Merge pull request #11023 from IoannisPanagiotas/kmeans-relax-config-for-props

Kmeans relax config for  1d props

Author: IoannisPanagiotas

algo/src/test/java/org/neo4j/gds/kmeans/KmeansStreamConfigTest.java

@@ -72,7 +72,7 @@ void shouldFailOnInvalidPropertyValueTypes(ValueType valueType) {
         assertThatIllegalArgumentException()
             .isThrownBy(() -> streamConfig.graphStoreValidation(graphStoreMock, List.of(), List.of()))
             .withMessageContaining("Unsupported node property value type")
-            .withMessageContaining("Value type required: [DOUBLE_ARRAY] or [FLOAT_ARRAY].");
+            .withMessageContaining("Value type required: [DOUBLE], [DOUBLE_ARRAY] or [FLOAT_ARRAY].");
     }
 
     @ParameterizedTest
@@ -93,13 +93,14 @@ void shouldAcceptValidPropertyValueTypes(ValueType valueType) {
     static Stream<Arguments> validNodePropertyValueTypes() {
         return Stream.of(
             Arguments.of(ValueType.DOUBLE_ARRAY),
-            Arguments.of(ValueType.FLOAT_ARRAY)
+            Arguments.of(ValueType.FLOAT_ARRAY),
+            Arguments.of(ValueType.DOUBLE)
         );
     }
 
     static Stream<Arguments> invalidNodePropertyValueTypes() {
         return Arrays.stream(ValueType.values())
-            .filter(t -> t != ValueType.DOUBLE_ARRAY && t != ValueType.FLOAT_ARRAY)
+            .filter(t -> t != ValueType.DOUBLE_ARRAY && t != ValueType.FLOAT_ARRAY && t != ValueType.DOUBLE)
             .map(Arguments::of);
     }
 

doc/modules/ROOT/pages/algorithms/kmeans.adoc

@@ -53,6 +53,7 @@ This sampling strategy tries to spread the initial clusters more evenly so as to
 
 
 It is also possible to explicitly give the list of initial centroids to the algorithm via the `seedCentroids` parameter. In this case, the value of the `initialSampler` parameter is ignored, even if changed in the configuration.
+Note that although the algorithm supports clustering single double numbers, when  seeded through the `seedCentroids` parameter, the initial centroids must be passed as arrays e.g., [[1.5],[2.5]] for initializing  with 1.5 and 2.5 respectively.
 
 [[algorithm-k-means-considerations]]
 == Considerations

procedures/facade-api/configs/community-configs/src/main/java/org/neo4j/gds/kmeans/KmeansBaseConfig.java

@@ -69,13 +69,14 @@ default void nodePropertyTypeValidation(
         Collection<RelationshipType> selectedRelationshipTypes
     ) {
         var valueType = graphStore.nodeProperty(nodeProperty()).valueType();
-        if (valueType == ValueType.DOUBLE_ARRAY || valueType == ValueType.FLOAT_ARRAY) {
+        if (valueType == ValueType.DOUBLE_ARRAY || valueType == ValueType.FLOAT_ARRAY || valueType == ValueType.DOUBLE) {
             return;
         }
         throw new IllegalArgumentException(
             StringFormatting.formatWithLocale(
-                "Unsupported node property value type [%s]. Value type required: [%s] or [%s].",
+                "Unsupported node property value type [%s]. Value type required: [%s], [%s] or [%s].",
                 valueType,
+                ValueType.DOUBLE,
                 ValueType.DOUBLE_ARRAY,
                 ValueType.FLOAT_ARRAY
             )