wiki Activity

Merge branch 'master' of https://code-repo.d4science.org/D-Net/dnet-hadoop into orcid-no-doi

This commit is contained in:
Enrico Ottonello 2020-12-02 23:21:27 +01:00
parent 1b1e9ea67c 6ba8037cc7
commit 53b22c1937
14 changed files with 623 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
dhp-schemas/src/main/java/eu/dnetlib/dhp/schema/common/ModelConstants.java
View File

@ -7,6 +7,10 @@ import eu.dnetlib.dhp.schema.oaf.Qualifier;
public class ModelConstants {
public static final String ORCID = "orcid";
public static final String ORCID_PENDING = "orcid_pending";
public static final String ORCID_CLASSNAME = "Open Researcher and Contributor ID";
public static final String DNET_SUBJECT_TYPOLOGIES = "dnet:subject_classification_typologies";
public static final String DNET_RESULT_TYPOLOGIES = "dnet:result_typologies";
public static final String DNET_PUBLICATION_RESOURCE = "dnet:publication_resource";

2
dhp-workflows/dhp-doiboost/src/test/java/eu/dnetlib/doiboost/crossref/CrossrefMappingTest.scala
View File

@ -158,7 +158,7 @@ class CrossrefMappingTest {
rels.foreach(s => logger.info(s.getTarget))
assertEquals(rels.size, 3 )
assertEquals(rels.size, 6 )
}

2
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/PropagationConstant.java
View File

@ -44,8 +44,6 @@ public class PropagationConstant {
public static final String PROPAGATION_ORCID_TO_RESULT_FROM_SEM_REL_CLASS_ID = "authorpid:result";
public static final String PROPAGATION_ORCID_TO_RESULT_FROM_SEM_REL_CLASS_NAME = "Propagation of authors pid to result through semantic relations";
public static final String PROPAGATION_AUTHOR_PID = "ORCID";
public static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
private static final String cfHbforResultQuery = "select distinct r.id, inst.collectedfrom.key cf, inst.hostedby.key hb "

4
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/orcidtoresultfromsemrel/PrepareResultOrcidAssociationStep1.java
View File

@ -22,6 +22,7 @@ import com.fasterxml.jackson.databind.ObjectMapper;
import com.google.gson.Gson;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Relation;
import eu.dnetlib.dhp.schema.oaf.Result;
@ -102,7 +103,8 @@ public class PrepareResultOrcidAssociationStep1 {
+ " FROM result "
+ " LATERAL VIEW EXPLODE (author) a AS MyT "
+ " LATERAL VIEW EXPLODE (MyT.pid) p AS MyP "
+ " WHERE lower(MyP.qualifier.classid) = 'orcid') tmp "
+ " WHERE lower(MyP.qualifier.classid) = '" + ModelConstants.ORCID + "' or "
+ " lower(MyP.qalifier.classid) = '" + ModelConstants.ORCID_PENDING + "') tmp "
+ " GROUP BY id) r_t "
+ " JOIN ("
+ " SELECT source, target "

6
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/orcidtoresultfromsemrel/SparkOrcidToResultFromSemRelJob.java
View File

@ -23,6 +23,7 @@ import com.google.common.collect.Lists;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.PacePerson;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Author;
import eu.dnetlib.dhp.schema.oaf.Result;
import eu.dnetlib.dhp.schema.oaf.StructuredProperty;
@ -176,7 +177,7 @@ public class SparkOrcidToResultFromSemRelJob {
if (toaddpid) {
StructuredProperty p = new StructuredProperty();
p.setValue(autoritative_author.getOrcid());
p.setQualifier(getQualifier(PROPAGATION_AUTHOR_PID, PROPAGATION_AUTHOR_PID));
p.setQualifier(getQualifier(ModelConstants.ORCID_PENDING, ModelConstants.ORCID_CLASSNAME));
p
.setDataInfo(
getDataInfo(
@ -201,7 +202,8 @@ public class SparkOrcidToResultFromSemRelJob {
return false;
}
for (StructuredProperty pid : pids.get()) {
if (PROPAGATION_AUTHOR_PID.equals(pid.getQualifier().getClassid())) {
if (ModelConstants.ORCID_PENDING.equals(pid.getQualifier().getClassid().toLowerCase()) ||
ModelConstants.ORCID.equals(pid.getQualifier().getClassid().toLowerCase())) {
return true;
}
}

