More cleanup

Signed-off-by: Atri Sharma <atrisharma@Atris-Mac-Studio.local>

Author: Atri Sharma

server/src/main/java/org/opensearch/action/search/StreamSearchQueryThenFetchAsyncAction.java

@@ -152,7 +152,7 @@ protected void onStreamResult(SearchPhaseResult result, SearchShardIterator shar
      */
     @Override
     protected void onShardResult(SearchPhaseResult result, SearchShardIterator shardIt) {
-        // Safety log: track final shard response receipt in coordinator
+        // Trace final shard responses to diagnose coordinator sequencing.
         if (logger.isTraceEnabled()) {
             logger.trace(
                 "COORDINATOR: received final shard result from shard={}, target={}, totalOps={}, expectedOps={}",
@@ -196,12 +196,11 @@ void successfulShardExecution(SearchShardIterator shardsIt) {
 
     /**
      * Handle successful stream execution callback
-     * Since partials are no longer fed into the reducer, this callback is not
-     * needed for coordination.
+     * Partials are not fed into the reducer, so coordinator completion is driven
+     * by {@link #successfulShardExecution(SearchShardIterator)}.
      */
     private void successfulStreamExecution() {
-        // No-op: partials are bypassed from reducer, completion is handled by
-        // successfulShardExecution only
+        // No-op.
     }
 
 }

server/src/main/java/org/opensearch/search/aggregations/bucket/terms/StreamStringTermsAggregator.java

@@ -514,9 +514,7 @@ public StreamingCostMetrics estimateStreamingCost(SearchContext context) {
                 return StreamingCostMetrics.nonStreamable();
             }
 
-            // Calculate metrics (simplified for now to match StreamNumericTermsAggregator
-            // pattern essentially)
-            // But we need total bucket count from ordinals
+            // Estimate streaming cost using segment ordinals.
             ValuesSource.Bytes.WithOrdinals ordinalValuesSource = (ValuesSource.Bytes.WithOrdinals) valuesSource;
             List<LeafReaderContext> leaves = context.searcher().getIndexReader().leaves();
             long maxCardinality = 0;

server/src/main/java/org/opensearch/search/query/StreamingSearchMode.java

@@ -11,38 +11,28 @@
 import org.opensearch.common.annotation.ExperimentalApi;
 
 /**
- * Defines the different streaming search strategies based on the design-by-case approach.
- * Each mode optimizes for different use cases and performance characteristics.
+ * Streaming search mode selector.
+ *
+ * <p>Only {@link #NO_SCORING} is active in the current streaming implementation.
+ * Other values are retained for request parsing/backward compatibility.
  *
  * @opensearch.internal
  */
 @ExperimentalApi
 public enum StreamingSearchMode {
 
     /**
-     * Case 1: No scoring, no sorting - fastest TTFB
-     * - Shard collector: StreamingUnsortedCollector
-     * - Ring buffer with batch emission
-     * - Round-robin merge at coordinator
-     * - Best for: simple filtering, counting, exists queries
+     * No scoring and no sorting.
      */
     NO_SCORING("no_scoring"),
 
     /**
-     * Case 2: Full scoring + explicit sort - production ready
-     * - Shard collector: StreamingSortedCollector
-     * - WAND/Block-Max WAND with windowed top-K heap
-     * - K-way streaming merge at coordinator
-     * - Best for: scored searches with sorting
+     * Retained for backward compatibility.
      */
     SCORED_SORTED("scored_sorted"),
 
     /**
-     * Case 3: Full scoring, no sorting - moderate performance
-     * - Shard collector: StreamingScoredUnsortedCollector
-     * - Ring buffer with scoring
-     * - No merge needed at coordinator
-     * - Best for: scored searches without sorting
+     * Retained for backward compatibility.
      */
     SCORED_UNSORTED("scored_unsorted");
 
@@ -98,8 +88,6 @@ public boolean requiresSorting() {
         return false; // No sorting is supported in the current mode
     }
 
-    // Confidence-based mode removed in this branch
-
     @Override
     public String toString() {
         return value;