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

avoid NPEs merging titles

This commit is contained in:
Claudio Atzori 2020-06-12 10:45:11 +02:00
parent d1d92c4d8c
commit ba8a024af9
2 changed files with 5 additions and 96 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
dhp-schemas/src/main/java/eu/dnetlib/dhp/schema/oaf/Result.java
View File

@ -254,28 +254,25 @@ public class Result extends OafEntity implements Serializable {
final StructuredProperty p = baseMainTitle; final StructuredProperty p = baseMainTitle;
title = title.stream().filter(t -> t != p).collect(Collectors.toList()); title = title.stream().filter(t -> t != p).collect(Collectors.toList());
} }
//
//
// title.remove(baseMainTitle);
} }
StructuredProperty newMainTitle = null; StructuredProperty newMainTitle = null;
if (r.getTitle() != null) { if (r.getTitle() != null) {
newMainTitle = getMainTitle(r.getTitle()); newMainTitle = getMainTitle(r.getTitle());
if (newMainTitle != null) { if (newMainTitle != null && title != null) {
final StructuredProperty p = newMainTitle; final StructuredProperty p = newMainTitle;
title = title.stream().filter(t -> t != p).collect(Collectors.toList()); title = title.stream().filter(t -> t != p).collect(Collectors.toList());
} }
// r.getTitle().remove(newMainTitle);
} }
if (newMainTitle != null && compareTrust(this, r) < 0) if (newMainTitle != null && compareTrust(this, r) < 0) {
baseMainTitle = newMainTitle; baseMainTitle = newMainTitle;
}
title = mergeLists(title, r.getTitle()); title = mergeLists(title, r.getTitle());
if (title != null && baseMainTitle != null) if (title != null && baseMainTitle != null) {
title.add(baseMainTitle); title.add(baseMainTitle);
}
relevantdate = mergeLists(relevantdate, r.getRelevantdate()); relevantdate = mergeLists(relevantdate, r.getRelevantdate());

88
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/NormalizeEmptyFields.java
View File

@ -1,88 +0,0 @@
package eu.dnetlib.dhp.oa.graph.clean;
import eu.dnetlib.dhp.oa.graph.raw.common.OafMapperUtils;
import eu.dnetlib.dhp.oa.graph.raw.common.VocabularyGroup;
import eu.dnetlib.dhp.schema.oaf.Oaf;
import eu.dnetlib.dhp.schema.oaf.Qualifier;
import org.apache.commons.lang3.StringUtils;
import org.apache.spark.api.java.function.MapFunction;
import java.lang.reflect.Field;
import java.util.Arrays;
import java.util.LinkedList;
import java.util.List;
import java.util.Objects;
public class NormalizeEmptyFields<T extends Oaf> implements MapFunction<T, T> {
private VocabularyGroup vocabularies;
public NormalizeEmptyFields(VocabularyGroup vocabularies) {
this.vocabularies = vocabularies;
}
@Override
public T call(T value) throws Exception {
doNormalize(value);
return value;
}
private void doNormalize(Object o) {
if (Objects.isNull(o)) {
return;
}
if (o instanceof Iterable) {
for (Object oi : (Iterable) o) {
doNormalize(oi);
}
} else {
Class clazz = o.getClass();
if (clazz.isPrimitive()
|| o instanceof Integer
|| o instanceof Double
|| o instanceof Float
|| o instanceof Long
|| o instanceof Boolean
|| o instanceof String) {
return;
} else {
try {
for (Field field : getAllFields(new LinkedList<>(), clazz)) {
field.setAccessible(true);
Object value = field.get(o);
if (value instanceof Qualifier && Objects.isNull(value)) {
field.set(o, OafMapperUtils.unknown("", ""));
} else if (value instanceof Field && Objects.isNull(value)) {
} else {
doNormalize(value);
}
}
} catch (IllegalAccessException | IllegalArgumentException e) {
throw new RuntimeException(e);
}
}
}
}
private static List<Field> getAllFields(List<Field> fields, Class<?> clazz) {
fields.addAll(Arrays.asList(clazz.getDeclaredFields()));
final Class<?> superclass = clazz.getSuperclass();
if (Objects.nonNull(superclass) && superclass.getPackage().equals(Oaf.class.getPackage())) {
getAllFields(fields, superclass);
}
return fields;
}
public VocabularyGroup getVocabularies() {
return vocabularies;
}
}