6
dhp-workflows/dhp-enrichment/src/test/java/eu/dnetlib/dhp/orcidtoresultfromsemrel/OrcidPropagationJobTest.java
View File

@ -19,8 +19,11 @@ import org.junit.jupiter.api.Test;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.cloudera.org.codehaus.jackson.map.jsontype.impl.ClassNameIdResolver;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.PropagationConstant;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.Dataset;
public class OrcidPropagationJobTest {
@ -166,7 +169,8 @@ public class OrcidPropagationJobTest {
propagatedAuthors
.filter(
"id = '50|dedup_wf_001::95b033c0c3961f6a1cdcd41a99a9632e' "
+ "and name = 'Vajinder' and surname = 'Kumar' and pidType = 'ORCID'")
+ "and name = 'Vajinder' and surname = 'Kumar' and pidType = '" +
ModelConstants.ORCID_PENDING + "'")
.count());
Assertions.assertEquals(1, propagatedAuthors.filter("pid = '0000-0002-8825-3517'").count());

28
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctions.java
View File

@ -189,6 +189,16 @@ public class CleaningFunctions {
author.setRank(i++);
}
}
final Set<String> collectedFrom = Optional
.ofNullable(r.getCollectedfrom())
.map(
c -> c
.stream()
.map(KeyValue::getKey)
.collect(Collectors.toCollection(HashSet::new)))
.orElse(new HashSet<>());
for (Author a : r.getAuthor()) {
if (Objects.isNull(a.getPid())) {
a.setPid(Lists.newArrayList());
@ -201,13 +211,29 @@ public class CleaningFunctions {
.filter(p -> Objects.nonNull(p.getQualifier()))
.filter(p -> StringUtils.isNotBlank(p.getValue()))
.map(p -> {
// hack to distinguish orcid from orcid_pending
String pidProvenance = Optional
.ofNullable(p.getDataInfo())
.map(
d -> Optional
.ofNullable(d.getProvenanceaction())
.map(Qualifier::getClassid)
.orElse(""))
.orElse("");
if (pidProvenance.equals(ModelConstants.SYSIMPORT_CROSSWALK_ENTITYREGISTRY)) {
p.getQualifier().setClassid(ModelConstants.ORCID);
} else {
p.getQualifier().setClassid(ModelConstants.ORCID_PENDING);
}
p.setValue(p.getValue().trim().replaceAll(ORCID_PREFIX_REGEX, ""));
return p;
})
.collect(
Collectors
.toMap(
StructuredProperty::getValue, Function.identity(), (p1, p2) -> p1,
p -> p.getQualifier().getClassid() + p.getValue(),
Function.identity(),
(p1, p2) -> p1,
LinkedHashMap::new))
.values()
.stream()

17
dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java
View File

@ -358,6 +358,23 @@ public class MappersTest {
System.out.println(p.getTitle().get(0).getValue());
}
@Test
void testBologna() throws IOException {
final String xml = IOUtils.toString(getClass().getResourceAsStream("oaf-bologna.xml"));
final List<Oaf> list = new OafToOafMapper(vocs, false).processMdRecord(xml);
System.out.println("***************");
System.out.println(new ObjectMapper().writeValueAsString(list));
System.out.println("***************");
final Publication p = (Publication) list.get(0);
assertValidId(p.getId());
assertValidId(p.getCollectedfrom().get(0).getKey());
System.out.println(p.getTitle().get(0).getValue());
assertTrue(StringUtils.isNotBlank(p.getTitle().get(0).getValue()));
System.out.println(p.getTitle().get(0).getValue());
}
private void assertValidId(final String id) {
assertEquals(49, id.length());
assertEquals('|', id.charAt(2));

22
dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/clean/result.json
View File

