From 431cbe9955530806bd202cf76b30358ea55f7d6a Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Tue, 23 Mar 2021 09:28:58 +0100
Subject: [PATCH] handle missing instance.pid during bulk cleaning

---
 .../dhp/schema/oaf/CleaningFunctions.java     | 26 +++++++++++--------
 .../oa/graph/clean/CleaningFunctionTest.java  |  3 ++-
 2 files changed, 17 insertions(+), 12 deletions(-)

diff --git a/dhp-common/src/main/java/eu/dnetlib/dhp/schema/oaf/CleaningFunctions.java b/dhp-common/src/main/java/eu/dnetlib/dhp/schema/oaf/CleaningFunctions.java
index 1cee3058e..412ed408e 100644
--- a/dhp-common/src/main/java/eu/dnetlib/dhp/schema/oaf/CleaningFunctions.java
+++ b/dhp-common/src/main/java/eu/dnetlib/dhp/schema/oaf/CleaningFunctions.java
@@ -149,17 +149,21 @@ public class CleaningFunctions {
 			if (Objects.nonNull(r.getInstance())) {
 
 				for (Instance i : r.getInstance()) {
-					final Set<StructuredProperty> pids = Sets.newHashSet(i.getPid());
-					i
-						.setAlternateIdentifier(
-							Optional
-								.ofNullable(i.getAlternateIdentifier())
-								.map(
-									altId -> altId
-										.stream()
-										.filter(p -> !pids.contains(p))
-										.collect(Collectors.toList()))
-								.orElse(Lists.newArrayList()));
+					Optional
+						.ofNullable(i.getPid())
+						.ifPresent(pid -> {
+							final Set<StructuredProperty> pids = Sets.newHashSet(i.getPid());
+							i
+								.setAlternateIdentifier(
+									Optional
+										.ofNullable(i.getAlternateIdentifier())
+										.map(
+											altId -> altId
+												.stream()
+												.filter(p -> !pids.contains(p))
+												.collect(Collectors.toList()))
+										.orElse(Lists.newArrayList()));
+						});
 
 					if (Objects.isNull(i.getAccessright()) || StringUtils.isBlank(i.getAccessright().getClassid())) {
 						i
diff --git a/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctionTest.java b/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctionTest.java
index 3a7a3ee19..0860c8bde 100644
--- a/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctionTest.java
+++ b/dhp-workflows/dhp-graph-mapper/src/test/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctionTest.java
@@ -19,6 +19,7 @@ import org.mockito.junit.jupiter.MockitoExtension;
 import com.fasterxml.jackson.databind.ObjectMapper;
 
 import eu.dnetlib.dhp.common.vocabulary.VocabularyGroup;
+import eu.dnetlib.dhp.schema.common.ModelConstants;
 import eu.dnetlib.dhp.schema.oaf.*;
 import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpException;
 import eu.dnetlib.enabling.is.lookup.rmi.ISLookUpService;
@@ -78,7 +79,7 @@ public class CleaningFunctionTest {
 		assertEquals("CLOSED", p_out.getInstance().get(0).getAccessright().getClassid());
 		assertEquals("Closed Access", p_out.getInstance().get(0).getAccessright().getClassname());
 
-		Set<String> pidTerms = vocabularies.getTerms("dnet:pid_types");
+		Set<String> pidTerms = vocabularies.getTerms(ModelConstants.DNET_PID_TYPES);
 		assertTrue(
 			p_out
 				.getPid()