lsmyrnaios
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

Fix: do not import joda 

formatting
This commit is contained in:
Giambattista Bloisi 2024-12-05 15:21:32 +01:00
parent 89b7bc84f2
commit fed13e083e
6 changed files with 83 additions and 59 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

88
dhp-pace-core/src/main/java/eu/dnetlib/pace/tree/DateRange.java
View File

@ -1,10 +1,5 @@
package eu.dnetlib.pace.tree;
import com.wcohen.ss.AbstractStringDistance;
import eu.dnetlib.pace.config.Config;
import eu.dnetlib.pace.tree.support.AbstractStringComparator;
import eu.dnetlib.pace.tree.support.ComparatorClass;
import org.joda.time.DateTime;
package eu.dnetlib.pace.tree;
import java.time.DateTimeException;
import java.time.LocalDate;
@ -13,55 +8,60 @@ import java.time.format.DateTimeFormatter;
import java.util.Locale;
import java.util.Map;
import com.wcohen.ss.AbstractStringDistance;
import eu.dnetlib.pace.config.Config;
import eu.dnetlib.pace.tree.support.AbstractStringComparator;
import eu.dnetlib.pace.tree.support.ComparatorClass;
@ComparatorClass("dateRange")
public class DateRange extends AbstractStringComparator {
int YEAR_RANGE;
int YEAR_RANGE;
public DateRange(Map<String, String> params) {
super(params, new com.wcohen.ss.JaroWinkler());
YEAR_RANGE = Integer.parseInt(params.getOrDefault("year_range", "3"));
}
public DateRange(Map<String, String> params) {
super(params, new com.wcohen.ss.JaroWinkler());
YEAR_RANGE = Integer.parseInt(params.getOrDefault("year_range", "3"));
}
public DateRange(final double weight) {
super(weight, new com.wcohen.ss.JaroWinkler());
}
public DateRange(final double weight) {
super(weight, new com.wcohen.ss.JaroWinkler());
}
protected DateRange(final double weight, final AbstractStringDistance ssalgo) {
super(weight, ssalgo);
}
protected DateRange(final double weight, final AbstractStringDistance ssalgo) {
super(weight, ssalgo);
}
public static boolean isNumeric(String str) {
return str.matches("\\d+"); //match a number with optional '-' and decimal.
}
public static boolean isNumeric(String str) {
return str.matches("\\d+"); // match a number with optional '-' and decimal.
}
@Override
public double distance(final String a, final String b, final Config conf) {
if (a.isEmpty() || b.isEmpty()) {
return -1.0; // return -1 if a field is missing
}
@Override
public double distance(final String a, final String b, final Config conf) {
if (a.isEmpty() || b.isEmpty()) {
return -1.0; // return -1 if a field is missing
}
try {
DateTimeFormatter formatter = DateTimeFormatter.ofPattern("yyyy-MM-dd", Locale.ENGLISH);
LocalDate d1 = LocalDate.parse(a, formatter);
LocalDate d2 = LocalDate.parse(b, formatter);
Period period = Period.between(d1, d2);
try {
DateTimeFormatter formatter = DateTimeFormatter.ofPattern("yyyy-MM-dd", Locale.ENGLISH);
LocalDate d1 = LocalDate.parse(a, formatter);
LocalDate d2 = LocalDate.parse(b, formatter);
Period period = Period.between(d1, d2);
return period.getYears() <= YEAR_RANGE? 1.0 : 0.0;
}
catch (DateTimeException e) {
return -1.0;
}
return period.getYears() <= YEAR_RANGE ? 1.0 : 0.0;
} catch (DateTimeException e) {
return -1.0;
}
}
}
@Override
public double getWeight() {
return super.weight;
}
@Override
public double getWeight() {
return super.weight;
}
@Override
protected double normalize(final double d) {
return d;
}
@Override
protected double normalize(final double d) {
return d;
}
}

2
dhp-pace-core/src/main/java/eu/dnetlib/pace/tree/JsonListMatch.java
View File

@ -62,7 +62,7 @@ public class JsonListMatch extends AbstractListComparator {
Set<String> types = Sets.intersection(typesA, typesB);
if (types.isEmpty()) // if no common type, it is impossible to compare
if (types.isEmpty()) // if no common type, it is impossible to compare
return -1;
ca = ca.stream().filter(s -> types.contains(s.split("::")[0])).collect(Collectors.toSet());

30
dhp-pace-core/src/test/java/eu/dnetlib/pace/comparators/ComparatorTest.java
View File

