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

Merge branch 'graph_cleaning' into stable_ids

This commit is contained in:
Claudio Atzori 2020-11-03 11:52:47 +01:00
parent 3fcd669e99 5310e56dba
commit 8471888ad3
1 changed files with 18 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctions.java
View File

@ -18,6 +18,7 @@ import eu.dnetlib.dhp.schema.oaf.*;
public class CleaningFunctions { public class CleaningFunctions {
public static final String ORCID_PREFIX_REGEX = "^http(s?):\\/\\/orcid\\.org\\/"; public static final String ORCID_PREFIX_REGEX = "^http(s?):\\/\\/orcid\\.org\\/";
public static final String NONE = "none";
public static <T extends Oaf> T fixVocabularyNames(T value) { public static <T extends Oaf> T fixVocabularyNames(T value) {
if (value instanceof Datasource) { if (value instanceof Datasource) {
@ -106,6 +107,23 @@ public class CleaningFunctions {
.filter(sp -> StringUtils.isNotBlank(sp.getQualifier().getClassid())) .filter(sp -> StringUtils.isNotBlank(sp.getQualifier().getClassid()))
.collect(Collectors.toList())); .collect(Collectors.toList()));
} }
if (Objects.nonNull(r.getPid())) {
r
.setPid(
r
.getPid()
.stream()
.filter(Objects::nonNull)
.filter(sp -> StringUtils.isNotBlank(StringUtils.trim(sp.getValue())))
.filter(sp -> NONE.equalsIgnoreCase(sp.getValue()))
.filter(sp -> Objects.nonNull(sp.getQualifier()))
.filter(sp -> StringUtils.isNotBlank(sp.getQualifier().getClassid()))
.map(sp -> {
sp.setValue(StringUtils.trim(sp.getValue()));
return sp;
})
.collect(Collectors.toList()));
}
if (Objects.isNull(r.getResourcetype()) || StringUtils.isBlank(r.getResourcetype().getClassid())) { if (Objects.isNull(r.getResourcetype()) || StringUtils.isBlank(r.getResourcetype().getClassid())) {
r r
.setResourcetype( .setResourcetype(