Merge branch 'main' into data-stream-mappings-no-doc

Author: masseyke

server/src/main/java/org/elasticsearch/index/codec/vectors/DefaultIVFVectorsReader.java

@@ -18,7 +18,7 @@
 import org.apache.lucene.util.ArrayUtil;
 import org.apache.lucene.util.Bits;
 import org.apache.lucene.util.VectorUtil;
-import org.apache.lucene.util.hnsw.NeighborQueue;
+import org.elasticsearch.index.codec.vectors.cluster.NeighborQueue;
 import org.elasticsearch.index.codec.vectors.reflect.OffHeapStats;
 import org.elasticsearch.simdvec.ES91OSQVectorsScorer;
 import org.elasticsearch.simdvec.ES92Int7VectorsScorer;
@@ -243,22 +243,19 @@ public boolean hasNext() {
 
             @Override
             public CentroidOffsetAndLength nextPostingListOffsetAndLength() throws IOException {
-                int centroidOrdinal = neighborQueue.pop();
-                updateQueue(); // add one children if available so the queue remains fully populated
+                int centroidOrdinal = nextCentroid();
                 centroids.seek(childrenFileOffsets + (long) Long.BYTES * 2 * centroidOrdinal);
                 long postingListOffset = centroids.readLong();
                 long postingListLength = centroids.readLong();
                 return new CentroidOffsetAndLength(postingListOffset, postingListLength);
             }
 
-            private void updateQueue() throws IOException {
+            private int nextCentroid() throws IOException {
                 if (currentParentQueue.size() > 0) {
-                    // add a children from the current parent queue
-                    float score = currentParentQueue.topScore();
-                    int children = currentParentQueue.pop();
-                    neighborQueue.add(children, score);
+                    // return next centroid and maybe add a children from the current parent queue
+                    return neighborQueue.popAndAddRaw(currentParentQueue.popRaw());
                 } else if (parentsQueue.size() > 0) {
-                    // add a new parent from the current parent queue
+                    // current parent queue is empty, populate it again with the next parent
                     int pop = parentsQueue.pop();
                     populateOneChildrenGroup(
                         currentParentQueue,
@@ -273,7 +270,9 @@ private void updateQueue() throws IOException {
                         globalCentroidDp,
                         scores
                     );
-                    updateQueue();
+                    return nextCentroid();
+                } else {
+                    return neighborQueue.pop();
                 }
             }
         };

server/src/main/java/org/elasticsearch/index/codec/vectors/cluster/NeighborQueue.java

@@ -137,6 +137,25 @@ public int pop() {
         return decodeNodeId(heap.pop());
     }
 
+    /** Removes the top element and returns it */
+    public long popRaw() {
+        return heap.pop();
+    }
+
+    /**
+     * if the new element is the new top then return its node id. Otherwise,
+     * removes the current top element, returns its node id and adds the new element
+     * to the queue.
+     * */
+    public int popAndAddRaw(long raw) {
+        long top = heap.top();
+        if (raw < top) {
+            return decodeNodeId(raw);
+        }
+        heap.updateTop(raw);
+        return decodeNodeId(top);
+    }
+
     public void clear() {
         heap.clear();
     }