wip

ramls/settings/subjectSource.json

@@ -10,6 +10,10 @@
       "description": "label for the subject source",
       "type": "string"
     },
+    "code": {
+      "description": "label for the subject source code",
+      "type": "string"
+    },
     "source": {
       "type": "string",
       "description": "label indicating where the subject source entry originates from, i.e. 'folio' or 'local'",

src/main/java/org/folio/processing/mapping/defaultmapper/processor/functions/NormalizationFunction.java

@@ -442,6 +442,25 @@ public String apply(RuleExecutionContext context) {
     }
   },
 
+  SET_SUBJECT_SOURCE_ID_BY_CODE() {
+    @Override
+    public String apply(RuleExecutionContext context) {
+      List<SubjectSource> subjectSources = context.getMappingParameters().getSubjectSources();
+      String sourceCode = context.getSubFieldValue();
+
+      if (sourceCode == null || subjectSources == null) {
+        return StringUtils.EMPTY;
+      }
+
+      return subjectSources.stream()
+        .filter(subjectSource -> subjectSource.getCode() != null
+          && subjectSource.getCode().equalsIgnoreCase(sourceCode)) // todo: should we compare ignoring case, should we trim?
+        .findFirst()
+        .map(SubjectSource::getId)
+        .orElse(StringUtils.EMPTY);
+    }
+  },
+
   SET_SUBJECT_TYPE_ID() {
     private static final String NAME_PARAMETER = "name";
 

src/test/java/org/folio/processing/mapping/InstanceMappingTest.java

@@ -6,11 +6,14 @@
 import static org.junit.Assert.assertNull;
 import static org.junit.Assert.assertTrue;
 
+import com.fasterxml.jackson.core.type.TypeReference;
 import com.fasterxml.jackson.databind.ObjectMapper;
+import io.vertx.core.json.Json;
 import io.vertx.core.json.JsonArray;
 import io.vertx.core.json.JsonObject;
 import java.io.ByteArrayInputStream;
 import java.io.ByteArrayOutputStream;
+import java.io.File;
 import java.io.IOException;
 import java.nio.charset.StandardCharsets;
 import java.util.ArrayList;
@@ -69,6 +72,7 @@ public class InstanceMappingTest {
   private static final String BIB_WITH_720_FIELDS = "src/test/resources/org/folio/processing/mapping/instance/720_fields_samples.mrc";
   private static final String BIB_WITH_FIELDS_FOR_ALTERNATIVE_MAPPING = "src/test/resources/org/folio/processing/mapping/instance/fields_for_alternative_mapping_samples.mrc";
   private static final String BIB_WITH_FIELDS_FOR_ALTERNATIVE_MAPPING_WITH_PUNCTUATIONS = "src/test/resources/org/folio/processing/mapping/instance/fields_for_alternative_mapping_samples_with_punctuations.mrc";
+  public static final String BIB_WITH_SUBJECT_SOURCES_CODE_IN_2_SUBFIELD = "src/test/resources/org/folio/processing/mapping/instance/subject_source_codes_in_2_subfield.mrc";
   private static final String CLASSIFICATIONS_TEST = "src/test/resources/org/folio/processing/mapping/instance/classificationsTest.mrc";
   private static final String INSTANCES_CLASSIFICATIONS_PATH = "src/test/resources/org/folio/processing/mapping/instance/classificationsTestInstance.json";
   private static final String DEFAULT_MAPPING_RULES_PATH = "src/test/resources/org/folio/processing/mapping/instance/rules.json";
@@ -594,6 +598,58 @@ public void testMarcToInstanceWithRepeatableSubjectsButWithoutIndicators() throw
     });
   }
 
