From 2e9e13444d2c4933c2e012a084e052e9ad6e24c4 Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Thu, 1 Oct 2020 12:50:40 +0200
Subject: [PATCH] author pids made unique by value

---
 .../dhp/oa/graph/clean/CleaningFunctions.java | 24 ++++++++++++++++++
 .../eu/dnetlib/dhp/oa/graph/clean/result.json | 25 +++++++++++++++++--
 2 files changed, 47 insertions(+), 2 deletions(-)

diff --git a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctions.java b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctions.java
index 3a0eace1f..f615d69f2 100644
--- a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctions.java
+++ b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/CleaningFunctions.java
@@ -1,9 +1,12 @@
 
 package eu.dnetlib.dhp.oa.graph.clean;
 
+import java.util.LinkedHashMap;
 import java.util.Objects;
+import java.util.function.Function;
 import java.util.stream.Collectors;
 
+import com.clearspring.analytics.util.Lists;
 import org.apache.commons.lang3.StringUtils;
 
 import eu.dnetlib.dhp.oa.graph.raw.AbstractMdRecordToOafMapper;
@@ -13,6 +16,8 @@ import eu.dnetlib.dhp.schema.oaf.*;
 
 public class CleaningFunctions {
 
+	public static final String ORCID_PREFIX_REGEX = "^http(s?):\\/\\/orcid\\.org\\/";
+
 	public static <T extends Oaf> T fixVocabularyNames(T value) {
 		if (value instanceof Datasource) {
 			// nothing to clean here
@@ -139,6 +144,25 @@ public class CleaningFunctions {
 						author.setRank(i++);
 					}
 				}
+				for(Author a : r.getAuthor()) {
+					if (Objects.isNull(a.getPid())) {
+						a.setPid(Lists.newArrayList());
+					} else {
+						a.setPid(
+							a.getPid().stream()
+								.filter(p -> Objects.nonNull(p.getQualifier()))
+								.filter(p -> StringUtils.isNotBlank(p.getValue()))
+								.map(p -> {
+									p.setValue(p.getValue().trim().replaceAll(ORCID_PREFIX_REGEX, ""));
+									return p;
+								})
+								.collect(Collectors.toMap(StructuredProperty::getValue, Function.identity(), (p1, p2) -> p1, LinkedHashMap::new))
+								.values()
+								.stream()
+								.collect(Collectors.toList()));
+					}
+				}
+
 			}
 			if (value instanceof Publication) {
 
diff --git a/dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/clean/result.json b/dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/clean/result.json
index f51eed067..5c903cd0e 100644
--- a/dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/clean/result.json
+++ b/dhp-workflows/dhp-graph-mapper/src/test/resources/eu/dnetlib/dhp/oa/graph/clean/result.json
@@ -27,6 +27,28 @@
             "schemename": "dnet:pid_types"
           },
           "value": "0000-0001-9613-6639"
+        },
+        {
+          "dataInfo": {
+            "deletedbyinference": false,
+            "inferenceprovenance": "",
+            "inferred": false,
+            "invisible": false,
+            "provenanceaction": {
+              "classid": "sysimport:crosswalk:datasetarchive",
+              "classname": "sysimport:crosswalk:datasetarchive",
+              "schemeid": "dnet:provenanceActions",
+              "schemename": "dnet:provenanceActions"
+            },
+            "trust": "0.9"
+          },
+          "qualifier": {
+            "classid": "ORCID12",
+            "classname": "ORCID12",
+            "schemeid": "dnet:pid_types",
+            "schemename": "dnet:pid_types"
+          },
+          "value": "https://orcid.org/0000-0001-9613-6639"
         }
       ],
       "rank": 1,
@@ -91,8 +113,7 @@
       ],
       "fullname": "Barry, Peter S.",
       "name": "Peter S.",
-      "pid": [
-      ],
+      "pid": null,
       "rank": 3,
       "surname": "Barry"
     },