@ -49,6 +49,28 @@
"schemename": "dnet:pid_types"
},
"value": "https://orcid.org/0000-0001-9613-6639"
},
{
"dataInfo": {
"deletedbyinference": false,
"inferenceprovenance": "",
"inferred": false,
"invisible": false,
"provenanceaction": {
"classid": "sysimport:crosswalk:entityregistry",
"classname": "sysimport:crosswalk:entityregistry",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
},
"trust": "0.9"
},
"qualifier": {
"classid": "orcid",
"classname": "ORCID12",
"schemeid": "dnet:pid_types",
"schemename": "dnet:pid_types"
},
"value": "0000-0001-9613-6639"
}
],
"rank": 1,

1
dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/clean/terms.txt
View File

@ -1031,6 +1031,7 @@ dnet:pid_types @=@ dnet:pid_types @=@ jprn @=@ JPRN Identifier
dnet:pid_types @=@ dnet:pid_types @=@ mag_id @=@ Microsoft Academic Graph Identifier
dnet:pid_types @=@ dnet:pid_types @=@ oai @=@ Open Archives Initiative
dnet:pid_types @=@ dnet:pid_types @=@ orcid @=@ Open Researcher and Contributor ID
dnet:pid_types @=@ dnet:pid_types @=@ orcid_pending @=@ Open Researcher and Contributor ID
dnet:pid_types @=@ dnet:pid_types @=@ PANGAEA @=@ PANGAEA
dnet:pid_types @=@ dnet:pid_types @=@ epo_nr_epodoc @=@ Patent application number in EPODOC format
dnet:pid_types @=@ dnet:pid_types @=@ UNKNOWN @=@ UNKNOWN

115
dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/raw/oaf-bologna.xml Normal file
View File

