issue #174: PICA FRBR functions

pkiraly · Oct 5, 2022 · 6f8712d · 6f8712d
1 parent db64588
commit 6f8712d
Show file tree

Hide file tree

Showing 2 changed files with 33 additions and 5 deletions.
diff --git a/src/main/java/de/gwdg/metadataqa/marc/cli/FunctionalAnalysis.java b/src/main/java/de/gwdg/metadataqa/marc/cli/FunctionalAnalysis.java
@@ -47,7 +47,7 @@ public FunctionalAnalysis(String[] args) throws ParseException {
     parameters = new CompletenessParameters(args);
     options = parameters.getOptions();
     readyToProcess = true;
-    frbrFunctionLister = new FrbrFunctionLister(parameters.getMarcVersion());
+    frbrFunctionLister = new FrbrFunctionLister(parameters.getSchemaType(), parameters.getMarcVersion());
 
     logger.info(frbrFunctionLister.getBaseline().toString());
   }
@@ -212,7 +212,7 @@ else if (parameters.getSchemaType().equals(SchemaType.PICA))
       writer.write("frbrfunction" + separator + "count" + separator + "fields\n");
       for (FRBRFunction function : FRBRFunction.values()) {
         if (function.getParent() != null) {
-          List<String> paths = functions.get(function);
+          List<String> paths = functions.getOrDefault(function, new ArrayList<>());
           List<Object> cells = new ArrayList<>();
           cells.add(function.toString());
           cells.add(paths.size());

diff --git a/src/main/java/de/gwdg/metadataqa/marc/utils/FrbrFunctionLister.java b/src/main/java/de/gwdg/metadataqa/marc/utils/FrbrFunctionLister.java
@@ -2,6 +2,7 @@
 
 import de.gwdg.metadataqa.marc.Utils;
 import de.gwdg.metadataqa.marc.definition.*;
+import de.gwdg.metadataqa.marc.definition.bibliographic.SchemaType;
 import de.gwdg.metadataqa.marc.definition.structure.ControlFieldDefinition;
 import de.gwdg.metadataqa.marc.definition.structure.ControlfieldPositionDefinition;
 import de.gwdg.metadataqa.marc.definition.structure.DataFieldDefinition;
@@ -25,10 +26,11 @@
 public class FrbrFunctionLister {
 
   private static final Logger logger = Logger.getLogger(FrbrFunctionLister.class.getCanonicalName());
+  private SchemaType schemaType; // = SchemaType.MARC21;
+  private MarcVersion marcVersion; // = MarcVersion.MARC21;
 
   private Counter<FRBRFunction> baselineCounter = new Counter<>();
   private int elementsWithoutFunctions;
-  private MarcVersion marcVersion;
 
   private Map<FRBRFunction, FunctionValue> collector;
   private Map<FRBRFunction, Counter<FunctionValue>> histogram;
@@ -38,14 +40,23 @@ public class FrbrFunctionLister {
   private AppendableHashMap<FRBRFunction, String> marcPathByFunction;
   private Map<FRBRFunction, List<String>> picaPathByFunction;
 
-  public FrbrFunctionLister(MarcVersion marcVersion) {
-    this.marcVersion = marcVersion;
+  public FrbrFunctionLister(SchemaType schemaType, MarcVersion marcVersion) {
+    this.schemaType = schemaType == null ? SchemaType.MARC21 : schemaType;
+    this.marcVersion = marcVersion == null ? MarcVersion.MARC21 : marcVersion;
 
     prepareBaseline();
     prepareCollector();
     prepareHistogram();
   }
 
+  public FrbrFunctionLister(MarcVersion marcVersion) {
+    this(SchemaType.MARC21, marcVersion);
+  }
+
+  public FrbrFunctionLister(SchemaType schemaType) {
+    this(schemaType, MarcVersion.MARC21);
+  }
+
   public Map<FRBRFunction, Counter<FunctionValue>> getHistogram() {
     return histogram;
   }
@@ -55,6 +66,23 @@ public void prepareBaseline() {
     functionByMarcPath = new TreeMap<>();
     marcPathByFunction = new AppendableHashMap<>();
 
+    if (schemaType.equals(SchemaType.MARC21)) {
+      prepareBaselineForMarc21();
+    } else if (schemaType.equals(SchemaType.PICA)) {
+      prepareBaselineForMarc21();
+      initializePica();
+      functionByMarcPath = new TreeMap<>();
+      marcPathByFunction = new AppendableHashMap<>();
+      prepareBaselineForPica();
+    }
+  }
+
+  private void prepareBaselineForPica() {
+    for (Map.Entry<String, List<FRBRFunction>> entry : functionByPicaPath.entrySet())
+      registerFunctions(entry.getValue(), entry.getKey());
+  }
+
+  private void prepareBaselineForMarc21() {
     for (ControlfieldPositionDefinition subfield : MarcDefinition.getLeaderPositions())
       registerFunctions(subfield.getFrbrFunctions(), subfield.getPath(false));