(converter) Ensure features is mutable on DetailsWithWords as this is assumed later

code/processes/converting-process/ft-dom-classifier/resources/request-classifier.xml

@@ -15,6 +15,7 @@
     <classifier target="url-regex" rule="tracking">/ccm/collect$</classifier>
     <classifier target="url-regex" rule="tracking">^/[0-9]+\.js$</classifier>
     <classifier target="url-regex" rule="tracking">^/[a-z0-9]\.gif$</classifier>
+    <classifier target="url-regex" rule="tracking">^/pixel\.gif$</classifier>
     <classifier target="url-regex" rule="ads">/pagead/</classifier>
     <classifier target="url-regex" rule="ads">/google-ads/</classifier>
 

code/processes/converting-process/java/nu/marginalia/converting/processor/plugin/HtmlDocumentProcessorPlugin.java

@@ -161,7 +161,6 @@ public class HtmlDocumentProcessorPlugin extends AbstractDocumentProcessorPlugin
 
         final Set<HtmlFeature> features = featureExtractor.getFeatures(url, doc, documentHeaders, dld);
 
-
         if (!documentLengthLogic.validateLength(dld, specialization.lengthModifier() * documentClass.lengthLimitModifier())) {
             features.add(HtmlFeature.SHORT_DOCUMENT);
         }

code/processes/converting-process/java/nu/marginalia/converting/processor/plugin/PdfDocumentProcessorPlugin.java

@@ -115,7 +115,9 @@ public class PdfDocumentProcessorPlugin extends AbstractDocumentProcessorPlugin
 
         ret.quality = -5;
 
-        ret.features = Set.of(HtmlFeature.PDF);
+        ret.features = new HashSet<>(); // must be mutable!
+        ret.features.add(HtmlFeature.PDF);
+
         ret.description = getDescription(doc);
         ret.hashCode = dld.localitySensitiveHashCode();