forked from D-Net/dnet-hadoop
295 lines
10 KiB
Java
295 lines
10 KiB
Java
|
|
package eu.dnetlib.dhp.rdfconverter.utils;
|
|
|
|
import java.io.IOException;
|
|
import java.io.StringReader;
|
|
import java.io.StringWriter;
|
|
import java.util.*;
|
|
import java.util.stream.Collectors;
|
|
|
|
import org.eclipse.rdf4j.model.Model;
|
|
import org.eclipse.rdf4j.rio.RDFFormat;
|
|
import org.eclipse.rdf4j.rio.RDFWriter;
|
|
import org.eclipse.rdf4j.rio.Rio;
|
|
import org.slf4j.Logger;
|
|
import org.slf4j.LoggerFactory;
|
|
|
|
import com.fasterxml.jackson.databind.DeserializationFeature;
|
|
import com.fasterxml.jackson.databind.ObjectMapper;
|
|
import com.fasterxml.jackson.databind.module.SimpleModule;
|
|
import com.github.jsonldjava.core.JsonLdOptions;
|
|
import com.github.jsonldjava.core.JsonLdProcessor;
|
|
import com.github.jsonldjava.utils.JsonUtils;
|
|
|
|
import eu.dnetlib.dhp.rdfconverter.bioschema.model.BioSchemaProtein;
|
|
import eu.dnetlib.dhp.rdfconverter.bioschema.model.DataciteProtein;
|
|
|
|
public class RDFConverter {
|
|
|
|
private static final Logger log = LoggerFactory.getLogger(RDFConverter.class);
|
|
|
|
public ArrayList<String> nQuadsFile2DataciteJson(String nquads, String profile) throws Exception {
|
|
if (profile.equals("Protein")) {
|
|
return nQuadsFile2DataciteJson(nquads);
|
|
}
|
|
throw new RuntimeException("Profile not supported");
|
|
}
|
|
|
|
private ArrayList<String> nQuadsFile2DataciteJson(String nquads) throws Exception {
|
|
StringReader reader = new StringReader(nquads);
|
|
Model model = Rio.parse(reader, "", RDFFormat.NQUADS);
|
|
StringWriter jsonLDWriter = new StringWriter();
|
|
RDFWriter rdfRecordWriter = Rio.createWriter(RDFFormat.JSONLD, jsonLDWriter);
|
|
Rio.write(model, rdfRecordWriter);
|
|
String jsonLDBuffer = jsonLDWriter.toString();
|
|
Object jsonObject = JsonUtils.fromString(jsonLDBuffer);
|
|
Object compact = JsonLdProcessor.compact(jsonObject, new HashMap<>(), new JsonLdOptions());
|
|
String compactContent = JsonUtils.toString(compact);
|
|
log.debug("jsonld: " + compactContent);
|
|
|
|
ObjectMapper objectMapper = new ObjectMapper();
|
|
objectMapper.enable(DeserializationFeature.ACCEPT_SINGLE_VALUE_AS_ARRAY);
|
|
objectMapper.enable(DeserializationFeature.ACCEPT_EMPTY_STRING_AS_NULL_OBJECT);
|
|
objectMapper.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
|
|
objectMapper.configure(DeserializationFeature.FAIL_ON_INVALID_SUBTYPE, false);
|
|
objectMapper.configure(DeserializationFeature.FAIL_ON_MISSING_CREATOR_PROPERTIES, false);
|
|
BioSchemaProtein bioSchemaProtein = objectMapper.readValue(compactContent, BioSchemaProtein.class);
|
|
BioSchemaProtein.DateTimeType retrievedOnType = bioSchemaProtein.getRetrievedOn();
|
|
|
|
List<String> citations = bioSchemaProtein.getEntryList().stream().map(entry -> {
|
|
if (entry.getCitation() != null) {
|
|
BioSchemaProtein.Citation citationInfo = entry.getCitation();
|
|
return citationInfo.getId();
|
|
}
|
|
return null;
|
|
}).filter(id -> id != null).collect(Collectors.toList());
|
|
|
|
ArrayList<String> results = new ArrayList<String>();
|
|
final List<DataciteProtein> dataciteProteins = new ArrayList<>();
|
|
final Map<String, List<BioSchemaProtein.Link>> sequenceAnnotations = new HashMap<>();
|
|
final Map<String, String> propertyValues = new HashMap<>();
|
|
final Map<String, BioSchemaProtein.DefinedTerm> definedTerms = new HashMap<>();
|
|
final Map<String, BioSchemaProtein.Entry> bioschemaProteins = new HashMap<>();
|
|
final List<String> mainTitles = new ArrayList<>();
|
|
|
|
bioSchemaProtein.getEntryList().stream().forEach(entry -> {
|
|
|
|
if (entry.getType() != null
|
|
&& entry
|
|
.getType()
|
|
