From b0203ff5ccd29e834047effd050abf758c8f1720 Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Mon, 5 Jul 2021 14:35:26 +0200
Subject: [PATCH] added more cleaning/identifier utilities

---
 .../schema/oaf/utils/CleaningFunctions.java   | 29 ++++++++++++-------
 .../schema/oaf/utils/IdentifierFactory.java   | 20 +++++++------
 2 files changed, 30 insertions(+), 19 deletions(-)

diff --git a/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/CleaningFunctions.java b/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/CleaningFunctions.java
index 56ee75a..183214c 100644
--- a/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/CleaningFunctions.java
+++ b/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/CleaningFunctions.java
@@ -46,18 +46,27 @@ public class CleaningFunctions {
 	 * @return the PID containing the normalised value.
 	 */
 	public static StructuredProperty normalizePidValue(StructuredProperty pid) {
-		String value = Optional
-			.ofNullable(pid.getValue())
-			.map(String::trim)
-			.orElseThrow(() -> new IllegalArgumentException("PID value cannot be empty"));
-		switch (pid.getQualifier().getClassid()) {
+		pid.setValue(
+				normalizePidValue(
+					pid.getQualifier().getClassid(),
+					pid.getValue()));
 
-			// TODO add cleaning for more PID types as needed
-			case "doi":
-				pid.setValue(value.toLowerCase().replaceFirst(DOI_PREFIX_REGEX, DOI_PREFIX));
-				break;
-		}
 		return pid;
 	}
 
+	public static String normalizePidValue(String pidType, String pidValue) {
+		String value = Optional
+				.ofNullable(pidValue)
+				.map(String::trim)
+				.orElseThrow(() -> new IllegalArgumentException("PID value cannot be empty"));
+
+		switch (pidType) {
+
+			// TODO add cleaning for more PID types as needed
+			case "doi":
+				return value.toLowerCase().replaceFirst(DOI_PREFIX_REGEX, DOI_PREFIX);
+		}
+		return value;
+	}
+
 }
diff --git a/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/IdentifierFactory.java b/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/IdentifierFactory.java
index 43cdbc1..d0baec5 100644
--- a/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/IdentifierFactory.java
+++ b/src/main/java/eu/dnetlib/dhp/schema/oaf/utils/IdentifierFactory.java
@@ -164,9 +164,7 @@ public class IdentifierFactory implements Serializable {
 					.stream()
 					// filter away PIDs provided by a DS that is not considered an authority for the
 					// given PID Type
-					.filter(p -> {
-						return shouldFilterPid(collectedFrom, p, mapHandles);
-					})
+					.filter(p -> shouldFilterPid(collectedFrom, p, mapHandles))
 					.map(CleaningFunctions::normalizePidValue)
 					.filter(CleaningFunctions::pidFilter))
 			.orElse(Stream.empty());
@@ -193,13 +191,17 @@ public class IdentifierFactory implements Serializable {
 	}
 
 	private static <T extends OafEntity> String idFromPid(T entity, StructuredProperty s, boolean md5) {
+		return idFromPid(ModelSupport.getIdPrefix(entity.getClass()), s.getQualifier().getClassid(), s.getValue(), md5);
+	}
+
+	public static String idFromPid(String numericPrefix, String pidType, String pidValue, boolean md5) {
 		return new StringBuilder()
-			.append(ModelSupport.getIdPrefix(entity.getClass()))
-			.append(ID_PREFIX_SEPARATOR)
-			.append(createPrefix(s.getQualifier().getClassid()))
-			.append(ID_SEPARATOR)
-			.append(md5 ? md5(s.getValue()) : s.getValue())
-			.toString();
+				.append(numericPrefix)
+				.append(ID_PREFIX_SEPARATOR)
+				.append(createPrefix(pidType))
+				.append(ID_SEPARATOR)
+				.append(md5 ? md5(pidValue) : pidValue)
+				.toString();
 	}
 
 	// create the prefix (length = 12)