+  @Test
+  public void testMarcToSubjectSourceMappingByCodeFromSubfield() throws IOException {
+    MarcReader reader = new MarcStreamReader(new ByteArrayInputStream(
+      TestUtil.readFileFromPath(BIB_WITH_SUBJECT_SOURCES_CODE_IN_2_SUBFIELD).getBytes(StandardCharsets.UTF_8)));
+
+    JsonObject mappingRules = new JsonObject(TestUtil.readFileFromPath(DEFAULT_MAPPING_RULES_PATH));
+    List<SubjectSource> subjectSources = new ObjectMapper()
+      .readValue(new File(DEFAULT_SUBJECT_SOURCES_PATH), new TypeReference<>() {});
+
+    final String FIRST_LIBRARY_SOURCE_ID = "e894d0dc-621d-4b1d-98f6-6f7120eb0d40";
+    final String SECOND_LIBRARY_SOURCE_ID = "e894d0dc-621d-4b1d-98f6-6f7120eb0d41";
+    final String THIRD_LIBRARY_SOURCE_ID = "e894d0dc-621d-4b1d-98f6-6f7120eb0d42";
+    final String FOURTH_LIBRARY_SOURCE_ID = "e894d0dc-621d-4b1d-98f6-6f7120eb0d45";
+    final String FIFTH_LIBRARY_SOURCE_ID = "e894d0dc-621d-4b1d-98f6-6f7120eb0d46";
+
+    Map<String, String> subjectValueToSourceId = Map.of(
+      "Subject heading 600", FIRST_LIBRARY_SOURCE_ID,
+      "Subject heading 610", SECOND_LIBRARY_SOURCE_ID,
+      "Subject heading 611", THIRD_LIBRARY_SOURCE_ID,
+      "Subject heading 630", FOURTH_LIBRARY_SOURCE_ID,
+      "Subject heading 647", FIFTH_LIBRARY_SOURCE_ID,
+      "Subject heading 648", FIRST_LIBRARY_SOURCE_ID,
+      "Subject heading 650", SECOND_LIBRARY_SOURCE_ID,
+      "Subject heading 651", THIRD_LIBRARY_SOURCE_ID,
+      "Subject heading 655", FOURTH_LIBRARY_SOURCE_ID
+    );
+
+    ValidatorFactory factory = Validation.buildDefaultValidatorFactory();
+    while (reader.hasNext()) {
+      ByteArrayOutputStream os = new ByteArrayOutputStream();
+      MarcJsonWriter writer = new MarcJsonWriter(os);
+      Record record = reader.next();
+      writer.write(record);
+      JsonObject marc = new JsonObject(os.toString());
+      Instance instance = mapper.mapRecord(marc, new MappingParameters().withSubjectSources(subjectSources), mappingRules);
+
+      assertNotNull(instance.getSubjects());
+      assertEquals(9, instance.getSubjects().size());
+      instance.getSubjects().forEach(subject -> {
+        assertNotNull(subject.getValue());
+        assertEquals(subjectValueToSourceId.get(subject.getValue()), subject.getSourceId());
+      });
+
+
+
+//      Validator validator = factory.getValidator();
+//      Set<ConstraintViolation<Instance>> violations = validator.validate(instance);
+//      assertTrue(violations.isEmpty());
+
+    }
+  }
+
   @Test
   public void testMarc720ToInstanceContributors() throws IOException {
     MarcReader reader = new MarcStreamReader(new ByteArrayInputStream(TestUtil.readFileFromPath(BIB_WITH_720_FIELDS)

src/test/resources/org/folio/processing/mapping/instance/subjectSources.json

@@ -1,14 +1,18 @@
-[ {
+[
+  {
     "id" : "e894d0dc-621d-4b1d-98f6-6f7120eb0d40",
     "name" : "Library of Congress Subject Headings",
+    "code" : "lcsh",
     "source" : "folio"
   }, {
     "id" : "e894d0dc-621d-4b1d-98f6-6f7120eb0d41",
     "name" : "Library of Congress Children's and Young Adults' Subject Headings",
+    "code" : "cyac",
     "source" : "folio"
   }, {
     "id" : "e894d0dc-621d-4b1d-98f6-6f7120eb0d42",
     "name" : "Medical Subject Headings",
+    "code" : "mesh",
     "source" : "folio"
   }, {
     "id" : "e894d0dc-621d-4b1d-98f6-6f7120eb0d43",
@@ -21,13 +25,12 @@
   }, {
     "id" : "e894d0dc-621d-4b1d-98f6-6f7120eb0d45",
     "name" : "Canadian Subject Headings",
+    "code" : "cash",
     "source" : "folio"
   }, {
     "id" : "e894d0dc-621d-4b1d-98f6-6f7120eb0d46",
     "name" : "Répertoire de vedettes-matière",
+    "code" : "rvm",
     "source" : "folio"
-  }, {
-    "id" : "e894d0dc-621d-4b1d-98f6-6f7120eb0d47",
-    "name" : "Source specified in subfield $2",
-    "source" : "folio"
-  } ]
+  }
+]

src/test/resources/org/folio/processing/mapping/instance/subject_source_codes_in_2_subfield.mrc

@@ -0,0 +1 @@
+01001nam a2200301   4500001000800000003000400008005001700012008004100029010001600070020001500086035001900101035001800120040001300138041000800151050002600159082001400185100001800199245010000217250002300317260005600340300003000396350001100426500003500437504003200472600003200504600008000536655008300616in23605ICU19971010231800.0741204r1975    nyu      b    00000 eng u  a   74011027  a0151238154  a(ICU)BID297516  a(OCoLC)948019  aDLCcDLC0 aeng0 aPT2613.R338bZ78 1975  a833/.9/140 aReddick, John14aThe 'Danzig trilogy' of Günter Grass;ba study of The tin drum, Cat and mouse, and Dog years.  a[1st American ed.]0 aNew York,bHarcourt Brace Jovanovichc[1975, c1974]  aviii, 289, [1] p.c22 cm.  a$13.50  a"A Helen and Kurt Wolff book."  aBibliography: p. 287-[290].17aGrass, Günter,d1927-201517aGrass, Günter,d1927-2015.2fast0http://id.worldcat.org/fast/fst00057296 7aCriticism, interpretation, etc.2fast0http://id.worldcat.org/fast/fst01411635