[MRG] Remove random_seed in fit and use the one in init (#224)

* Remove random_seed in fit and use the one in init

* update tests with the new API

* Update test_RCA in sklearn_compat

* Update test_SDML in sklearn_compat

* Remove testing of pca_comps since it's deprecated

* Fix sklearn test

* Fix random_seed for test_iris in TestRCA

* Relaunch CI

* Augment tolerance rather than fix random_seed

* Add ChangedBehaviorWarning if the random_state is left default

* Update the merge

* Address #224 (review)

Author: wdevazelhes

metric_learn/constraints.py

@@ -6,6 +6,7 @@
 import warnings
 from six.moves import xrange
 from scipy.sparse import coo_matrix
+from sklearn.utils import check_random_state
 
 __all__ = ['Constraints']
 
@@ -23,7 +24,8 @@ def __init__(self, partial_labels):
     self.known_label_idx, = np.where(partial_labels >= 0)
     self.known_labels = partial_labels[self.known_label_idx]
 
-  def adjacency_matrix(self, num_constraints, random_state=np.random):
+  def adjacency_matrix(self, num_constraints, random_state=None):
+    random_state = check_random_state(random_state)
     a, b, c, d = self.positive_negative_pairs(num_constraints,
                                               random_state=random_state)
     row = np.concatenate((a, c))
@@ -35,7 +37,8 @@ def adjacency_matrix(self, num_constraints, random_state=np.random):
     return adj + adj.T
 
   def positive_negative_pairs(self, num_constraints, same_length=False,
-                              random_state=np.random):
+                              random_state=None):
+    random_state = check_random_state(random_state)
     a, b = self._pairs(num_constraints, same_label=True,
                        random_state=random_state)
     c, d = self._pairs(num_constraints, same_label=False,
@@ -68,13 +71,14 @@ def _pairs(self, num_constraints, same_label=True, max_iter=10,
     ab = np.array(list(ab)[:num_constraints], dtype=int)
     return self.known_label_idx[ab.T]
 
-  def chunks(self, num_chunks=100, chunk_size=2, random_state=np.random):
+  def chunks(self, num_chunks=100, chunk_size=2, random_state=None):
     """
     the random state object to be passed must be a numpy random seed
     """
+    random_state = check_random_state(random_state)
     chunks = -np.ones_like(self.known_label_idx, dtype=int)
     uniq, lookup = np.unique(self.known_labels, return_inverse=True)
-    all_inds = [set(np.where(lookup==c)[0]) for c in xrange(len(uniq))]
+    all_inds = [set(np.where(lookup == c)[0]) for c in xrange(len(uniq))]
     idx = 0
     while idx < num_chunks and all_inds:
       if len(all_inds) == 1:

metric_learn/itml.py

@@ -6,6 +6,7 @@
 import warnings
 import numpy as np
 from six.moves import xrange
+from sklearn.exceptions import ChangedBehaviorWarning
 from sklearn.metrics import pairwise_distances
 from sklearn.utils.validation import check_array
 from sklearn.base import TransformerMixin
@@ -298,7 +299,6 @@ class ITML_Supervised(_BaseITML, TransformerMixin):
            A positive definite (PD) matrix of shape
            (n_features, n_features), that will be used as such to set the
            prior.
-
   A0 : Not used
     .. deprecated:: 0.5.0
        `A0` was deprecated in version 0.5.0 and will
@@ -310,7 +310,9 @@ class ITML_Supervised(_BaseITML, TransformerMixin):
       tuples will be formed like this: X[indices].
   random_state : int or numpy.RandomState or None, optional (default=None)
       A pseudo random number generator object or a seed for it if int. If
-      ``prior='random'``, ``random_state`` is used to set the prior.
+      ``prior='random'``, ``random_state`` is used to set the prior. In any
+        case, `random_state` is also used to randomly sample constraints from
+        labels.
 
 
   Attributes
@@ -350,7 +352,7 @@ def __init__(self, gamma=1., max_iter=1000, convergence_threshold=1e-3,
     self.num_constraints = num_constraints
     self.bounds = bounds
 
-  def fit(self, X, y, random_state=np.random, bounds=None):
+  def fit(self, X, y, random_state='deprecated', bounds=None):
     """Create constraints from labels and learn the ITML model.
 
 
@@ -362,8 +364,11 @@ def fit(self, X, y, random_state=np.random, bounds=None):
     y : (n) array-like
         Data labels.
 