.equals("https://schema.org/Protein")) {
|
|
|
|
DataciteProtein dataciteProtein = new DataciteProtein();
|
|
|
|
citations.forEach(citation -> {
|
|
addRelatedIdentifier(dataciteProtein, citation, "IsCitedBy");
|
|
});
|
|
|
|
DataciteProtein.Types types = new DataciteProtein.Types();
|
|
types.setResourceType("Protein");
|
|
types.setResourceTypeGeneral("Dataset");
|
|
dataciteProtein.setTypes(types);
|
|
|
|
DataciteProtein.DataciteDate dataciteDate = new DataciteProtein.DataciteDate();
|
|
dataciteDate.setDate(retrievedOnType.getValue());
|
|
dataciteDate.setDateType("Collected");
|
|
dataciteProtein.getDates().add(dataciteDate);
|
|
|
|
DataciteProtein.Identifier identifier = new DataciteProtein.Identifier();
|
|
identifier.setIdentifier(entry.getId());
|
|
identifier.setIdentifierType("URL");
|
|
dataciteProtein.getIdentifiers().add(identifier);
|
|
|
|
if (entry.getIdentifier() != null) {
|
|
addAlternateIdentifier(dataciteProtein, entry.getIdentifier());
|
|
}
|
|
|
|
if (entry.getDescription() != null) {
|
|
DataciteProtein.Description description = new DataciteProtein.Description();
|
|
description.setDescription(entry.getDescription());
|
|
dataciteProtein.getDescriptions().add(description);
|
|
}
|
|
|
|
if (entry.getIsEncodedByBioChemEntity() != null) {
|
|
addRelatedIdentifier(dataciteProtein, entry.getIsEncodedByBioChemEntity(), "");
|
|
}
|
|
|
|
if (entry.getUrl() != null) {
|
|
addAlternateIdentifier(dataciteProtein, entry.getUrl());
|
|
}
|
|
|
|
if (entry.getAlternateName() != null) {
|
|
DataciteProtein.Title title = new DataciteProtein.Title();
|
|
title.setTitle(entry.getAlternateName());
|
|
title.setTitleType("AlternativeTitle");
|
|
dataciteProtein.getTitles().add(title);
|
|
}
|
|
|
|
if (entry.getBioChemInteraction() != null) {
|
|
entry.getBioChemInteraction().stream().filter(Objects::nonNull).forEach(bc -> {
|
|
addRelatedIdentifier(dataciteProtein, bc.getId(), "");
|
|
});
|
|
}
|
|
|
|
if (entry.getBioChemSimilarity() != null) {
|
|
entry.getBioChemSimilarity().stream().filter(Objects::nonNull).forEach(bc -> {
|
|
addRelatedIdentifier(dataciteProtein, bc.getId(), "");
|
|
});
|
|
}
|
|
|
|
if (entry.getHasMolecularFunction() != null) {
|
|
addRelatedIdentifier(dataciteProtein, entry.getHasMolecularFunction(), "");
|
|
}
|
|
|
|
if (entry.getIsInvolvedInBiologicalProcess() != null) {
|
|
addRelatedIdentifier(dataciteProtein, entry.getIsInvolvedInBiologicalProcess(), "");
|
|
}
|
|
|
|
if (entry.getIsEncodedByBioChemEntity() != null) {
|
|
addRelatedIdentifier(dataciteProtein, entry.getIsEncodedByBioChemEntity(), "");
|
|
}
|
|
|
|
if (entry.getIsPartOfBioChemEntity() != null) {
|
|
addRelatedIdentifier(dataciteProtein, entry.getIsPartOfBioChemEntity().getUrl(), "");
|
|
}
|
|
|
|
if (entry.getSameAs() != null) {
|
|
entry.getSameAs().stream().filter(Objects::nonNull).forEach(sameAs -> {
|
|
addRelatedIdentifier(dataciteProtein, sameAs.getId(), "IsIdenticalTo");
|
|
});
|
|
}
|
|
|
|
if (entry.getAssociatedDisease() != null) {
|
|
entry.getAssociatedDisease().stream().filter(Objects::nonNull).forEach(ad -> {
|
|
addRelatedIdentifier(dataciteProtein, ad.getName(), "IsIdenticalTo");
|
|
});
|
|
}
|
|
|
|
String proteinId = "";
|
|
try {
|
|
String[] identifierParts = dataciteProtein.getIdentifiers().get(0).getIdentifier().split("/");
|
|
proteinId = identifierParts[identifierParts.length - 1];
|
|
} catch (Exception e) {
|
|
log.error("Identifier not found", e.getMessage());
|
|
}
|
|
|
|
mainTitles.add(" " + entry.getName() + " ");
|
|
|
|
dataciteProtein.setId(proteinId);
|
|
bioschemaProteins.put(entry.getId(), entry);
|
|
dataciteProteins.add(dataciteProtein);
|
|
}
|
|
if (entry.getType() != null
|
|
