adding check_array everywhere

fancyimpute/bayesian_ridge_regression.py

@@ -17,6 +17,8 @@
 from numpy.linalg import norm, inv, multi_dot
 from numpy.random import multivariate_normal
 
+from sklearn.utils import check_array
+
 
 class BayesianRidgeRegression(object):
     """
@@ -44,6 +46,8 @@ def __init__(self, lambda_reg=0.001, add_ones=False, normalize_lambda=True):
         self.normalize_lambda = normalize_lambda
 
     def fit(self, X, y, inverse_covariance=None):
+        X = check_array(X)
+
         if self.add_ones:
             X_ones = self.add_column_of_ones(X)
         else:

fancyimpute/iterative_svd.py

@@ -13,6 +13,7 @@
 from __future__ import absolute_import, print_function, division
 
 from sklearn.decomposition import TruncatedSVD
+from sklearn.utils import check_array
 import numpy as np
 
 from .solver import Solver
@@ -53,6 +54,8 @@ def _converged(self, X_old, X_new, missing_mask):
         return (ssd / old_norm_squared) < self.convergence_threshold
 
     def solve(self, X, missing_mask):
+        X = check_array(X, force_all_finite=False)
+
         observed_mask = ~missing_mask
         X_filled = X
         for i in range(self.max_iters):

fancyimpute/knn.py

@@ -14,6 +14,7 @@
 import numpy as np
 
 from knnimpute import knn_impute_few_observed, knn_impute_with_argpartition
+from sklearn.utils import check_array
 
 from .solver import Solver
 
@@ -83,6 +84,8 @@ def __init__(
             self._impute_fn = knn_impute_few_observed
 
     def solve(self, X, missing_mask):
+        X = check_array(X, force_all_finite=False)
+
         if self.orientation == "columns":
             X = X.T
             missing_mask = missing_mask.T

fancyimpute/matrix_factorization.py

@@ -17,7 +17,7 @@
 from keras.callbacks import EarlyStopping
 from keras.layers import Input
 from keras.models import Model
-from sklearn.utils import shuffle
+from sklearn.utils import shuffle, check_array
 
 from .common import import_from
 from .scaler import Scaler
@@ -68,6 +68,8 @@ def __init__(
         self.verbose = verbose
 
     def solve(self, X, missing_mask):
+        X = check_array(X, force_all_finite=False)
+
         # shape data to fit into keras model
         (n_samples, n_features) = X.shape
         observed_mask = ~missing_mask

fancyimpute/mice.py

@@ -17,6 +17,8 @@
 from six.moves import range
 import numpy as np
 
+from sklearn.utils import check_array
+
 from .bayesian_ridge_regression import BayesianRidgeRegression
 from .solver import Solver
 
@@ -290,7 +292,7 @@ def multiple_imputations(self, X):
         belong in X.
         """
         start_t = time()
-        X = np.asarray(X)
+        X = check_array(X, force_all_finite=False)
         self._check_input(X)
         missing_mask = np.isnan(X)
         self._check_missing_value_mask(missing_mask)
@@ -330,7 +332,7 @@ def multiple_imputations(self, X):
     def complete(self, X):
         if self.verbose:
             print("[MICE] Completing matrix with shape %s" % (X.shape,))
-        X_completed = np.array(X.copy())
+        X_completed = check_array(X, force_all_finite=False, copy=True)
         imputed_arrays, missing_mask = self.multiple_imputations(X)
         # average the imputed values for each feature
         average_imputated_values = imputed_arrays.mean(axis=0)

fancyimpute/nuclear_norm_minimization.py

@@ -16,6 +16,7 @@
 
 from .solver import Solver
 
+from sklearn.utils import check_array
 
 class NuclearNormMinimization(Solver):
     """
@@ -109,6 +110,8 @@ def _create_objective(self, m, n):
         return S, objective
 
     def solve(self, X, missing_mask):
+        X = check_array(X, force_all_finite=False)
+
         m, n = X.shape
         S, objective = self._create_objective(m, n)
         constraints = self._constraints(

fancyimpute/similarity_weighted_averaging.py

@@ -13,6 +13,8 @@
 from __future__ import absolute_import, print_function, division
 from collections import defaultdict
 
+from sklearn.utils import check_array
+
 import numpy as np
 
 from .dictionary_helpers import (
@@ -151,6 +153,8 @@ def complete_dict(
         return result
 
     def complete(self, X):
+        X = check_array(X, force_all_finite=False)
+
         if self.verbose:
             print(
                 ("[SimilarityWeightedAveraging] Creating dictionary from matrix "

fancyimpute/soft_impute.py

@@ -15,6 +15,7 @@
 from six.moves import range
 import numpy as np
 from sklearn.utils.extmath import randomized_svd
+from sklearn.utils import check_array
 
 from .common import masked_mae
 from .solver import Solver
@@ -134,6 +135,8 @@ def _max_singular_value(self, X_filled):
         return s[0]
 
     def solve(self, X, missing_mask):
+        X = check_array(X, force_all_finite=False)
+
         X_init = X.copy()
 
         X_filled = X

fancyimpute/solver.py

@@ -15,6 +15,8 @@
 import numpy as np
 from six.moves import range
 
+from sklearn.utils import check_array
+
 from .common import generate_random_column_samples
 
 
@@ -91,6 +93,8 @@ def fill(
         inplace : bool
             Modify matrix or fill a copy
         """
+        X = check_array(X, force_all_finite=False)
+
         if not inplace:
             X = X.copy()
 
@@ -120,7 +124,7 @@ def prepare_input_data(self, X):
         Check to make sure that the input matrix and its mask of missing
         values are valid. Returns X and missing mask.
         """
-        X = np.asarray(X)
+        X = check_array(X, force_all_finite=False)
         if X.dtype != "f" and X.dtype != "d":
             X = X.astype(float)
 
@@ -142,7 +146,7 @@ def clip(self, X):
 
     def project_result(self, X):
         """
-        First undo normaliztion and then clip to the user-specified min/max
+        First undo normalization and then clip to the user-specified min/max
         range.
         """
         X = np.asarray(X)