Improve handling of ad-hoc OBO namespaces in ShortFormAnnotator

.github/workflows/test.yml

@@ -3,15 +3,19 @@ on:
   push:
     branches:
       - dev
+  pull_request:
+    branches:
+      - dev
+
 jobs:
   testDataload:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v3
-      - name: Set up JDK 11
+      - name: Set up Java
         uses: actions/setup-java@v3
         with:
-          java-version: '14'
+          java-version: '25'
           distribution: 'adopt'
       - name: Build with Maven
         run: mvn clean package
@@ -23,5 +27,3 @@ jobs:
       - uses: actions/checkout@v3
       - name: Run API tests
         run: ./test_api.sh
-
-

dataload/rdf2json/dependency-reduced-pom.xml

@@ -53,6 +53,26 @@
       <type>pom</type>
       <scope>compile</scope>
     </dependency>
+    <dependency>
+      <groupId>org.junit.jupiter</groupId>
+      <artifactId>junit-jupiter</artifactId>
+      <version>5.10.2</version>
+      <scope>test</scope>
+      <exclusions>
+        <exclusion>
+          <artifactId>junit-jupiter-api</artifactId>
+          <groupId>org.junit.jupiter</groupId>
+        </exclusion>
+        <exclusion>
+          <artifactId>junit-jupiter-params</artifactId>
+          <groupId>org.junit.jupiter</groupId>
+        </exclusion>
+        <exclusion>
+          <artifactId>junit-jupiter-engine</artifactId>
+          <groupId>org.junit.jupiter</groupId>
+        </exclusion>
+      </exclusions>
+    </dependency>
   </dependencies>
   <properties>
     <maven.compiler.target>11</maven.compiler.target>

dataload/rdf2json/pom.xml

@@ -44,6 +44,13 @@
             <artifactId>logback-classic</artifactId>
             <version>1.4.11</version>
         </dependency>
+
+        <dependency>
+            <groupId>org.junit.jupiter</groupId>
+            <artifactId>junit-jupiter</artifactId>
+            <version>5.10.2</version>
+            <scope>test</scope>
+        </dependency>
     </dependencies>
 
     <properties>

dataload/rdf2json/src/main/java/uk/ac/ebi/rdf2json/annotators/ShortFormAnnotator.java

@@ -12,6 +12,7 @@
 
 public class ShortFormAnnotator {
 	private static final Logger logger = LoggerFactory.getLogger(ShortFormAnnotator.class);
+	private static final String oboPurlPrefix = "http://purl.obolibrary.org/obo/";
 
 	public static void annotateShortForms(OntologyGraph graph) {
 
@@ -36,7 +37,7 @@ public static void annotateShortForms(OntologyGraph graph) {
 				preferredPrefix = graph.config.get("id").toString().toUpperCase();
 			}
 
-			String shortForm = extractShortForm(graph, ontologyBaseUris, preferredPrefix, c.uri);
+			String shortForm = extractShortForm(ontologyBaseUris, preferredPrefix, c.uri);
 
 			/*
 			CURIEs are formed by following rules:
@@ -75,16 +76,26 @@ public static void annotateShortForms(OntologyGraph graph) {
 
 	}
 
-	private static String extractShortForm(OntologyGraph graph, Set<String> ontologyBaseUris, String preferredPrefix,
+	public static String extractShortForm(Set<String> ontologyBaseUris, String preferredPrefix,
 			String uri) {
 
 		if (uri.startsWith("urn:")) {
 			return uri.substring(4);
 		}
 
-		// if(uri.startsWith("http://purl.obolibrary.org/obo/")) {
-		// return uri.substring("http://purl.obolibrary.org/obo/".length());
-		// }
+		// Check if it's an ad-hoc member of an OBO space, for example, how
+		// http://obolibrary.org/obo/mesh#C is an ad-hoc member of the OBO
+		// mesh space or http://purl.obolibrary.org/obo/chebi#mass as an ad-hoc
+		// member of the OBO chebi space (not the same as the CHEBI space).
+		//
+		// For example, http://purl.obolibrary.org/obo/chebi#mass becomes 
+		// obo:chebi#mass
+		//
+		// See further discussion at https://github.com/EBISPOT/ols4/issues/935.
+		String oboAdHocURIPrefix = oboPurlPrefix + preferredPrefix.toLowerCase() + "#";
+		if (uri.startsWith(oboAdHocURIPrefix)) {
+			return "obo:" + uri.substring(oboPurlPrefix.length());
+		}
 
 		for (String baseUri : ontologyBaseUris) {
 			if (uri.startsWith(baseUri) && preferredPrefix != null) {

dataload/rdf2json/src/test/java/TestShortFormAnnotator.java

@@ -0,0 +1,41 @@
+import org.junit.jupiter.api.Test;
+import static uk.ac.ebi.rdf2json.annotators.ShortFormAnnotator.extractShortForm;
+import java.util.*;
+import static org.junit.jupiter.api.Assertions.assertEquals;
+
+public class TestShortFormAnnotator {
+    static Set<String> uriPrefixes = new HashSet<>() {{
+        add("http://purl.obolibrary.org/obo/TEST_");
+    }};
+
+    @Test
+    public void testURN() {
+        // Test slicing off front of URN
+        assertEquals("garbage", extractShortForm(uriPrefixes, "TEST", "urn:garbage"));
+    }
+
+    @Test
+    public void testOBODefaultNamespace() {
+        // Test OBO default namespaces (added in https://github.com/EBISPOT/ols4/pull/937)
+        assertEquals("obo:test#abc", extractShortForm(uriPrefixes, "TEST", "http://purl.obolibrary.org/obo/test#abc"));
+
+        // Reverts to the guess workflow
+        assertEquals("abc", extractShortForm(uriPrefixes, "XXX", "http://purl.obolibrary.org/obo/test#abc"));
+    }
+
+    @Test
+    public void testRegular() {
+        // Test regular parsing based, note that it takes the prefix given
+        assertEquals("XXX_1234567", extractShortForm(uriPrefixes, "XXX", "http://purl.obolibrary.org/obo/TEST_1234567"));
+        assertEquals("TEST_1234567", extractShortForm(uriPrefixes, "TEST", "http://purl.obolibrary.org/obo/TEST_1234567"));
+    }
+
+    @Test
+    public void testGuesses() {
+        // Guesses
+        assertEquals("1234567", extractShortForm(uriPrefixes, "XXX", "http://example.org/1234567"));
+        assertEquals("1234567", extractShortForm(uriPrefixes, "TEST", "http://example.org/1234567"));
+        assertEquals("1234567", extractShortForm(uriPrefixes, "XXX", "http://example.org/any.html#1234567"));
+        assertEquals("1234567", extractShortForm(uriPrefixes, "TEST", "http://example.org/any.html#1234567"));
+    }
+}