@ -72,14 +72,34 @@ public class ComparatorTest extends AbstractPaceTest {
CodeMatch codeMatch = new CodeMatch(params);
// names have different codes
assertEquals(0.0, codeMatch.distance("physical oceanography at ctd station june 1998 ev02a", "physical oceanography at ctd station june 1998 ir02", conf));
assertEquals(
0.0,
codeMatch
.distance(
"physical oceanography at ctd station june 1998 ev02a",
"physical oceanography at ctd station june 1998 ir02", conf));
// names have same code
assertEquals(1.0, codeMatch.distance("physical oceanography at ctd station june 1998 ev02a", "physical oceanography at ctd station june 1998 ev02a", conf));
assertEquals(
1.0,
codeMatch
.distance(
"physical oceanography at ctd station june 1998 ev02a",
"physical oceanography at ctd station june 1998 ev02a", conf));
// code is not in both names
assertEquals(-1, codeMatch.distance("physical oceanography at ctd station june 1998", "physical oceanography at ctd station june 1998 ev02a", conf));
assertEquals(1.0, codeMatch.distance("physical oceanography at ctd station june 1998", "physical oceanography at ctd station june 1998", conf));
assertEquals(
-1,
codeMatch
.distance(
"physical oceanography at ctd station june 1998",
"physical oceanography at ctd station june 1998 ev02a", conf));
assertEquals(
1.0,
codeMatch
.distance(
"physical oceanography at ctd station june 1998", "physical oceanography at ctd station june 1998",
conf));
}
@Test
@ -275,7 +295,7 @@ public class ComparatorTest extends AbstractPaceTest {
Arrays
.asList(
"{\"datainfo\":{\"deletedbyinference\":false,\"inferenceprovenance\":null,\"inferred\":false,\"invisible\":false,\"provenanceaction\":{\"classid\":\"sysimport:actionset\",\"classname\":\"Harvested\",\"schemeid\":\"dnet:provenanceActions\",\"schemename\":\"dnet:provenanceActions\"},\"trust\":\"0.9\"},\"qualifier\":{\"classid\":\"grid\",\"classname\":\"GRID Identifier\",\"schemeid\":\"dnet:pid_types\",\"schemename\":\"dnet:pid_types\"},\"value\":\"grid_1\"}",
"{\"datainfo\":{\"deletedbyinference\":false,\"inferenceprovenance\":null,\"inferred\":false,\"invisible\":false,\"provenanceaction\":{\"classid\":\"sysimport:actionset\",\"classname\":\"Harvested\",\"schemeid\":\"dnet:provenanceActions\",\"schemename\":\"dnet:provenanceActions\"},\"trust\":\"0.9\"},\"qualifier\":{\"classid\":\"ror\",\"classname\":\"Research Organization Registry\",\"schemeid\":\"dnet:pid_types\",\"schemename\":\"dnet:pid_types\"},\"value\":\"ror_1\"}"),
"{\"datainfo\":{\"deletedbyinference\":false,\"inferenceprovenance\":null,\"inferred\":false,\"invisible\":false,\"provenanceaction\":{\"classid\":\"sysimport:actionset\",\"classname\":\"Harvested\",\"schemeid\":\"dnet:provenanceActions\",\"schemename\":\"dnet:provenanceActions\"},\"trust\":\"0.9\"},\"qualifier\":{\"classid\":\"ror\",\"classname\":\"Research Organization Registry\",\"schemeid\":\"dnet:pid_types\",\"schemename\":\"dnet:pid_types\"},\"value\":\"ror_1\"}"),
"authors");
List<String> b = createFieldList(
Arrays

4
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/community/ResultTagger.java
View File

@ -130,7 +130,7 @@ public class ResultTagger implements Serializable {
// log.info("Remove constraints for " + communityId);
if (conf.getRemoveConstraintsMap().keySet().contains(communityId) &&
conf.getRemoveConstraintsMap().get(communityId).getCriteria() != null &&
!conf.getRemoveConstraintsMap().get(communityId).getCriteria().isEmpty() &&
!conf.getRemoveConstraintsMap().get(communityId).getCriteria().isEmpty() &&
conf
.getRemoveConstraintsMap()
.get(communityId)
@ -228,7 +228,7 @@ public class ResultTagger implements Serializable {
.forEach(communityId -> {
if (!removeCommunities.contains(communityId) &&
conf.getSelectionConstraintsMap().get(communityId).getCriteria() != null &&
!conf.getSelectionConstraintsMap().get(communityId).getCriteria().isEmpty() &&
!conf.getSelectionConstraintsMap().get(communityId).getCriteria().isEmpty() &&
conf
.getSelectionConstraintsMap()
.get(communityId)

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

@ -915,7 +915,8 @@ class MappersTest {
@Test
void testODFRecord_guidelines4() throws IOException {
final String xml = IOUtils.toString(Objects.requireNonNull(getClass().getResourceAsStream("odf_guidelines4.xml")));
final String xml = IOUtils
.toString(Objects.requireNonNull(getClass().getResourceAsStream("odf_guidelines4.xml")));
final List<Oaf> list = new OdfToOafMapper(vocs, false, true).processMdRecord(xml);
final Publication p = (Publication) list.get(0);

15
dhp-workflows/dhp-graph-provision/src/main/java/eu/dnetlib/dhp/oa/provision/model/ProvisionModelSupport.java
View File

@ -5,7 +5,6 @@ import java.io.StringReader;
import java.util.*;
import java.util.stream.Collectors;
import eu.dnetlib.dhp.schema.solr.PersonTopic;
import org.apache.commons.lang3.StringUtils;
import org.dom4j.Document;
import org.dom4j.DocumentException;
@ -40,6 +39,7 @@ import eu.dnetlib.dhp.schema.solr.OpenAccessColor;
import eu.dnetlib.dhp.schema.solr.OpenAccessRoute;
import eu.dnetlib.dhp.schema.solr.Organization;
import eu.dnetlib.dhp.schema.solr.Person;
import eu.dnetlib.dhp.schema.solr.PersonTopic;
import eu.dnetlib.dhp.schema.solr.Pid;
import eu.dnetlib.dhp.schema.solr.Project;
import eu.dnetlib.dhp.schema.solr.Result;
@ -216,11 +216,14 @@ public class ProvisionModelSupport {
}
private static List<PersonTopic> mapPersonTopics(List<eu.dnetlib.dhp.schema.oaf.PersonTopic> subjects) {
return Optional.ofNullable(subjects)
.map(ss -> ss.stream()
.map(ProvisionModelSupport::mapPersonTopic)
.collect(Collectors.toList()))
.orElse(null);
return Optional
.ofNullable(subjects)
.map(
ss -> ss
.stream()
.map(ProvisionModelSupport::mapPersonTopic)
.collect(Collectors.toList()))
.orElse(null);
}
private static PersonTopic mapPersonTopic(eu.dnetlib.dhp.schema.oaf.PersonTopic pt) {