forked from D-Net/dnet-hadoop
[cleaning] subjects declared as belonging to specific vocabularies whose values are not found in the vocab are set to type keyword
This commit is contained in:
parent
6163ecbf63
commit
b47aaf4dd1
|
@ -211,7 +211,7 @@ public class GraphCleaningFunctions extends CleaningFunctions {
|
|||
.orElse(s.getValue()),
|
||||
Function.identity(),
|
||||
(s1, s2) -> Collections
|
||||
.min(Lists.newArrayList(s1, s1), new SubjectProvenanceComparator())))
|
||||
.min(Lists.newArrayList(s1, s2), new SubjectProvenanceComparator())))
|
||||
.values());
|
||||
r.setSubject(subjects);
|
||||
}
|
||||
|
|
|
@ -3,6 +3,7 @@ package eu.dnetlib.dhp.oa.graph.clean;
|
|||
|
||||
import java.io.Serializable;
|
||||
import java.util.HashMap;
|
||||
import java.util.Objects;
|
||||
import java.util.concurrent.atomic.AtomicReference;
|
||||
|
||||
import org.apache.commons.lang3.SerializationUtils;
|
||||
|
@ -10,6 +11,7 @@ import org.apache.commons.lang3.StringUtils;
|
|||
|
||||
import eu.dnetlib.dhp.common.FunctionalInterfaceSupport.SerializableConsumer;
|
||||
import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
|
||||
import eu.dnetlib.dhp.common.vocabulary.VocabularyTerm;
|
||||
import eu.dnetlib.dhp.schema.common.ModelConstants;
|
||||
import eu.dnetlib.dhp.schema.oaf.*;
|
||||
|
||||
|
@ -31,29 +33,30 @@ public class CleaningRuleMap extends HashMap<Class<?>, SerializableConsumer<Obje
|
|||
}
|
||||
|
||||
private static void cleanSubject(VocabularyGroup vocabularies, Subject subject) {
|
||||
if (cleanSubjectForVocabulary(ModelConstants.DNET_SUBJECT_FOS_CLASSID, vocabularies, subject)) {
|
||||
return;
|
||||
} else {
|
||||
cleanSubjectForVocabulary(ModelConstants.DNET_SUBJECT_FOS_CLASSID, vocabularies, subject);
|
||||
// TODO cleaning based on different subject vocabs can be added here
|
||||
}
|
||||
}
|
||||
|
||||
private static boolean cleanSubjectForVocabulary(String vocabularyId, VocabularyGroup vocabularies,
|
||||
private static void cleanSubjectForVocabulary(String vocabularyId, VocabularyGroup vocabularies,
|
||||
Subject subject) {
|
||||
AtomicReference<Boolean> modified = new AtomicReference<>(false);
|
||||
|
||||
vocabularies.find(vocabularyId).ifPresent(vocabulary -> {
|
||||
if (!ModelConstants.DNET_SUBJECT_KEYWORD.equalsIgnoreCase(subject.getQualifier().getClassid())) {
|
||||
return;
|
||||
}
|
||||
if (ModelConstants.DNET_SUBJECT_KEYWORD.equalsIgnoreCase(subject.getQualifier().getClassid())) {
|
||||
Qualifier newValue = vocabulary.lookup(subject.getValue());
|
||||
if (!ModelConstants.UNKNOWN.equals(newValue.getClassid())) {
|
||||
subject.setValue(newValue.getClassid());
|
||||
subject.getQualifier().setClassid(vocabularyId);
|
||||
subject.getQualifier().setClassname(vocabulary.getName());
|
||||
modified.set(true);
|
||||
}
|
||||
} else if (vocabularyId.equals(subject.getQualifier().getClassid())) {
|
||||
Qualifier syn = vocabulary.getSynonymAsQualifier(subject.getValue());
|
||||
VocabularyTerm term = vocabulary.getTerm(subject.getValue());
|
||||
if (Objects.isNull(syn) && Objects.isNull(term)) {
|
||||
subject.getQualifier().setClassid(ModelConstants.DNET_SUBJECT_KEYWORD);
|
||||
subject.getQualifier().setClassname(ModelConstants.DNET_SUBJECT_KEYWORD);
|
||||
}
|
||||
}
|
||||
});
|
||||
return modified.get();
|
||||
}
|
||||
|
||||
private static void cleanRelation(VocabularyGroup vocabularies, Relation r) {
|
||||
|
|
|
@ -43,7 +43,7 @@ public class CleanCountrySparkJob implements Serializable {
|
|||
|
||||
String jsonConfiguration = IOUtils
|
||||
.toString(
|
||||
CleanContextSparkJob.class
|
||||
CleanCountrySparkJob.class
|
||||
.getResourceAsStream(
|
||||
"/eu/dnetlib/dhp/oa/graph/input_clean_country_parameters.json"));
|
||||
final ArgumentApplicationParser parser = new ArgumentApplicationParser(jsonConfiguration);
|
||||
|
|
|
@ -278,6 +278,16 @@ public class GraphCleaningFunctionsTest {
|
|||
s -> "0102 computer and information sciences".equals(s.getValue()) &
|
||||
ModelConstants.DNET_SUBJECT_FOS_CLASSID.equals(s.getQualifier().getClassid())));
|
||||
|
||||
List<Subject> s1 = p_cleaned
|
||||
.getSubject()
|
||||
.stream()
|
||||
.filter(s -> s.getValue().equals("In Situ Hybridization"))
|
||||
.collect(Collectors.toList());
|
||||
assertNotNull(s1);
|
||||
assertEquals(1, s1.size());
|
||||
assertEquals(ModelConstants.DNET_SUBJECT_KEYWORD, s1.get(0).getQualifier().getClassid());
|
||||
assertEquals(ModelConstants.DNET_SUBJECT_KEYWORD, s1.get(0).getQualifier().getClassname());
|
||||
|
||||
// TODO add more assertions to verity the cleaned values
|
||||
System.out.println(MAPPER.writeValueAsString(p_cleaned));
|
||||
}
|
||||
|
|
|
@ -706,6 +706,28 @@
|
|||
"source": [
|
||||
],
|
||||
"subject": [
|
||||
{
|
||||
"dataInfo": {
|
||||
"provenanceaction": {
|
||||
"classid": "sysimport:crosswalk:repository",
|
||||
"classname": "sysimport:crosswalk:repository",
|
||||
"schemeid": "dnet:provenanceActions",
|
||||
"schemename": "dnet:provenanceActions"
|
||||
},
|
||||
"deletedbyinference": false,
|
||||
"inferred": false,
|
||||
"inferenceprovenance": "",
|
||||
"invisible": false,
|
||||
"trust": "0.9"
|
||||
},
|
||||
"qualifier": {
|
||||
"classid": "FOS",
|
||||
"classname": "Fields of Science and Technology classification",
|
||||
"schemeid": "dnet:result_subject",
|
||||
"schemename": "dnet:result_subject"
|
||||
},
|
||||
"value": "In Situ Hybridization"
|
||||
},
|
||||
{
|
||||
"dataInfo": {
|
||||
"deletedbyinference": false,
|
||||
|
|
Loading…
Reference in New Issue