@ -0,0 +1,115 @@
<?xml version="1.0" encoding="UTF-8"?>
<record xmlns:oaf="http://namespace.openaire.eu/oaf"
xmlns:oai="http://www.openarchives.org/OAI/2.0/" xmlns:xs="http://www.w3.org/2001/XMLSchema"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:dri="http://www.driver-repository.eu/namespace/dri"
xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance"
xmlns:dr="http://www.driver-repository.eu/namespace/dr"
xmlns:dc="http://purl.org/dc/elements/1.1/">
<header xmlns="http://namespace.openaire.eu/">
<dri:objIdentifier>od_________3::4f7038e665fdd3800297735f087a530c</dri:objIdentifier>
<dri:recordIdentifier>oai:amsacta.unibo.it:6382</dri:recordIdentifier>
<dri:dateOfCollection/>
<dri:mdFormat/>
<dri:mdFormatInterpretation/>
<dri:repositoryId/>
<dr:objectIdentifier/>
<dr:dateOfCollection>2020-05-21T05:26:15.93Z</dr:dateOfCollection>
<dr:dateOfTransformation>2020-08-01T11:06:26.977Z</dr:dateOfTransformation>
<oaf:datasourceprefix>od_________3</oaf:datasourceprefix>
</header>
<metadata xmlns="http://namespace.openaire.eu/">
<dc:title>CONSOLE Project - Deliverable 5.1 - "Guidelines for Community of Practice (CoP)
management at local level"</dc:title>
<dc:creator>Blanco-Velázquez, Francisco José</dc:creator>
<dc:creator>Runge, Tania</dc:creator>
<dc:creator>Anaya-Romero, María</dc:creator>
<dc:date>2020-05-11</dc:date>
<dc:description>The Community of Practice (CoP) is foreseen to play a key role in boosting
innovation in the effective and long-lasting delivery of agri-environmental-climate
public goods (AECPGs). The CONSOLE CoP will be organized around practitioners
experienced in the provision of AECPGs and those interested in it and will be nourished
throughout the project lifetime. In line with the definition of Wenger1 the CoP is
defined as a group of people (the community) who share a common interest and who learn
how to perform better through regular interaction and exchange of experiences. The idea
is to set up a pan-European CoP with national and/or local (regional) sub-groups managed
by the CONSOLE partners with the aim of developing improved and novel contract solutions
in collaboration with its members. This document sets out: (a) the purpose and
objectives of the CoP in CONSOLE, (b) the setting up and management of the CoP at
European, national and local level, (c) the process for motivating individuals to
participate. The CONSOLE CoP is intended to facilitate knowledge exchange and mutual
learning, mainly through virtual contacts. Participation in the CoP is based on sharing
and reciprocity principle. A core objective of these guidelines is to ensure a sound
management and facilitation of the CoP by all CONSOLE partners in view of optimizing the
input from CoP members in the project activities. Members within a national or local
CONSOLE CoP are: 1) CONSOLE partners; 2) practitioners, mainly farmers and foresters,
who test and implement practically the contractual models, and 3) experts, that may have
punctual interventions in the CoP. A vibrant CoP with active involvement of its members
is crucial for the assessment and testing of improved and novel voluntary measures for
the delivery of AECPGs. For each of the CONSOLE countries one national contact person is
nominated to take over the role as national focal point for the CoP activities in his
country and to serve as facilitator of the CoP. These facilitators are responsible to
ensure participation along the various project tasks foreseen within several WPs and to
overcome potential language barriers. The national contact person may be supported by
other CONSOLE partners from his country for local activities. At local level the CoP
benefits from existing contacts of CONSOLE partners to practitioners, including the
experts interviewed for the case studies analysis within WP2. The forming/development of
a CoP requires promoting exchanges taking on board the interests and needs of the actors
involved and to animate them to share their expertise. Collaborative learning within the
CoP supported by dedicated training activities will be crucial to ensure the intended
major transition towards smarter AECPGs-related practices in Europe. These guidelines
focus on the identification of the various tasks where CoP participation is foreseen at
local (regional) level and to provide support for the identification of potential
members. In the deliverable D5.2 “Guidelines for testing the solutions catalogue by CoP
and partners” further details about the involvement of the CoP will be
provided.</dc:description>
<dc:format>application/pdf</dc:format>
<dc:identifier>http://amsacta.unibo.it/6382/</dc:identifier>
<dc:language>eng</dc:language>
<dc:relation>info:eu-repo/grantAgreement/EC/H2020/817949/EU/CONtract SOLutions for Effective
and lasting delivery of agri-environmental-climate public goods by EU agriculture and
forestry/CONSOLE</dc:relation>
<dc:relation>info:eu-repo/semantics/altIdentifier/doi/10.6092/unibo/amsacta/6382</dc:relation>
<dc:relation>info:eu-repo/semantics/reference/url/https://console-project.eu/</dc:relation>
<dc:source>Blanco-Velázquez, Francisco José ; Runge, Tania ; Anaya-Romero, María (2020)
CONSOLE Project - Deliverable 5.1 - "Guidelines for Community of Practice (CoP)
management at local level". p. 21. DOI 10.6092/unibo/amsacta/6382
&lt;http://doi.org/10.6092/unibo/amsacta/6382&gt;.</dc:source>
<dc:subject classid="keyword" classname="keyword" schemeid="dnet:result_subject"
schemename="dnet:result_subject">AGR/01 Economia ed estimo rurale</dc:subject>
<dc:subject classid="keyword" classname="keyword" schemeid="dnet:result_subject"
schemename="dnet:result_subject">SECS-S/01 Statistica</dc:subject>
<dc:type>info:eu-repo/semantics/book</dc:type>
<dc:type>info:eu-repo/semantics/publishedVersion</dc:type>
<dr:CobjCategory type="publication">0002</dr:CobjCategory>
<oaf:dateAccepted>2020-05-11</oaf:dateAccepted>
<oaf:projectid>corda__h2020::817949</oaf:projectid>
<oaf:embargoenddate/>
<oaf:collectedDatasourceid>opendoar____::3</oaf:collectedDatasourceid>
<oaf:accessrights>OPEN</oaf:accessrights>
<oaf:hostedBy name="AMS Acta" id="opendoar____::3"/>
<oaf:collectedFrom name="AMS Acta" id="opendoar____::3"/>
<oaf:identifier identifierType="doi">10.6092/unibo/amsacta/6382</oaf:identifier>
<oaf:journal issn="" eissn="" vol="" iss="" sp="" ep=""/>
</metadata>
<about>
<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 harvestDate="2020-05-21T05:26:15.93Z" altered="true">
<baseURL>http%3A%2F%2Famsacta.unibo.it%2Fcgi%2Fopenaire3</baseURL>
<identifier>oai:amsacta.unibo.it:6382</identifier>
<datestamp>2020-05-13T09:27:00Z</datestamp>
<metadataNamespace>http://www.openarchives.org/OAI/2.0/oai_dc/</metadataNamespace>
</originDescription>
</provenance>
<oaf:datainfo>
<oaf:inferred>false</oaf:inferred>
<oaf:deletedbyinference>false</oaf:deletedbyinference>
<oaf:trust>0.9</oaf:trust>
<oaf:inferenceprovenance/>
<oaf:provenanceaction schemename="dnet:provenanceActions"
schemeid="dnet:provenanceActions" classname="sysimport:crosswalk:repository"
classid="sysimport:crosswalk:repository"/>
</oaf:datainfo>
</about>
</record>

