miriam.baglioni
/
dnet-hadoop
forked from D-Net/dnet-hadoop
2
0
Fork 0
Code Issues 2 Pull Requests Releases Wiki Activity

added set of same type of entries -url cf hb- before creating extras to have them distinct

This commit is contained in:
Miriam Baglioni 2020-06-22 17:45:38 +02:00
parent 004bf225cb
commit 1566fd590e
1 changed files with 12 additions and 6 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/dump/gcat/Mapper.java
View File

@ -19,6 +19,9 @@ public class Mapper implements Serializable {
final CatalogueEntry out = new CatalogueEntry();
Optional<Qualifier> ort = Optional.ofNullable(input.getResulttype());
List<KeyValue> externals = new ArrayList<>();
Set<String> urlSet = new HashSet<>();
Set<String> cfSet = new HashSet<>();
Set<String> hbSet = new HashSet<>();
if (ort.isPresent()) {
switch (ort.get().getClassid()) {
case "publication":
@ -43,13 +46,13 @@ public class Mapper implements Serializable {
eu.dnetlib.dhp.schema.oaf.Software is = (eu.dnetlib.dhp.schema.oaf.Software) input;
Optional
.ofNullable(is.getCodeRepositoryUrl())
.ifPresent(value -> externals.add(KeyValue.newInstance("url", value.getValue())));
.ifPresent(value -> urlSet.add(value.getValue()));
Optional
.ofNullable(is.getDocumentationUrl())
.ifPresent(
value -> value
.stream()
.map(v -> externals.add(KeyValue.newInstance("url", v.getValue()))));
.map(v -> urlSet.add(v.getValue())));
Optional
.ofNullable(is.getProgrammingLanguage())
@ -80,7 +83,7 @@ public class Mapper implements Serializable {
.ofNullable(input.getCollectedfrom())
.ifPresent(
value -> value
.forEach(v -> externals.add(KeyValue.newInstance("collected from", v.getValue()))));
.forEach(v -> cfSet.add(v.getValue())));
Optional
.ofNullable(input.getContributor())
@ -124,13 +127,12 @@ public class Mapper implements Serializable {
Optional
.ofNullable(v.getHostedby())
.ifPresent(hb -> externals.add(KeyValue.newInstance("hosted by", hb.getValue())));
.ifPresent(hb -> hbSet.add(hb.getValue()));
final HashSet<String> urlSet = new HashSet<>();
Optional
.ofNullable(v.getUrl())
.ifPresent(u -> u.forEach(url -> urlSet.add(url)));
urlSet.forEach(url -> externals.add(KeyValue.newInstance("url", url)));
}));
@ -180,6 +182,10 @@ public class Mapper implements Serializable {
.newInstance("subject", s.getQualifier().getClassid() + ":" + s.getValue()))));
externals.add(KeyValue.newInstance("resource type", input.getResourcetype().getClassid()));
cfSet.forEach(cf -> externals.add(KeyValue.newInstance("collected from", cf)));
hbSet.forEach(hb -> externals.add(KeyValue.newInstance("hosted by", hb)));
urlSet.forEach(url -> externals.add(KeyValue.newInstance("url", url)));
out.setExtras(externals);
}