VUFIND-1710 Add support for custom delimiters in topic browse

Ostensibly this just required a new TopicNormalizer, but the specific
use case added a small complication.

What want is to have our topic headings contain a UTF-8 marker (an
en-space) to act as a delimiter between different terms.  They'll be
stripped out by the new TopicNormalizer for the purposes of searching
and sorting, but VuFind will use the delimiter to display a
user-selectable delimiter when showing topics.

But, this extra delimiter would cause the lookup against the authority
index to fail, since the authority index is always space-delimited.
Further, we can't change the authority index to use the same delimiter
as topics, because the authority index is a mixture of topics and
names.

So, in addition to the new TopicNormalizer, this commit extends the
Normalizer interface to add a `headingForAuthQuery` method.  By
default this is a no-op (so existing code doesn't change), but the
TopicNormalizer uses this to replace the UTF-8 delimiter with a
regular space to allow the authority index lookup to succeed.

src/main/java/org/vufind/solr/handler/Browse.java

@@ -5,6 +5,8 @@
 import java.util.List;
 import java.util.Map;
 
+import org.vufind.util.Normalizer;
+
 /**
  * Class that performs the alphabetical browse of an index and produces a
  * {@code BrowseList} object.
@@ -26,14 +28,16 @@ public Browse(HeadingsDB headings, BibDB bibdb, AuthDB auth,
         this.maxBibListSize = maxBibListSize;
     }
 
-    private void populateItem(BrowseItem item, String fields) throws Exception
+    private void populateItem(BrowseItem item, String fields, Normalizer normalizer) throws Exception
     {
         Map<String, List<Collection<String>>> bibinfo =
             bibDB.matchingExtras(item.getHeading(), fields, maxBibListSize);
         item.setExtras(bibinfo);
         item.setCount(bibDB.recordCount(item.getHeading()));
 
-        Map<String, List<String>> authFields = authDB.getFields(item.getHeading());
+        String headingForAuthLookup = normalizer.headingForAuthQuery(item.getHeading());
+
+        Map<String, List<String>> authFields = authDB.getFields(headingForAuthLookup);
 
         List<String> seeAlsoList = new ArrayList<String>();
         for (String value : authFields.get("seeAlso")) {
@@ -63,7 +67,7 @@ public int getId(String from) throws Exception
     }
 
 
-    public BrowseList getList(int rowid, int offset, int rows, String extras)
+    public BrowseList getList(int rowid, int offset, int rows, String extras, Normalizer normalizer)
     throws Exception
     {
         BrowseList result = new BrowseList();
@@ -79,11 +83,11 @@ public BrowseList getList(int rowid, int offset, int rows, String extras)
 
             BrowseItem item = new BrowseItem(sort_key, heading);
 
-            populateItem(item, extras);
+            populateItem(item, extras, normalizer);
 
             result.add(item);
         }
 
         return result;
     }
-}
+}

src/main/java/org/vufind/solr/handler/BrowseRequestHandler.java

@@ -24,6 +24,7 @@
 import org.apache.solr.search.SolrIndexSearcher;
 import org.apache.solr.security.AuthorizationContext;
 import org.apache.solr.util.RefCounted;
+import org.vufind.util.Normalizer;
 import org.vufind.util.NormalizerFactory;
 
 /*
@@ -191,7 +192,9 @@ public void handleRequestBody(org.apache.solr.request.SolrQueryRequest req,
 
             Log.info("Browsing from: " + rowid);
 
-            BrowseList list = browse.getList(rowid, offset, rows, fields);
+            Normalizer normalizer = NormalizerFactory.getNormalizer(source.normalizer);
+
+            BrowseList list = browse.getList(rowid, offset, rows, fields, normalizer);
 
             Map<String,Object> result = new HashMap<>();
 
@@ -200,7 +203,7 @@ public void handleRequestBody(org.apache.solr.request.SolrQueryRequest req,
             result.put("startRow", rowid);
             result.put("offset", offset);
 
-            new MatchTypeResponse(from, list, rowid, rows, offset, NormalizerFactory.getNormalizer(source.normalizer)).addTo(result);
+            new MatchTypeResponse(from, list, rowid, rows, offset, normalizer).addTo(result);
 
             rsp.add("Browse", result);
         } finally {

src/main/java/org/vufind/util/Normalizer.java

@@ -23,4 +23,7 @@ public interface Normalizer
      */
     public byte[] normalize(String s);
 
-}
+    default public String headingForAuthQuery(String heading) {
+        return heading;
+    }
+}

src/main/java/org/vufind/util/TopicNormalizer.java

@@ -0,0 +1,27 @@
+package org.vufind.util;
+
+public class TopicNormalizer implements Normalizer
+{
+    private Normalizer defaultNormalizer;
+
+    public TopicNormalizer() {
+        defaultNormalizer = new ICUCollatorNormalizer();
+    }
+
+    // Separator defined by VuFind's marc.properties for topic_browse field.
+    final static String TOPIC_TERM_SEPARATOR = "\u2002";
+
+    @Override
+    public byte[] normalize(String s)
+    {
+        // Treat topic term separators as spaces when producing a sort key
+        return defaultNormalizer.normalize(s.replace(TOPIC_TERM_SEPARATOR, " "));
+    }
+
+    @Override
+    public String headingForAuthQuery(String heading) {
+        // Auth DB uses single spaces as delimiters while topics use em spaces
+        return heading.replace(TOPIC_TERM_SEPARATOR, " ");
+    }
+
+}