&& entry
|
|
.getType()
|
|
.equals("https://schema.org/SequenceAnnotation")) {
|
|
sequenceAnnotations.put(entry.getId(), entry.getSequenceAnnotation());
|
|
}
|
|
if (entry.getType() != null
|
|
&& entry
|
|
.getType()
|
|
.equals("https://schema.org/PropertyValue")) {
|
|
propertyValues.put(entry.getId(), entry.getPropertyValue().getId());
|
|
}
|
|
if (entry.getType() != null
|
|
&& entry
|
|
.getType()
|
|
.equals("https://schema.org/DefinedTerm")) {
|
|
BioSchemaProtein.DefinedTerm term = new BioSchemaProtein.DefinedTerm();
|
|
term.setId(entry.getId());
|
|
term.setTermCode(entry.getTermCode());
|
|
term.setName(entry.getName());
|
|
definedTerms.put(term.getId(), term);
|
|
}
|
|
if (entry.getType() != null
|
|
&& entry
|
|
.getType()
|
|
.equals("https://schema.org/CollectionPage")) {
|
|
mainTitles.add(entry.getName());
|
|
}
|
|
});
|
|
bioschemaProteins
|
|
.entrySet()
|
|
.stream()
|
|
.filter(bioProtein -> Objects.nonNull(bioProtein.getValue().getHasSequenceAnnotation()))
|
|
.forEach(bioProtein -> {
|
|
List<String> seqAnnIds = bioProtein
|
|
.getValue()
|
|
.getHasSequenceAnnotation()
|
|
.stream()
|
|
.map(s -> s.getId())
|
|
.collect(Collectors.toList());
|
|
List<BioSchemaProtein.Link> propertyIds = sequenceAnnotations
|
|
.entrySet()
|
|
.stream()
|
|
.filter(s -> seqAnnIds.contains(s.getKey()))
|
|
.flatMap(s -> {
|
|
return s.getValue().stream();
|
|
})
|
|
.collect(Collectors.toList());
|
|
|
|
propertyIds
|
|
.stream()
|
|
.map(propertyId -> propertyValues.get(propertyId.getId()))
|
|
.filter(term -> Objects.nonNull(term))
|
|
.map(term -> definedTerms.get(term))
|
|
.filter(term -> Objects.nonNull(term))
|
|
.forEach(term -> {
|
|
dataciteProteins
|
|
.stream()
|
|
.filter(
|
|
dcProtein -> dcProtein
|
|
.getIdentifiers()
|
|
.get(0)
|
|
.getIdentifier()
|
|
.equals(bioProtein.getKey()))
|
|
.filter(dcProtein -> Objects.nonNull(dcProtein))
|
|
.forEach(d -> {
|
|
DataciteProtein.Subject subject = new DataciteProtein.Subject();
|
|
subject.setSchemeURI(term.getId());
|
|
subject.setValue(term.getName());
|
|
subject.setSubjectScheme(term.getTermCode());
|
|
d.getSubjects().add(subject);
|
|
});
|
|
});
|
|
});
|
|
|
|
dataciteProteins.forEach(d -> {
|
|
DataciteProtein.Title title = new DataciteProtein.Title();
|
|
title.setTitle(d.getId() + " - ".concat(mainTitles.get(0)));
|
|
d.getTitles().add(title);
|
|
});
|
|
|
|
ObjectMapper mapper = new ObjectMapper();
|
|
dataciteProteins.forEach(d -> {
|
|
StringWriter writer = new StringWriter();
|
|
try {
|
|
mapper.writeValue(writer, d);
|
|
} catch (IOException e) {
|
|
throw new RuntimeException(e);
|
|
}
|
|
results.add(writer.toString());
|
|
});
|
|
return results;
|
|
}
|
|
|
|
private void addRelatedIdentifier(DataciteProtein DataciteProtein, String relatedIdentifierValue,
|
|
String relationType) {
|
|
DataciteProtein.RelatedIdentifier relatedIdentifier = new DataciteProtein.RelatedIdentifier();
|
|
relatedIdentifier.setRelatedIdentifier(relatedIdentifierValue);
|
|
if (!relationType.isEmpty()) {
|
|
relatedIdentifier.setRelationType(relationType);
|
|
}
|
|
if (relatedIdentifierValue.contains("http://") || relatedIdentifierValue.contains("https://")) {
|
|
relatedIdentifier.setRelatedIdentifierType("URL");
|
|
}
|
|
DataciteProtein.getRelatedIdentifiers().add(relatedIdentifier);
|
|
}
|
|
|
|
private void addAlternateIdentifier(DataciteProtein DataciteProtein, String alternateIdentifierValue) {
|
|
DataciteProtein.AlternateIdentifier alternateIdentifier = new DataciteProtein.AlternateIdentifier();
|
|
alternateIdentifier.setAlternateIdentifier(alternateIdentifierValue);
|
|
DataciteProtein.getAlternateIdentifiers().add(alternateIdentifier);
|
|
}
|
|
}
|