-    random_state : numpy.random.RandomState, optional
-        If provided, controls random number generation.
+    random_state : Not used
+      .. deprecated:: 0.5.0
+        `random_state` in the `fit` function was deprecated in version 0.5.0
+        and will be removed in 0.6.0. Set `random_state` at initialization
+        instead (when instantiating a new `ITML_Supervised` object).
 
     bounds : array-like of two numbers
         Bounds on similarity, aside slack variables, s.t.
@@ -384,6 +389,18 @@ def fit(self, X, y, random_state=np.random, bounds=None):
                     ' It has been deprecated in version 0.5.0 and will be'
                     ' removed in 0.6.0. Use the "bounds" parameter of this '
                     'fit method instead.', DeprecationWarning)
+    if random_state != 'deprecated':
+      warnings.warn('"random_state" parameter in the `fit` function is '
+                    'deprecated. Set `random_state` at initialization '
+                    'instead (when instantiating a new `ITML_Supervised` '
+                    'object).', DeprecationWarning)
+    else:
+      warnings.warn('As of v0.5.0, `ITML_Supervised` now uses the '
+                    '`random_state` given at initialization to sample '
+                    'constraints, not the default `np.random` from the `fit` '
+                    'method, since this argument is now deprecated. '
+                    'This warning will disappear in v0.6.0.',
+                    ChangedBehaviorWarning)
     X, y = self._prepare_inputs(X, y, ensure_min_samples=2)
     num_constraints = self.num_constraints
     if num_constraints is None:
@@ -392,6 +409,6 @@ def fit(self, X, y, random_state=np.random, bounds=None):
 
     c = Constraints(y)
     pos_neg = c.positive_negative_pairs(num_constraints,
-                                        random_state=random_state)
+                                        random_state=self.random_state)
     pairs, y = wrap_pairs(X, pos_neg)
     return _BaseITML._fit(self, pairs, y, bounds=bounds)

metric_learn/lsml.py

@@ -286,7 +286,8 @@ class LSML_Supervised(_BaseLSML, TransformerMixin):
   random_state : int or numpy.RandomState or None, optional (default=None)
       A pseudo random number generator object or a seed for it if int. If
       ``init='random'``, ``random_state`` is used to set the random
-      prior.
+      prior. In any case, `random_state` is also used to randomly sample
+      constraints from labels.
 
   Attributes
   ----------
@@ -308,7 +309,7 @@ def __init__(self, tol=1e-3, max_iter=1000, prior=None,
     self.num_constraints = num_constraints
     self.weights = weights
 
-  def fit(self, X, y, random_state=np.random):
+  def fit(self, X, y, random_state='deprecated'):
     """Create constraints from labels and learn the LSML model.
 
     Parameters
@@ -319,13 +320,28 @@ def fit(self, X, y, random_state=np.random):
     y : (n) array-like
         Data labels.
 
-    random_state : numpy.random.RandomState, optional
-        If provided, controls random number generation.
+    random_state : Not used
+      .. deprecated:: 0.5.0
+        `random_state` in the `fit` function was deprecated in version 0.5.0
+        and will be removed in 0.6.0. Set `random_state` at initialization
+        instead (when instantiating a new `LSML_Supervised` object).
     """
     if self.num_labeled != 'deprecated':
       warnings.warn('"num_labeled" parameter is not used.'
                     ' It has been deprecated in version 0.5.0 and will be'
                     ' removed in 0.6.0', DeprecationWarning)
+    if random_state != 'deprecated':
+      warnings.warn('"random_state" parameter in the `fit` function is '
+                    'deprecated. Set `random_state` at initialization '
+                    'instead (when instantiating a new `LSML_Supervised` '
+                    'object).', DeprecationWarning)
+    else:
+      warnings.warn('As of v0.5.0, `LSML_Supervised` now uses the '
+                    '`random_state` given at initialization to sample '
+                    'constraints, not the default `np.random` from the `fit` '
+                    'method, since this argument is now deprecated. '
+                    'This warning will disappear in v0.6.0.',
+                    ChangedBehaviorWarning)
     X, y = self._prepare_inputs(X, y, ensure_min_samples=2)
     num_constraints = self.num_constraints
     if num_constraints is None:
@@ -334,6 +350,6 @@ def fit(self, X, y, random_state=np.random):
 
     c = Constraints(y)
     pos_neg = c.positive_negative_pairs(num_constraints, same_length=True,
-                                        random_state=random_state)
+                                        random_state=self.random_state)
     return _BaseLSML._fit(self, X[np.column_stack(pos_neg)],
                           weights=self.weights)

metric_learn/mmc.py

@@ -538,7 +538,8 @@ class MMC_Supervised(_BaseMMC, TransformerMixin):
   random_state : int or numpy.RandomState or None, optional (default=None)
       A pseudo random number generator object or a seed for it if int. If
       ``init='random'``, ``random_state`` is used to initialize the random
-      Mahalanobis matrix.
+      Mahalanobis matrix.  In any case, `random_state` is also used to
+      randomly sample constraints from labels.
 
