Copilot commented on code in PR #18090:
URL: https://github.com/apache/pinot/pull/18090#discussion_r3036852768
##########
pinot-segment-local/src/test/java/org/apache/pinot/segment/local/indexsegment/mutable/MutableSegmentImplTest.java:
##########
@@ -227,6 +230,29 @@ public void testUpdateIngestionTimestampWithoutIndexing() {
}
}
+ @Test
+ public void
testVectorIndexConfigOnMutableSegmentWithoutMutableVectorReader() {
+ Schema vectorSchema = new
Schema.SchemaBuilder().setSchemaName("vectorSchema")
+ .addSingleValueDimension("id", FieldSpec.DataType.INT)
+ .addMultiValueDimension("embedding", FieldSpec.DataType.FLOAT)
+ .build();
+ VectorIndexConfig vectorIndexConfig =
+ new VectorIndexConfig(false, "IVF_PQ", 4, 1,
VectorIndexConfig.VectorDistanceFunction.COSINE,
+ Map.of("nlist", "1", "pq_m", "2", "pq_nbits", "8"));
+ MutableSegmentImpl mutableSegment =
Review Comment:
The IVF_PQ config used in this test uses non-canonical property keys
(`pq_m`, `pq_nbits`) and omits required IVF_PQ properties like
`trainSampleSize` (and `pqM`/`pqNbits` in the expected camelCase form). This
makes the test config inconsistent with `VectorIndexConfigValidator`/docs and
could cause the test to become invalid as soon as config validation is enforced
for mutable segments.
Consider updating the properties map to use the supported keys (`nlist`,
`pqM`, `pqNbits`, `trainSampleSize`, optional `trainingSeed`) so the test
reflects a realistic table config.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