24
dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/raw/textgrid.xml
View File

@ -1,23 +1,24 @@
<?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:oai="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<record xmlns:datacite="http://datacite.org/schema/kernel-3"
xmlns:dr="http://www.driver-repository.eu/namespace/dr"
xmlns:oaf="http://namespace.openaire.eu/oaf" xmlns:oai="http://www.openarchives.org/OAI/2.0/">
<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:prov="http://www.openarchives.org/OAI/2.0/provenance">
xmlns:dri="http://www.driver-repository.eu/namespace/dri"
xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<dri:objIdentifier>r3f52792889d::000051aa1f61d77d2c0b340091f8024e</dri:objIdentifier>
<dri:recordIdentifier>textgrid:q9cv.0</dri:recordIdentifier>
<dri:dateOfCollection>2020-11-17T09:34:11.128+01:00</dri:dateOfCollection>
<oaf:datasourceprefix>r3f52792889d</oaf:datasourceprefix>
<identifier xmlns="http://www.openarchives.org/OAI/2.0/">textgrid:q9cv.0</identifier>
<datestamp xmlns="http://www.openarchives.org/OAI/2.0/">2012-01-21T13:35:20Z</datestamp>
<dr:dateOfTransformation>2020-11-17T09:46:21.551+01:00</dr:dateOfTransformation>
<dr:dateOfTransformation>2020-11-17T19:08:56.703+01:00</dr:dateOfTransformation>
</oai:header>
<metadata>
<datacite:resource xmlns="http://www.openarchives.org/OAI/2.0/"
xmlns:datacite="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:prov="http://www.openarchives.org/OAI/2.0/provenance">
xmlns:dri="http://www.driver-repository.eu/namespace/dri"
xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<datacite:identifier identifierType="Handle">hdl:11858/00-1734-0000-0003-7664-F</datacite:identifier>
<datacite:creators>
<datacite:creator>
@ -46,8 +47,8 @@
<datacite:date dateType="Updated">2012-01-21T13:35:20Z</datacite:date>
</datacite:dates>
<datacite:resourceType resourceTypeGeneral="Dataset"/>
<alternateIdentifiers>
<datacite:alternateIdentifier alternateIdentifierType="URI">textgrid:q9cv.0</datacite:alternateIdentifier>
<alternateIdentifiers xmlns="http://datacite.org/schema/kernel-3">
<datacite:alternateIdentifier alternateIdentifierType="URI" xmlns="http://www.openarchives.org/OAI/2.0/">textgrid:q9cv.0</datacite:alternateIdentifier>
<alternateIdentifier alternateIdentifierType="URL">http://hdl.handle.net/hdl:11858/00-1734-0000-0003-7664-F</alternateIdentifier>
</alternateIdentifiers>
<datacite:relatedIdentifiers>
@ -83,7 +84,7 @@
<oaf:identifier identifierType="handle">hdl:11858/00-1734-0000-0003-7664-F</oaf:identifier>
<dr:CobjCategory type="dataset">0021</dr:CobjCategory>
<oaf:refereed>0002</oaf:refereed>
<oaf:dateAccepted>2012-01-01</oaf:dateAccepted>
<oaf:dateAccepted>2012-01-21</oaf:dateAccepted>
<oaf:accessrights>OPEN</oaf:accessrights>
<oaf:license>http://creativecommons.org/licenses/by/3.0/de/legalcode</oaf:license>
<oaf:language>und</oaf:language>
@ -91,7 +92,8 @@
<oaf:collectedFrom id="re3data_____::r3d100011365" name="TextGrid Repository"/>
</metadata>
<about xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:dri="http://www.driver-repository.eu/namespace/dri" xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance">
xmlns:dri="http://www.driver-repository.eu/namespace/dri"
xmlns:prov="http://www.openarchives.org/OAI/2.0/provenance" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<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="2020-11-17T09:34:11.128+01:00">
<baseURL>https%3A%2F%2Fdev.textgridlab.org%2F1.0%2Ftgoaipmh%2Foai</baseURL>