   `MMC_Supervised` creates pairs of similar sample by taking same class
   samples, and pairs of dissimilar samples by taking different class
@@ -566,7 +567,7 @@ def __init__(self, max_iter=100, max_proj=10000, convergence_threshold=1e-6,
     self.num_labeled = num_labeled
     self.num_constraints = num_constraints
 
-  def fit(self, X, y, random_state=np.random):
+  def fit(self, X, y, random_state='deprecated'):
     """Create constraints from labels and learn the MMC model.
 
     Parameters
@@ -575,13 +576,28 @@ def fit(self, X, y, random_state=np.random):
         Input data, where each row corresponds to a single instance.
     y : (n) array-like
         Data labels.
-    random_state : numpy.random.RandomState, optional
-        If provided, controls random number generation.
+    random_state : Not used
+      .. deprecated:: 0.5.0
+        `random_state` in the `fit` function was deprecated in version 0.5.0
+        and will be removed in 0.6.0. Set `random_state` at initialization
+        instead (when instantiating a new `MMC_Supervised` object).
     """
     if self.num_labeled != 'deprecated':
       warnings.warn('"num_labeled" parameter is not used.'
                     ' It has been deprecated in version 0.5.0 and will be'
                     ' removed in 0.6.0', DeprecationWarning)
+    if random_state != 'deprecated':
+      warnings.warn('"random_state" parameter in the `fit` function is '
+                    'deprecated. Set `random_state` at initialization '
+                    'instead (when instantiating a new `MMC_Supervised` '
+                    'object).', DeprecationWarning)
+    else:
+      warnings.warn('As of v0.5.0, `MMC_Supervised` now uses the '
+                    '`random_state` given at initialization to sample '
+                    'constraints, not the default `np.random` from the `fit` '
+                    'method, since this argument is now deprecated. '
+                    'This warning will disappear in v0.6.0.',
+                    ChangedBehaviorWarning)
     X, y = self._prepare_inputs(X, y, ensure_min_samples=2)
     num_constraints = self.num_constraints
     if num_constraints is None:
@@ -590,6 +606,6 @@ def fit(self, X, y, random_state=np.random):
 
     c = Constraints(y)
     pos_neg = c.positive_negative_pairs(num_constraints,
-                                        random_state=random_state)
+                                        random_state=self.random_state)
     pairs, y = wrap_pairs(X, pos_neg)
     return _BaseMMC._fit(self, pairs, y)

metric_learn/rca.py

@@ -184,11 +184,17 @@ class RCA_Supervised(RCA):
         be removed in 0.6.0. Use `n_components` instead.
 
   num_chunks: int, optional
+
   chunk_size: int, optional
+
   preprocessor : array-like, shape=(n_samples, n_features) or callable
       The preprocessor to call to get tuples from indices. If array-like,
       tuples will be formed like this: X[indices].
 
+  random_state : int or numpy.RandomState or None, optional (default=None)
+      A pseudo random number generator object or a seed for it if int.
+      It is used to randomly sample constraints from labels.
+
   Attributes
   ----------
   transformer_ : `numpy.ndarray`, shape=(n_components, n_features)
@@ -197,13 +203,15 @@ class RCA_Supervised(RCA):
 
   def __init__(self, num_dims='deprecated', n_components=None,
                pca_comps='deprecated', num_chunks=100, chunk_size=2,
-               preprocessor=None):
+               preprocessor=None, random_state=None):
+    """Initialize the supervised version of `RCA`."""
     RCA.__init__(self, num_dims=num_dims, n_components=n_components,
                  pca_comps=pca_comps, preprocessor=preprocessor)
     self.num_chunks = num_chunks
     self.chunk_size = chunk_size
+    self.random_state = random_state
 
-  def fit(self, X, y, random_state=np.random):
+  def fit(self, X, y, random_state='deprecated'):
     """Create constraints from labels and learn the RCA model.
     Needs num_constraints specified in constructor.
 
