D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 12 Projects 2 Releases 2 Wiki Activity

filter out authors pid (only orcid). Added check to get unique provenance for context id. filtr out countries with code UNKNOWN

This commit is contained in:
Miriam Baglioni 2020-08-07 17:48:18 +02:00
parent 5ceb8c5f0a
commit 1b3ad1bce6
1 changed files with 119 additions and 38 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

127
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/dump/ResultMapper.java
View File

@ -5,6 +5,8 @@ import java.io.Serializable;
import java.util.*;
import java.util.stream.Collectors;
import org.apache.commons.lang3.StringUtils;
import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.dump.oaf.*;
import eu.dnetlib.dhp.schema.dump.oaf.community.CommunityResult;
@ -125,10 +127,6 @@ public class ResultMapper implements Serializable {
break;
}
Optional
.ofNullable(input.getFulltext())
.ifPresent(fts -> out.setFulltext(fts.stream().map(ft -> ft.getValue()).collect(Collectors.toList())));
Optional
.ofNullable(input.getAuthor())
.ifPresent(ats -> out.setAuthor(ats.stream().map(at -> getAuthor(at)).collect(Collectors.toList())));
@ -155,14 +153,19 @@ public class ResultMapper implements Serializable {
.ifPresent(value -> value.stream().forEach(c -> contributorList.add(c.getValue())));
out.setContributor(contributorList);
List<Country> countryList = new ArrayList<>();
// List<Country> countryList = new ArrayList<>();
Optional
.ofNullable(input.getCountry())
.ifPresent(
value -> value
value -> out
.setCountry(
value
.stream()
.forEach(
.map(
c -> {
if (c.getClassid().equals((ModelConstants.UNKNOWN))) {
return null;
}
Country country = new Country();
country.setCode(c.getClassid());
country.setLabel(c.getClassname());
@ -175,13 +178,14 @@ public class ResultMapper implements Serializable {
.newInstance(
provenance
.getProvenanceaction()
.getClassname(),
.getClassid(),
c.getDataInfo().getTrust())));
countryList
.add(country);
}));
return country;
})
.filter(Objects::nonNull)
.collect(Collectors.toList())));
out.setCountry(countryList);
// out.setCountry(countryList);
final List<String> coverageList = new ArrayList<>();
Optional
@ -380,7 +384,8 @@ public class ResultMapper implements Serializable {
.orElse(null))
.filter(Objects::nonNull)
.collect(Collectors.toSet()));
context.setProvenance(provenance);
context.setProvenance(getUniqueProvenance(provenance));
}
return context;
}
@ -406,6 +411,50 @@ public class ResultMapper implements Serializable {
}
private static List<Provenance> getUniqueProvenance(List<Provenance> provenance) {
Provenance iProv = new Provenance();
// iProv.setProvenance(Constants.INFERRED);
Provenance hProv = new Provenance();
// hProv.setProvenance(Constants.HARVESTED);
Provenance lProv = new Provenance();
for (Provenance p : provenance) {
switch (p.getProvenance()) {
case Constants.HARVESTED:
hProv = getHighestTrust(hProv, p);
break;
case Constants.INFERRED:
iProv = getHighestTrust(iProv, p);
// To be removed as soon as the new beta run has been done
// this fixex issue of not set trust during bulktagging
if (StringUtils.isEmpty(iProv.getTrust())) {
iProv.setTrust(Constants.DEFAULT_TRUST);
}
break;
case Constants.USER_CLAIM:
lProv = getHighestTrust(lProv, p);
break;
}
}
return Arrays
.asList(iProv, hProv, lProv)
.stream()
.filter(p -> !StringUtils.isEmpty(p.getProvenance()))
.collect(Collectors.toList());
}
private static Provenance getHighestTrust(Provenance hProv, Provenance p) {
if (StringUtils.isNoneEmpty(hProv.getTrust(), p.getTrust()))
return hProv.getTrust().compareTo(p.getTrust()) > 0 ? hProv : p;
return (StringUtils.isEmpty(p.getTrust()) && !StringUtils.isEmpty(hProv.getTrust())) ? hProv : p;
}
private static Subject getSubject(StructuredProperty s) {
Subject subject = new Subject();
subject.setSubject(ControlledField.newInstance(s.getQualifier().getClassid(), s.getValue()));
@ -435,21 +484,53 @@ public class ResultMapper implements Serializable {
a.setName(oa.getName());
a.setSurname(oa.getSurname());
a.setRank(oa.getRank());
Optional
.ofNullable(oa.getPid())
.ifPresent(
value -> a
.setPids(
value
.stream()
.map(p -> getPid(p))
.collect(Collectors.toList())));
Optional<List<StructuredProperty>> oPids = Optional
.ofNullable(oa.getPid());
if (oPids.isPresent()) {
Pid pid = getOrcid(oPids.get());
if (pid != null) {
a.setPid(pid);
}
}
return a;
}
private static Pid getOrcid(List<StructuredProperty> p) {
for (StructuredProperty pid : p) {
if (pid.getQualifier().getClassid().equals(Constants.ORCID)) {
Optional<DataInfo> di = Optional.ofNullable(pid.getDataInfo());
if (di.isPresent()) {
return Pid
.newInstance(
ControlledField
.newInstance(
pid.getQualifier().getClassid(),
pid.getValue()),
Provenance
.newInstance(
di.get().getProvenanceaction().getClassname(),
di.get().getTrust()));
} else {
return Pid
.newInstance(
ControlledField
.newInstance(
pid.getQualifier().getClassid(),
pid.getValue())
);
}
}
}
return null;
}
private static Pid getPid(StructuredProperty p) {
Pid pid = new Pid();
pid.setPid(ControlledField.newInstance(p.getQualifier().getClassid(), p.getValue()));
pid.setId(ControlledField.newInstance(p.getQualifier().getClassid(), p.getValue()));
Optional<DataInfo> di = Optional.ofNullable(p.getDataInfo());
Provenance provenance = new Provenance();
if (di.isPresent()) {