33
dhp-workflows/dhp-graph-provision/src/test/java/eu/dnetlib/dhp/oa/provision/XmlRecordFactoryTest.java
View File

@ -5,23 +5,31 @@ import static org.junit.jupiter.api.Assertions.*;
import java.io.IOException;
import java.io.StringReader;
import java.util.List;
import org.apache.commons.io.IOUtils;
import org.apache.commons.lang3.StringUtils;
import org.dom4j.Document;
import org.dom4j.DocumentException;
import org.dom4j.io.SAXReader;
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.Disabled;
import org.junit.jupiter.api.Test;
import org.mockito.Mock;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.oa.provision.model.JoinedEntity;
import eu.dnetlib.dhp.oa.provision.utils.ContextMapper;
import eu.dnetlib.dhp.oa.provision.utils.XmlRecordFactory;
import eu.dnetlib.dhp.schema.oaf.Oaf;
import eu.dnetlib.dhp.schema.oaf.OafEntity;
import eu.dnetlib.dhp.schema.oaf.OafMapperUtils;
import eu.dnetlib.dhp.schema.oaf.Publication;
import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
//TODO to enable it we need to update the joined_entity.json test file
@Disabled
//@Disabled
public class XmlRecordFactoryTest {
private static final String otherDsTypeId = "scholarcomminfra,infospace,pubsrepository::mock,entityregistry,entityregistry::projects,entityregistry::repositories,websource";
@ -35,6 +43,27 @@ public class XmlRecordFactoryTest {
JoinedEntity je = new ObjectMapper().readValue(json, JoinedEntity.class);
assertNotNull(je);
Document doc = buildXml(je);
//// TODO specific test assertion on doc
}
@Test
void testBologna() throws IOException, DocumentException {
final String json = IOUtils.toString(getClass().getResourceAsStream("oaf-bologna.json"));
Publication oaf = new ObjectMapper().readValue(json, Publication.class);
assertNotNull(oaf);
JoinedEntity je = new JoinedEntity();
je.setEntity(oaf);
assertNotNull(je);
Document doc = buildXml(je);
// TODO specific test assertion on doc
System.out.println(doc.asXML());
}
private Document buildXml(JoinedEntity je) throws DocumentException {
ContextMapper contextMapper = new ContextMapper();
XmlRecordFactory xmlRecordFactory = new XmlRecordFactory(contextMapper, false, XmlConverterJob.schemaLocation,
@ -49,5 +78,7 @@ public class XmlRecordFactoryTest {
assertNotNull(doc);
// TODO add assertions based of values extracted from the XML record
return doc;
}
}

379
dhp-workflows/dhp-graph-provision/src/test/resources/eu/dnetlib/dhp/oa/provision/oaf-bologna.json Normal file
View File