@@ -212,10 +220,26 @@ def fit(self, X, y, random_state=np.random):
     X : (n x d) data matrix
         each row corresponds to a single instance
     y : (n) data labels
-    random_state : a random.seed object to fix the random_state if needed.
+    random_state : Not used
+      .. deprecated:: 0.5.0
+        `random_state` in the `fit` function was deprecated in version 0.5.0
+        and will be removed in 0.6.0. Set `random_state` at initialization
+        instead (when instantiating a new `RCA_Supervised` object).
     """
+    if random_state != 'deprecated':
+      warnings.warn('"random_state" parameter in the `fit` function is '
+                    'deprecated. Set `random_state` at initialization '
+                    'instead (when instantiating a new `RCA_Supervised` '
+                    'object).', DeprecationWarning)
+    else:
+      warnings.warn('As of v0.5.0, `RCA_Supervised` now uses the '
+                    '`random_state` given at initialization to sample '
+                    'constraints, not the default `np.random` from the `fit` '
+                    'method, since this argument is now deprecated. '
+                    'This warning will disappear in v0.6.0.',
+                    ChangedBehaviorWarning)
     X, y = self._prepare_inputs(X, y, ensure_min_samples=2)
     chunks = Constraints(y).chunks(num_chunks=self.num_chunks,
                                    chunk_size=self.chunk_size,
-                                   random_state=random_state)
+                                   random_state=self.random_state)
     return RCA.fit(self, X, chunks)

metric_learn/sdml.py

@@ -310,7 +310,8 @@ class SDML_Supervised(_BaseSDML, TransformerMixin):
   random_state : int or numpy.RandomState or None, optional (default=None)
       A pseudo random number generator object or a seed for it if int. If
       ``init='random'``, ``random_state`` is used to set the random
-      prior.
+      prior. In any case, `random_state` is also used to randomly sample
+      constraints from labels.
 
   Attributes
   ----------
@@ -336,7 +337,7 @@ def __init__(self, balance_param=0.5, sparsity_param=0.01, prior=None,
     self.num_labeled = num_labeled
     self.num_constraints = num_constraints
 
-  def fit(self, X, y, random_state=np.random):
+  def fit(self, X, y, random_state='deprecated'):
     """Create constraints from labels and learn the SDML model.
 
     Parameters
@@ -345,9 +346,11 @@ def fit(self, X, y, random_state=np.random):
         data matrix, where each row corresponds to a single instance
     y : array-like, shape (n,)
         data labels, one for each instance
-    random_state : {numpy.random.RandomState, int}, optional
-        Random number generator or random seed. If not given, the singleton
-        numpy.random will be used.
+    random_state : Not used
+      .. deprecated:: 0.5.0
+        `random_state` in the `fit` function was deprecated in version 0.5.0
+        and will be removed in 0.6.0. Set `random_state` at initialization
+        instead (when instantiating a new `SDML_Supervised` object).
 
     Returns
     -------
@@ -358,6 +361,18 @@ def fit(self, X, y, random_state=np.random):
       warnings.warn('"num_labeled" parameter is not used.'
                     ' It has been deprecated in version 0.5.0 and will be'
                     ' removed in 0.6.0', DeprecationWarning)
+    if random_state != 'deprecated':
+      warnings.warn('"random_state" parameter in the `fit` function is '
+                    'deprecated. Set `random_state` at initialization '
+                    'instead (when instantiating a new `SDML_Supervised` '
+                    'object).', DeprecationWarning)
+    else:
+      warnings.warn('As of v0.5.0, `SDML_Supervised` now uses the '
+                    '`random_state` given at initialization to sample '
+                    'constraints, not the default `np.random` from the `fit` '
+                    'method, since this argument is now deprecated. '
+                    'This warning will disappear in v0.6.0.',
+                    ChangedBehaviorWarning)
     X, y = self._prepare_inputs(X, y, ensure_min_samples=2)
     num_constraints = self.num_constraints
     if num_constraints is None:
@@ -366,6 +381,6 @@ def fit(self, X, y, random_state=np.random):
 
     c = Constraints(y)
     pos_neg = c.positive_negative_pairs(num_constraints,
-                                        random_state=random_state)
+                                        random_state=self.random_state)
     pairs, y = wrap_pairs(X, pos_neg)
     return _BaseSDML._fit(self, pairs, y)