Change json for pickle in binarization

skbold/core/mvp_between.py

@@ -2,7 +2,7 @@
 
 import os
 import re
-import json
+import pickle
 import warnings
 import os.path as op
 import pandas as pd
@@ -418,9 +418,9 @@ def add_y(self, file_path, col_name, sep='\t', index_col=0,
             self._undersample_majority()
 
     def apply_binarization_params(self, param_file, ensure_balanced=False):
-
-        with open(param_file, 'r') as fin:
-            params = json.load(fin, encoding='utf-8')
+        """ Applies binarization-parameters to y. """
+        with open(param_file, 'rb') as fin:
+            params = pickle.load(fin)
 
         if params['type'] == 'zscore':
             y_norm = (self.y - params['mean']) / params['std']
@@ -471,9 +471,8 @@ def binarize_y(self, params, save_path=None, ensure_balanced=False):
             self._undersample_majority()
 
         if save_path is not None:
-            with open(op.join(save_path, 'binarize_params.json'), 'w',
-                      encoding="utf-8") as w:
-                json.dump(labb.binarize_params, w, indent=4)
+            with open(op.join(save_path, 'binarize_params.pkl'), 'wb') as w:
+                pickle.dump(labb.binarize_params, w)
 
     def split(self, file_path, col_name, target, sep='\t', index_col=0):
         """ Splits an MvpBetween object based on some external index.

skbold/core/tests/test_mvp_between.py

@@ -109,15 +109,15 @@ def test_mvp_between_binarize_y(mvp1c, params):
     mvp1c.add_y(fpath, col_name='var_continuous', index_col=0)
     mvp1c.binarize_y(params, ensure_balanced=True, save_path=testdata_path)
     assert((np.unique(mvp1c.y) == [0, 1]).all())
-    assert(op.isfile(op.join(testdata_path, 'binarize_params.json')))
+    assert(op.isfile(op.join(testdata_path, 'binarize_params.pkl')))
 
 
 def test_mvp_between_apply_binarization_params(mvp1c):
     fpath = op.join(testdata_path, 'sample_behav.tsv')
     mvp1c.add_y(fpath, col_name='var_continuous', index_col=0)
     mvp1c.apply_binarization_params(op.join(testdata_path,
-                                            'binarize_params.json'))
-    os.remove(op.join(testdata_path, 'binarize_params.json'))
+                                            'binarize_params.pkl'))
+    os.remove(op.join(testdata_path, 'binarize_params.pkl'))
 
 
 @pytest.mark.parametrize("cols", ['confound_categorical',

skbold/data/test_data/sample_behav.tsv

@@ -1,14 +1,14 @@
 	var_categorical	var_continuous	var_multinomial	confound_categorical	confound_continuous	group
 sub001	1	23.123	2	0	384.123	train
 sub002	1	24.129	2	0	819.32	train
-sub003	1	20.001	2	1	381.33	
+sub003	1	20.001	2	1	381.33	train
 sub004	0	19.583	2	1	921.939	train
-sub005		9.209	2	1	201.193	train
+sub005	0	9.209	2	1	201.193	train
 sub006	0	58.284	1	0	403.201	train
 sub007	0	132.123	1	1	683.863	test
 sub008	999	492.593	1	1	290.193	test
 sub009	1	394.288	1	0	993.201	test
 sub010	1	918.13	3	0	381.382	test
-sub011	1	382.111		1	998.281	test
+sub011	1	382.111	3	1	998.281	test
 sub012	0	399.11	3	0	888.133	test
 sub013	1	393.99	3	0	133.281	test