@ -0,0 +1,379 @@
{
"collectedfrom": [
{
"key": "10|opendoar____::eccbc87e4b5ce2fe28308fd9f2a7baf3",
"value": "AMS Acta",
"dataInfo": null
}
],
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
},
"lastupdatetimestamp": 1606898557407,
"id": "50|od_________3::4f7038e665fdd3800297735f087a530c",
"originalId": [
"oai:amsacta.unibo.it:6382"
],
"pid": [
{
"value": "10.6092/unibo/amsacta/6382",
"qualifier": {
"classid": "doi",
"classname": "doi",
"schemeid": "dnet:pid_types",
"schemename": "dnet:pid_types"
},
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
}
],
"dateofcollection": "",
"dateoftransformation": "2020-08-01T11:06:26.977Z",
"extraInfo": [],
"oaiprovenance": {
"originDescription": {
"harvestDate": "2020-05-21T05:26:15.93Z",
"altered": true,
"baseURL": "http%3A%2F%2Famsacta.unibo.it%2Fcgi%2Fopenaire3",
"identifier": "oai:amsacta.unibo.it:6382",
"datestamp": "2020-05-13T09:27:00Z",
"metadataNamespace": "http://www.openarchives.org/OAI/2.0/oai_dc/"
}
},
"measures": null,
"author": [
{
"fullname": "Blanco-Velázquez, Francisco José",
"name": "Francisco José",
"surname": "Blanco-Velázquez",
"rank": 1,
"pid": [],
"affiliation": null
},
{
"fullname": "Runge, Tania",
"name": "Tania",
"surname": "Runge",
"rank": 2,
"pid": [],
"affiliation": null
},
{
"fullname": "Anaya-Romero, María",
"name": "María",
"surname": "Anaya-Romero",
"rank": 3,
"pid": [],
"affiliation": null
}
],
"resulttype": {
"classid": "publication",
"classname": "publication",
"schemeid": "dnet:result_typologies",
"schemename": "dnet:result_typologies"
},
"language": {
"classid": "eng",
"classname": "English",
"schemeid": "dnet:languages",
"schemename": "dnet:languages"
},
"country": [],
"subject": [
{
"value": "AGR/01 Economia ed estimo rurale",
"qualifier": {
"classid": "keyword",
"classname": "keyword",
"schemeid": "dnet:result_subject",
"schemename": "dnet:result_subject"
},
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
},
{
"value": "SECS-S/01 Statistica",
"qualifier": {
"classid": "keyword",
"classname": "keyword",
"schemeid": "dnet:result_subject",
"schemename": "dnet:result_subject"
},
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
}
],
"title": [
{
"value": "CONSOLE Project - Deliverable 5.1 - \"Guidelines for Community of Practice (CoP)\n management at local level\"",
"qualifier": {
"classid": "main title",
"classname": "main title",
"schemeid": "dnet:dataCite_title",
"schemename": "dnet:dataCite_title"
},
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
}
],
"relevantdate": [],
"description": [
{
"value": "The Community of Practice (CoP) is foreseen to play a key role in boosting\n innovation in the effective and long-lasting delivery of agri-environmental-climate\n public goods (AECPGs). The CONSOLE CoP will be organized around practitioners\n experienced in the provision of AECPGs and those interested in it and will be nourished\n throughout the project lifetime. In line with the definition of Wenger1 the CoP is\n defined as a group of people (the community) who share a common interest and who learn\n how to perform better through regular interaction and exchange of experiences. The idea\n is to set up a pan-European CoP with national and/or local (regional) sub-groups managed\n by the CONSOLE partners with the aim of developing improved and novel contract solutions\n in collaboration with its members. This document sets out: (a) the purpose and\n objectives of the CoP in CONSOLE, (b) the setting up and management of the CoP at\n European, national and local level, (c) the process for motivating individuals to\n participate. The CONSOLE CoP is intended to facilitate knowledge exchange and mutual\n learning, mainly through virtual contacts. Participation in the CoP is based on sharing\n and reciprocity principle. A core objective of these guidelines is to ensure a sound\n management and facilitation of the CoP by all CONSOLE partners in view of optimizing the\n input from CoP members in the project activities. Members within a national or local\n CONSOLE CoP are: 1) CONSOLE partners; 2) practitioners, mainly farmers and foresters,\n who test and implement practically the contractual models, and 3) experts, that may have\n punctual interventions in the CoP. A vibrant CoP with active involvement of its members\n is crucial for the assessment and testing of improved and novel voluntary measures for\n the delivery of AECPGs. For each of the CONSOLE countries one national