merge main

Author: john-wagster

docs/changelog/132959.yaml

@@ -0,0 +1,5 @@
+pr: 132959
+summary: Adds the `v_hamming` function for calculating the Hamming distance between two dense vectors
+area: ES|QL
+type: feature
+issues: [132056]

docs/reference/query-languages/esql/_snippets/functions/description/v_hamming.md

@@ -585,6 +585,9 @@ tests:
 - class: org.elasticsearch.xpack.esql.analysis.AnalyzerTests
   method: testNoDenseVectorFailsForMagnitude
   issue: https://github.com/elastic/elasticsearch/issues/133013
+- class: org.elasticsearch.xpack.ml.integration.TextEmbeddingQueryIT
+  method: testModelWithPrefixStrings
+  issue: https://github.com/elastic/elasticsearch/issues/133138
 
 # Examples:
 #

docs/reference/query-languages/esql/_snippets/functions/examples/v_hamming.md

@@ -262,7 +262,7 @@ public final void search(String field, float[] target, KnnCollector knnCollector
         // TODO do we need to handle nested doc counts similarly to how we handle
         // filtering? E.g. keep exploring until we hit an expected number of parent documents vs. child vectors?
         while (centroidPrefetchingIterator.hasNext()
-            && (maxVectorVisited > actualDocs || knnCollector.minCompetitiveSimilarity() == Float.NEGATIVE_INFINITY)) {
+            && (maxVectorVisited > expectedDocs || knnCollector.minCompetitiveSimilarity() == Float.NEGATIVE_INFINITY)) {
             // todo do we actually need to know the score???
             CentroidOffsetAndLength offsetAndLength = centroidPrefetchingIterator.nextPostingListOffsetAndLength();
             // todo do we need direct access to the raw centroid???, this is used for quantizing, maybe hydrating and quantizing