[S] Change in asterixdb[master]: [ASTERIXDB-3817][STO] Reject an unknown metric in the quantization codec

[AsterixDB Code Review <do...@asterix-gerrit.ics.uci.edu>]

From Hongyu Shi <le...@gmail.com>:

Hongyu Shi has uploaded this change for review. ( https://asterix-gerrit.ics.uci.edu/c/asterixdb/+/21867?usp=email )


Change subject: [ASTERIXDB-3817][STO] Reject an unknown metric in the quantization codec
......................................................................

[ASTERIXDB-3817][STO] Reject an unknown metric in the quantization codec

- user model changes: no
- storage format changes: no
- interface changes: no

Details:
fromDistanceMetric fell back to DOT_PRODUCT for a metric it did not recognize. The fallback is removed; an unknown metric is an error that names the value.

Change-Id: Ie91a7b27265a9259e1ec6ebdb96f46a1c700bf3b
---
M asterixdb/asterix-common/src/main/java/org/apache/asterix/common/vector/OptimizedScalarQuantizationCodec.java
1 file changed, 6 insertions(+), 6 deletions(-)



  git pull ssh://asterix-gerrit.ics.uci.edu:29418/asterixdb refs/changes/67/21867/1

diff --git a/asterixdb/asterix-common/src/main/java/org/apache/asterix/common/vector/OptimizedScalarQuantizationCodec.java b/asterixdb/asterix-common/src/main/java/org/apache/asterix/common/vector/OptimizedScalarQuantizationCodec.java
index ed6b51e..661eb09 100644
--- a/asterixdb/asterix-common/src/main/java/org/apache/asterix/common/vector/OptimizedScalarQuantizationCodec.java
+++ b/asterixdb/asterix-common/src/main/java/org/apache/asterix/common/vector/OptimizedScalarQuantizationCodec.java
@@ -78,15 +78,15 @@
      * Converts a distance metric string to SimilarityFunction enum.
      * 
      * @param distanceMetric Distance metric string (e.g., "euclidean", "cosine", "dot")
-     * @return Corresponding SimilarityFunction enum, or DOT_PRODUCT as default
+     * @return Corresponding SimilarityFunction enum
+     * @throws HyracksDataException if the metric is not one a vector index can be created with
      */
-    public static SimilarityFunction fromDistanceMetric(String distanceMetric) {
-        // Resolve the metric name through VectorSimilarityMetric, the single source of truth for vector
-        // metric aliases (it handles null/blank/casing). A null (unrecognized or blank) metric falls back
-        // to DOT_PRODUCT, preserving the previous default for unknown metrics.
+    public static SimilarityFunction fromDistanceMetric(String distanceMetric) throws HyracksDataException {
         VectorSimilarityMetric metric = VectorSimilarityMetric.fromAlias(distanceMetric);
         if (metric == null) {
-            return SimilarityFunction.DOT_PRODUCT;
+            // The codes must be built under the metric the index's distance function ranks by.
+            throw new RuntimeDataException(ErrorCode.ILLEGAL_STATE,
+                    "Unknown vector distance metric '" + distanceMetric + "' reached the quantizer");
         }
         return switch (metric) {
             case EUCLIDEAN -> SimilarityFunction.EUCLIDEAN;

-- 
To view, visit https://asterix-gerrit.ics.uci.edu/c/asterixdb/+/21867?usp=email
To unsubscribe, or for help writing mail filters, visit https://asterix-gerrit.ics.uci.edu/settings?usp=email

Gerrit-MessageType: newchange
Gerrit-Project: asterixdb
Gerrit-Branch: master
Gerrit-Change-Id: Ie91a7b27265a9259e1ec6ebdb96f46a1c700bf3b
Gerrit-Change-Number: 21867
Gerrit-PatchSet: 1
Gerrit-Owner: Hongyu Shi <le...@gmail.com>