Some cleanup

carl/distributions/base.py

@@ -11,10 +11,9 @@
 from scipy.optimize import minimize
 
 from sklearn.base import BaseEstimator
-from sklearn.utils import check_random_state as check_random_state_sklearn
+from sklearn.utils import check_random_state
 
 from theano.gof import graph
-from theano.tensor.shared_randomstreams import RandomStreams
 from theano.tensor.sharedvar import SharedVariable
 
 # ???: define the bounds of the parameters
@@ -47,15 +46,6 @@ def check_parameter(name, value):
     return value, parameters, constants, observeds
 
 
-def check_random_state_theano(random_state):
-    if isinstance(random_state, RandomStreams):
-        return random_state
-    elif isinstance(random_state, np.random.RandomState):
-        random_state = random_state.randint(np.iinfo(np.int32).max)
-
-    return RandomStreams(seed=random_state)
-
-
 def bound(expression, out, *predicates):
     guard = 1
     for p in predicates:
@@ -111,7 +101,7 @@ def make_(self, expression, name, args=None, kwargs=None):
         setattr(self, name, func)
 
     def rvs(self, n_samples, **kwargs):
-        rng = check_random_state_sklearn(self.random_state)
+        rng = check_random_state(self.random_state)
         p = rng.rand(n_samples, 1)
         return self.ppf(p, **kwargs)
 

carl/distributions/mixture.py

@@ -9,13 +9,22 @@
 import theano.tensor as T
 
 from theano.gof import graph
+from theano.tensor.shared_randomstreams import RandomStreams
 
 from . import DistributionMixin
-from .base import check_random_state_theano
 from .base import check_parameter
 from .base import bound
 
 
+def check_random_state(random_state):
+    if isinstance(random_state, RandomStreams):
+        return random_state
+    elif isinstance(random_state, np.random.RandomState):
+        random_state = random_state.randint(np.iinfo(np.int32).max)
+
+    return RandomStreams(seed=random_state)
+
+
 class Mixture(DistributionMixin):
     def __init__(self, components, weights=None,
                        random_state=None, optimizer=None):
@@ -86,7 +95,7 @@ def __init__(self, components, weights=None,
 
         # randc
         n_samples = T.iscalar()
-        rng = check_random_state_theano(self.random_state)
+        rng = check_random_state(self.random_state)
         self.randc_ = rng.multinomial(size=(n_samples,), pvals=self.weights)
         self.make_(self.randc_, "randc", args=[n_samples])