(pdf) Fix crash for some bold lines

This reverts commit 39ce40bfeb.

code/common/model/java/nu/marginalia/model/EdgeDomain.java

@@ -112,14 +112,6 @@ public class EdgeDomain implements Serializable {
         return topDomain;
     }
 
-    public String getDomainKey() {
-        int cutPoint = topDomain.indexOf('.');
-        if (cutPoint < 0) {
-            return topDomain;
-        }
-        return topDomain.substring(0, cutPoint).toLowerCase();
-    }
-
     /** If possible, try to provide an alias domain,
      * i.e. a domain name that is very likely to link to this one
      * */

code/common/model/test/nu/marginalia/model/EdgeDomainTest.java

@@ -8,14 +8,6 @@ import static org.junit.jupiter.api.Assertions.assertEquals;
 
 class EdgeDomainTest {
 
-    @Test
-    public void testSkepdic() throws URISyntaxException {
-        var domain = new EdgeUrl("http://www.skepdic.com/astrology.html");
-        assertEquals("skepdic", domain.getDomain().getDomainKey());
-        var domain2 = new EdgeUrl("http://skepdic.com/astrology.html");
-        assertEquals("skepdic", domain2.getDomain().getDomainKey());
-    }
-
     @Test
     public void testHkDomain() throws URISyntaxException {
         var domain = new EdgeUrl("http://l7072i3.l7c.net");

code/processes/converting-process/java/org/apache/pdfbox/text/HeadingAwarePDFTextStripper.java

@@ -2002,12 +2002,11 @@ public class HeadingAwarePDFTextStripper extends LegacyPDFStreamEngine
         float minFontWeight = Integer.MAX_VALUE;
         for (var word : line)
         {
-            int i = 0;
             for (var textPosition : word.getTextPositions())
             {
-                if (word.text.charAt(i++) == ' ') {
-                    continue;
-                }
+                // Skip empty text positions as they may have a different font
+                if (word.text.isBlank()) continue;
+
                 var font = textPosition.getFont();
                 if (font == null) continue;
                 var descriptor = font.getFontDescriptor();

code/services-application/search-service-legacy/java/nu/marginalia/search/results/UrlDeduplicator.java

@@ -61,7 +61,7 @@ public class UrlDeduplicator {
 
     private boolean limitResultsPerDomain(DecoratedSearchResultItem details) {
         final var domain = details.getUrl().getDomain();
-        final String key = domain.getDomainKey();
+        final String key = domain.toString();
 
         return keyCount.adjustOrPutValue(key, 1, 1) <= resultsPerKey;
     }

code/services-application/search-service/java/nu/marginalia/search/SearchResultClusterer.java

@@ -23,7 +23,7 @@ public class SearchResultClusterer {
     }
 
     /** No clustering, just return the results as is */
-    private static List<ClusteredUrlDetails> noOp(List<UrlDetails> results, int total) {
+    public static List<ClusteredUrlDetails> noOp(List<UrlDetails> results, int total) {
         if (results.isEmpty())
             return List.of();
 

code/services-application/search-service/java/nu/marginalia/search/SearchService.java

@@ -85,7 +85,6 @@ public class SearchService extends JoobyService {
         String emptySvg = "<svg xmlns=\"http://www.w3.org/2000/svg\"></svg>";
         jooby.get("/site/{domain}/favicon", ctx -> {
             String domain = ctx.path("domain").value();
-            logger.info("Finding icon for domain {}", domain);
             try {
                 DbDomainQueries.DomainIdWithNode domainIdWithNode = domainQueries.getDomainIdWithNode(new EdgeDomain(domain));
                 var faviconMaybe = faviconClient.getFavicon(domain, domainIdWithNode.nodeAffinity());

code/services-application/search-service/java/nu/marginalia/search/results/UrlDeduplicator.java

@@ -25,10 +25,11 @@ public class UrlDeduplicator {
     }
 
     public boolean shouldRemove(DecoratedSearchResultItem details) {
-//        if (!deduplicateOnSuperficialHash(details))
-//            return true;
-//        if (!deduplicateOnLSH(details))
-//            return true;
+
+        if (!deduplicateOnSuperficialHash(details))
+            return true;
+        if (!deduplicateOnLSH(details))
+            return true;
         if (!limitResultsPerDomain(details))
             return true;
 
@@ -61,7 +62,7 @@ public class UrlDeduplicator {
 
     private boolean limitResultsPerDomain(DecoratedSearchResultItem details) {
         final var domain = details.getUrl().getDomain();
-        final String key = domain.getDomainKey();
+        final String key = domain.toString();
 
         return keyCount.adjustOrPutValue(key, 1, 1) <= resultsPerKey;
     }

deploy.txt

@@ -8,3 +8,4 @@
 2025-05-05:  Deploy executor partition 4.
 2025-05-05:  Deploy control.
 2025-05-08:  Deploy assistant.
+2025-05-17:  Redeploy all.