TimeSeries Desc Sort gets skipped with Lucene 10 upgrade (#17329)

Ensuring time series desc sort optimisation invokes searchLeaf on CtxIdxSearcher

---------

Signed-off-by: expani <[email protected]>

Author: expani

server/src/main/java/org/opensearch/search/internal/ContextIndexSearcher.java

@@ -44,6 +44,7 @@
 import org.apache.lucene.search.Collector;
 import org.apache.lucene.search.CollectorManager;
 import org.apache.lucene.search.ConjunctionUtils;
+import org.apache.lucene.search.ConstantScoreQuery;
 import org.apache.lucene.search.DocIdSetIterator;
 import org.apache.lucene.search.Explanation;
 import org.apache.lucene.search.IndexSearcher;
@@ -257,8 +258,15 @@ public void search(
 
     @Override
     public void search(Query query, Collector collector) throws IOException {
-        super.search(query, collector);
-        searchContext.bucketCollectorProcessor().processPostCollection(collector);
+        // TODO : Remove when switching to use the @org.apache.lucene.search.IndexSearcher#search(Query, CollectorManager) variant from
+        // @org.opensearch.search.query.QueryPhase#searchWithCollector which then calls the overridden
+        // search(LeafReaderContextPartition[] partitions, Weight weight, Collector collector)
+        query = collector.scoreMode().needsScores() ? rewrite(query) : rewrite(new ConstantScoreQuery(query));
+        Weight weight = createWeight(query, collector.scoreMode(), 1);
+        LeafReaderContextPartition[] partitions = (getLeafContexts() == null)
+            ? new LeafReaderContextPartition[0]
+            : getLeafContexts().stream().map(LeafReaderContextPartition::createForEntireSegment).toArray(LeafReaderContextPartition[]::new);
+        search(partitions, weight, collector);
     }
 
     public void search(
@@ -297,6 +305,7 @@ protected void search(LeafReaderContextPartition[] partitions, Weight weight, Co
                     searchLeaf(partition.ctx, partition.minDocId, partition.maxDocId, weight, collector);
                 }
             }
+            // TODO : Make this a responsibility for the callers rather than implicitly getting it done here ?
             searchContext.bucketCollectorProcessor().processPostCollection(collector);
         } catch (Throwable t) {
             searchContext.indexShard().getSearchOperationListener().onFailedSliceExecution(searchContext);