commit 828f9feca37cbec169df7f1b6d68e277f0952d35
author Orhan Kislal <okislal@apache.org> Wed Oct 02 19:09:12 2019 -0400
committer kaknikhil <kaknikhil@users.noreply.github.com> Mon Oct 07 14:14:02 2019 -0700
tree a52406beaf25144e07992d56ce3fc40ad3c93c1e
parent 1b5ba4afd58ca9b263ccac47769ef281b45e3466

Kmeans: Use squared dist norm2 as a default for silh functions

JIRA: MADLIB-1383

src/ports/postgres/modules/kmeans/kmeans.sql_in

diff --git a/src/ports/postgres/modules/kmeans/kmeans.sql_in b/src/ports/postgres/modules/kmeans/kmeans.sql_in
index 50f6500..9e36a92 100644
--- a/src/ports/postgres/modules/kmeans/kmeans.sql_in
+++ b/src/ports/postgres/modules/kmeans/kmeans.sql_in
@@ -2092,7 +2092,7 @@
     rel_source VARCHAR,
     expr_point VARCHAR,
     centroids DOUBLE PRECISION[][],
-    fn_dist VARCHAR /*+ DEFAULT 'dist_norm2' */
+    fn_dist VARCHAR /*+ DEFAULT 'squared_dist_norm2' */
 ) RETURNS DOUBLE PRECISION
 LANGUAGE plpgsql VOLATILE
 AS $$
@@ -2164,7 +2164,7 @@
 LANGUAGE sql
 AS $$
     SELECT MADLIB_SCHEMA.simple_silhouette($1, $2, $3,
-        'MADLIB_SCHEMA.dist_norm2')
+        'MADLIB_SCHEMA.squared_dist_norm2')
 $$
 m4_ifdef(`__HAS_FUNCTION_PROPERTIES__', `READS SQL DATA', `');
 
@@ -2347,7 +2347,7 @@
     expr_point VARCHAR,
     centroids_table VARCHAR,
     centroids_col VARCHAR,
-    fn_dist VARCHAR /*+ DEFAULT 'dist_norm2' */
+    fn_dist VARCHAR /*+ DEFAULT 'squared_dist_norm2' */
 ) RETURNS VOID AS $$
     PythonFunction(kmeans, kmeans, simple_silhouette_points_str_wrapper)
 $$ LANGUAGE plpythonu VOLATILE
@@ -2363,7 +2363,7 @@
 ) RETURNS VOID
 AS $$
     SELECT MADLIB_SCHEMA.simple_silhouette_points($1, $2, $3, $4, $5, $6,
-        'MADLIB_SCHEMA.dist_norm2')
+        'MADLIB_SCHEMA.squared_dist_norm2')
 $$ LANGUAGE sql VOLATILE
 m4_ifdef(`\_\_HAS_FUNCTION_PROPERTIES\_\_', `MODIFIES SQL DATA', `');
 
@@ -2374,7 +2374,7 @@
     pid VARCHAR,
     expr_point VARCHAR,
     centroids DOUBLE PRECISION[],
-    fn_dist VARCHAR /*+ DEFAULT 'dist_norm2' */
+    fn_dist VARCHAR /*+ DEFAULT 'squared_dist_norm2' */
 ) RETURNS VOID AS $$
     PythonFunction(kmeans, kmeans, simple_silhouette_points_dbl_wrapper)
 $$ LANGUAGE plpythonu VOLATILE
@@ -2389,6 +2389,6 @@
 ) RETURNS VOID
 AS $$
     SELECT MADLIB_SCHEMA.simple_silhouette_points($1, $2, $3, $4, $5,
-        'MADLIB_SCHEMA.dist_norm2')
+        'MADLIB_SCHEMA.squared_dist_norm2')
 $$ LANGUAGE sql VOLATILE
 m4_ifdef(`\_\_HAS_FUNCTION_PROPERTIES\_\_', `MODIFIES SQL DATA', `');

src/ports/postgres/modules/kmeans/test/kmeans.sql_in

diff --git a/src/ports/postgres/modules/kmeans/test/kmeans.sql_in b/src/ports/postgres/modules/kmeans/test/kmeans.sql_in
index b0e5024..6e86c82 100644
--- a/src/ports/postgres/modules/kmeans/test/kmeans.sql_in
+++ b/src/ports/postgres/modules/kmeans/test/kmeans.sql_in
@@ -242,7 +242,7 @@
 SELECT * FROM simple_silhouette_points(
     'km_sample', 'silh_out', 'pid', 'points',
     ARRAY[[1,1,1,1,1,1,1,1,1,1,1,1,1],
-    [14.23, 1.71, 2.43, 15.6, 127, 2.8, 3.0600, 0.2800, 2.29, 5.64, 1.04, 3.92, 1065]]::DOUBLE PRECISION[][]);
+    [14.23, 1.71, 2.43, 15.6, 127, 2.8, 3.0600, 0.2800, 2.29, 5.64, 1.04, 3.92, 1065]]::DOUBLE PRECISION[][],'MADLIB_SCHEMA.dist_norm2');
 
 SELECT assert(relative_error(1, silh) < 1e-3,
     'Incorrect silhouette value')