dimitris.pierrakos
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

[EOSCTag] changed code to remove EOSC Jupyter Notebook and modified test to exclude galaxy + software from the tagging for Galaxy

This commit is contained in:
Miriam Baglioni 2022-05-17 15:09:22 +02:00
parent e4eac1d20b
commit f5207885e3
3 changed files with 10 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/bulktag/SparkEoscTag.java
View File

@ -95,13 +95,14 @@ public class SparkEoscTag {
if (containsCriteriaNotebook(s)) { if (containsCriteriaNotebook(s)) {
sbject.add(EOSC_NOTEBOOK); sbject.add(EOSC_NOTEBOOK);
if (sbject.stream().anyMatch(sb -> sb.getValue().equals("EOSC Jupyter Notebook"))){ if (sbject.stream().anyMatch(sb -> sb.getValue().equals("EOSC Jupyter Notebook"))) {
sbject = sbject.stream().map(sb -> { sbject = sbject.stream().map(sb -> {
if (sb.getValue().equals("EOSC Jupyter Notebook")){ if (sb.getValue().equals("EOSC Jupyter Notebook")) {
return null; return null;
} }
return sb; return sb;
}).filter(Objects::nonNull).collect(Collectors.toList()); }).filter(Objects::nonNull).collect(Collectors.toList());
s.setSubject(sbject);
} }
} }
if (containsCriteriaGalaxy(s)) { if (containsCriteriaGalaxy(s)) {

21
dhp-workflows/dhp-enrichment/src/test/java/eu/dnetlib/dhp/bulktag/EOSCTagJobTest.java
View File

@ -132,7 +132,7 @@ public class EOSCTagJobTest {
Assertions Assertions
.assertEquals( .assertEquals(
2, tmp 1, tmp
.filter(sw -> sw.getId().equals("50|od______1582::4132f5ec9496f0d6adc7b00a50a56ff4")) .filter(sw -> sw.getId().equals("50|od______1582::4132f5ec9496f0d6adc7b00a50a56ff4"))
.collect() .collect()
.get(0) .get(0)
@ -326,7 +326,7 @@ public class EOSCTagJobTest {
Assertions Assertions
.assertEquals( .assertEquals(
2, 1,
tmp tmp
.filter( .filter(
s -> s.getSubject().stream().anyMatch(sbj -> sbj.getValue().equals("EOSC::Galaxy Workflow"))) s -> s.getSubject().stream().anyMatch(sbj -> sbj.getValue().equals("EOSC::Galaxy Workflow")))
@ -352,21 +352,12 @@ public class EOSCTagJobTest {
Assertions Assertions
.assertEquals( .assertEquals(
6, tmp 5, tmp
.filter(sw -> sw.getId().equals("50|od______1582::501b25d420f808c8eddcd9b16e917f11")) .filter(sw -> sw.getId().equals("50|od______1582::501b25d420f808c8eddcd9b16e917f11"))
.collect() .collect()
.get(0) .get(0)
.getSubject() .getSubject()
.size()); .size());
Assertions
.assertTrue(
tmp
.filter(sw -> sw.getId().equals("50|od______1582::501b25d420f808c8eddcd9b16e917f11"))
.collect()
.get(0)
.getSubject()
.stream()
.anyMatch(s -> s.getValue().equals("EOSC::Galaxy Workflow")));
Assertions Assertions
.assertEquals( .assertEquals(
@ -394,7 +385,7 @@ public class EOSCTagJobTest {
Assertions Assertions
.assertEquals( .assertEquals(
2, 1,
orp orp
.filter( .filter(
s -> s.getSubject().stream().anyMatch(sbj -> sbj.getValue().equals("EOSC::Galaxy Workflow"))) s -> s.getSubject().stream().anyMatch(sbj -> sbj.getValue().equals("EOSC::Galaxy Workflow")))
@ -438,14 +429,14 @@ public class EOSCTagJobTest {
Assertions Assertions
.assertEquals( .assertEquals(
3, orp 2, orp
.filter(sw -> sw.getId().equals("50|od______2017::1e400f1747487fd15998735c41a55c72")) .filter(sw -> sw.getId().equals("50|od______2017::1e400f1747487fd15998735c41a55c72"))
.collect() .collect()
.get(0) .get(0)
.getSubject() .getSubject()
.size()); .size());
Assertions Assertions
.assertTrue( .assertFalse(
orp orp
.filter(sw -> sw.getId().equals("50|od______2017::1e400f1747487fd15998735c41a55c72")) .filter(sw -> sw.getId().equals("50|od______2017::1e400f1747487fd15998735c41a55c72"))
.collect() .collect()

2
dhp-workflows/dhp-enrichment/src/test/resources/eu/dnetlib/dhp/eosctag/jupyter/software/software_10.json
View File

File diff suppressed because one or more lines are too long