Change shuffle param behavior

Author: Alexsandruss

config_example.json

@@ -208,7 +208,6 @@
                     }
                 }
             ],
-            "shuffle": [""],
             "include-y": [""],
             "train-size": [0.75],
             "test-size": [0.25]

cuml/train_test_split.py

@@ -12,8 +12,8 @@
                     help='Size of training subset')
 parser.add_argument('--test-size', type=float, default=0.25,
                     help='Size of testing subset')
-parser.add_argument('--shuffle', default=False, action='store_true',
-                    help='Perform data shuffle before splitting')
+parser.add_argument('--do-not-shuffle', default=False, action='store_true',
+                    help='Do not perform data shuffle before splitting')
 params = parse_args(parser)
 
 # Load generated data
@@ -22,7 +22,7 @@
 tts_params = {
     'train_size': params.train_size,
     'test_size': params.test_size,
-    'shuffle': params.shuffle,
+    'shuffle': not params.do_not_shuffle,
     'random_state': params.seed
 }
 

sklearn/train_test_split.py

@@ -12,8 +12,8 @@
                     help='Size of training subset')
 parser.add_argument('--test-size', type=float, default=0.25,
                     help='Size of testing subset')
-parser.add_argument('--shuffle', default=False, action='store_true',
-                    help='Perform data shuffle before splitting')
+parser.add_argument('--do-not-shuffle', default=False, action='store_true',
+                    help='Do not perform data shuffle before splitting')
 parser.add_argument('--include-y', default=False, action='store_true',
                     help='Include label (Y) in splitting')
 parser.add_argument('--rng', default=None,
@@ -35,7 +35,7 @@
 tts_params = {
     'train_size': params.train_size,
     'test_size': params.test_size,
-    'shuffle': params.shuffle,
+    'shuffle': not params.do_not_shuffle,
     'random_state': params.seed
 }