contact person is\n nominated to take over the role as national focal point for the CoP activities in his\n country and to serve as facilitator of the CoP. These facilitators are responsible to\n ensure participation along the various project tasks foreseen within several WPs and to\n overcome potential language barriers. The national contact person may be supported by\n other CONSOLE partners from his country for local activities. At local level the CoP\n benefits from existing contacts of CONSOLE partners to practitioners, including the\n experts interviewed for the case studies analysis within WP2. The forming/development of\n a CoP requires promoting exchanges taking on board the interests and needs of the actors\n involved and to animate them to share their expertise. Collaborative learning within the\n CoP supported by dedicated training activities will be crucial to ensure the intended\n major transition towards smarter AECPGs-related practices in Europe. These guidelines\n focus on the identification of the various tasks where CoP participation is foreseen at\n local (regional) level and to provide support for the identification of potential\n members. In the deliverable D5.2 “Guidelines for testing the solutions catalogue by CoP\n and partners” further details about the involvement of the CoP will be\n provided.",
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
}
],
"dateofacceptance": {
"value": "2020-05-11",
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
},
"publisher": null,
"embargoenddate": null,
"source": [
{
"value": "Blanco-Velázquez, Francisco José ; Runge, Tania ; Anaya-Romero, María (2020)\n CONSOLE Project - Deliverable 5.1 - \"Guidelines for Community of Practice (CoP)\n management at local level\". p. 21. DOI 10.6092/unibo/amsacta/6382\n <http://doi.org/10.6092/unibo/amsacta/6382>.",
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
}
],
"fulltext": [],
"format": [
{
"value": "application/pdf",
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
}
],
"contributor": [],
"resourcetype": null,
"coverage": [],
"bestaccessright": {
"classid": "OPEN",
"classname": "Open Access",
"schemeid": "dnet:access_modes",
"schemename": "dnet:access_modes"
},
"context": [],
"externalReference": [],
"instance": [
{
"license": null,
"accessright": {
"classid": "OPEN",
"classname": "Open Access",
"schemeid": "dnet:access_modes",
"schemename": "dnet:access_modes"
},
"instancetype": {
"classid": "0002",
"classname": "Book",
"schemeid": "dnet:publication_resource",
"schemename": "dnet:publication_resource"
},
"hostedby": {
"key": "10|opendoar____::eccbc87e4b5ce2fe28308fd9f2a7baf3",
"value": "AMS Acta",
"dataInfo": null
},
"url": [
"http://amsacta.unibo.it/6382/"
],
"distributionlocation": "",
"collectedfrom": {
"key": "10|opendoar____::eccbc87e4b5ce2fe28308fd9f2a7baf3",
"value": "AMS Acta",
"dataInfo": null
},
"dateofacceptance": {
"value": "2020-05-11",
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
}
},
"processingchargeamount": null,
"processingchargecurrency": null,
"refereed": {
"classid": "UNKNOWN",
"classname": "Unknown",
"schemeid": "dnet:review_levels",
"schemename": "dnet:review_levels"
}
}
],
"journal": null
},
{
"collectedfrom": [
{
"key": "10|opendoar____::eccbc87e4b5ce2fe28308fd9f2a7baf3",
"value": "AMS Acta",
"dataInfo": null
}
],
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
},
"lastupdatetimestamp": 1606898557407,
"relType": "resultProject",
"subRelType": "outcome",
"relClass": "isProducedBy",
"source": "50|od_________3::4f7038e665fdd3800297735f087a530c",
"target": "40|corda__h2020::79a0e16c122c9a18eb60e4a5e64b620d",
"validated": null,
"validationDate": null,
"properties": []
},
{
"collectedfrom": [
{
"key": "10|opendoar____::eccbc87e4b5ce2fe28308fd9f2a7baf3",
"value": "AMS Acta",
"dataInfo": null
}
],
"dataInfo": {
"invisible": false,
"inferred": false,
"deletedbyinference": false,
"trust": "0.9",
"inferenceprovenance": "",
"provenanceaction": {
"classid": "sysimport:crosswalk:repository",
"classname": "sysimport:crosswalk:repository",
"schemeid": "dnet:provenanceActions",
"schemename": "dnet:provenanceActions"
}
},
"lastupdatetimestamp": 1606898557407,
"relType": "resultProject",
"subRelType": "outcome",
"relClass": "produces",
"source": "40|corda__h2020::79a0e16c122c9a18eb60e4a5e64b620d",
"target": "50|od_________3::4f7038e665fdd3800297735f087a530c",
"validated": null,
"validationDate": null,
"properties": []
}