pispis
/
dnet-hadoop
forked from antonis.lempesis/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

filter ORCID e MAG identifiers

This commit is contained in:
Michele Artini 2020-05-22 11:00:27 +02:00
parent 9de71e54a8
commit 9f2d0f1b08
2 changed files with 17 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/raw/OdfToOafMapper.java
View File

@ -102,7 +102,8 @@ public class OdfToOafMapper extends AbstractMdRecordToOafMapper {
final String type = ((Node) o).valueOf("./@nameIdentifierScheme") final String type = ((Node) o).valueOf("./@nameIdentifierScheme")
.trim() .trim()
.toUpperCase() .toUpperCase()
.replaceAll(" ", ""); .replaceAll(" ", "")
.replaceAll("_", "");
if (type.startsWith("ORCID")) { if (type.startsWith("ORCID")) {
final String cleanedId = id.replaceAll("http://orcid.org/", "").replaceAll("https://orcid.org/", ""); final String cleanedId = id.replaceAll("http://orcid.org/", "").replaceAll("https://orcid.org/", "");

23
dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/raw/MappersTest.java
View File

@ -21,7 +21,14 @@ import org.mockito.Mock;
import org.mockito.junit.jupiter.MockitoExtension; import org.mockito.junit.jupiter.MockitoExtension;
import eu.dnetlib.dhp.schema.common.ModelConstants; import eu.dnetlib.dhp.schema.common.ModelConstants;
import eu.dnetlib.dhp.schema.oaf.*; import eu.dnetlib.dhp.schema.oaf.Author;
import eu.dnetlib.dhp.schema.oaf.Dataset;
import eu.dnetlib.dhp.schema.oaf.Field;
import eu.dnetlib.dhp.schema.oaf.Oaf;
import eu.dnetlib.dhp.schema.oaf.Publication;
import eu.dnetlib.dhp.schema.oaf.Relation;
import eu.dnetlib.dhp.schema.oaf.Software;
import eu.dnetlib.dhp.schema.oaf.StructuredProperty;
@ExtendWith(MockitoExtension.class) @ExtendWith(MockitoExtension.class)
public class MappersTest { public class MappersTest {
@ -54,13 +61,13 @@ public class MappersTest {
assertTrue(StringUtils.isNotBlank(p.getTitle().get(0).getValue())); assertTrue(StringUtils.isNotBlank(p.getTitle().get(0).getValue()));
assertTrue(p.getAuthor().size() > 0); assertTrue(p.getAuthor().size() > 0);
Optional<Author> author = p final Optional<Author> author = p
.getAuthor() .getAuthor()
.stream() .stream()
.filter(a -> a.getPid() != null && !a.getPid().isEmpty()) .filter(a -> a.getPid() != null && !a.getPid().isEmpty())
.findFirst(); .findFirst();
assertTrue(author.isPresent()); assertTrue(author.isPresent());
StructuredProperty pid = author final StructuredProperty pid = author
.get() .get()
.getPid() .getPid()
.stream() .stream()
@ -121,13 +128,13 @@ public class MappersTest {
assertTrue(StringUtils.isNotBlank(d.getTitle().get(0).getValue())); assertTrue(StringUtils.isNotBlank(d.getTitle().get(0).getValue()));
assertTrue(d.getAuthor().size() > 0); assertTrue(d.getAuthor().size() > 0);
Optional<Author> author = d final Optional<Author> author = d
.getAuthor() .getAuthor()
.stream() .stream()
.filter(a -> a.getPid() != null && !a.getPid().isEmpty()) .filter(a -> a.getPid() != null && !a.getPid().isEmpty())
.findFirst(); .findFirst();
assertTrue(author.isPresent()); assertTrue(author.isPresent());
StructuredProperty pid = author final StructuredProperty pid = author
.get() .get()
.getPid() .getPid()
.stream() .stream()
@ -135,7 +142,7 @@ public class MappersTest {
.get(); .get();
assertEquals("0000-0001-9074-1619", pid.getValue()); assertEquals("0000-0001-9074-1619", pid.getValue());
assertEquals("ORCID", pid.getQualifier().getClassid()); assertEquals("ORCID", pid.getQualifier().getClassid());
assertEquals("ORCID", pid.getQualifier().getClassname()); assertEquals("Open Researcher and Contributor ID", pid.getQualifier().getClassname());
assertEquals(ModelConstants.DNET_PID_TYPES, pid.getQualifier().getSchemeid()); assertEquals(ModelConstants.DNET_PID_TYPES, pid.getQualifier().getSchemeid());
assertEquals(ModelConstants.DNET_PID_TYPES, pid.getQualifier().getSchemename()); assertEquals(ModelConstants.DNET_PID_TYPES, pid.getQualifier().getSchemename());
assertEquals("Baracchini, Theo", author.get().getFullname()); assertEquals("Baracchini, Theo", author.get().getFullname());
@ -143,13 +150,13 @@ public class MappersTest {
assertEquals("Theo", author.get().getName()); assertEquals("Theo", author.get().getName());
assertEquals(1, author.get().getAffiliation().size()); assertEquals(1, author.get().getAffiliation().size());
Optional<Field<String>> opAff = author final Optional<Field<String>> opAff = author
.get() .get()
.getAffiliation() .getAffiliation()
.stream() .stream()
.findFirst(); .findFirst();
assertTrue(opAff.isPresent()); assertTrue(opAff.isPresent());
Field<String> affiliation = opAff.get(); final Field<String> affiliation = opAff.get();
assertEquals("ISTI-CNR", affiliation.getValue()); assertEquals("ISTI-CNR", affiliation.getValue());
assertTrue(d.getSubject().size() > 0); assertTrue(d.getSubject().size() > 0);