Merge pull request #235 from delallea/minor

theano/compile/function_module.py

@@ -991,11 +991,11 @@ class FunctionMaker(object):
 
         # optimize the env
         compute_test_value_orig = theano.config.compute_test_value
+        add_stack_trace_on_call = gof.Op.add_stack_trace_on_call
         try:
             theano.config.compute_test_value = "off"
-            start_optimizer = time.time()
-            add_stack_trace_on_call = gof.Op.add_stack_trace_on_call
             gof.Op.add_stack_trace_on_call = False
+            start_optimizer = time.time()
             optimizer(env)
             end_optimizer = time.time()
 

theano/sandbox/cuda/elemwise.py

@@ -5,12 +5,14 @@ The elemwise fct are also used with scalar operation! So it can happen that ndim
 """
 
 
-import StringIO, sys
+import copy, logging, StringIO, sys
+
 import numpy
-from theano import Op, Type, Apply, Variable, Constant
-from theano import tensor, scalar, gof
 
-import logging, copy
+from theano import Apply, Constant, Op, Type, Variable
+from theano import gof, scalar, tensor
+
+
 _logger_name = 'theano.sandbox.cuda.elemwise'
 _logger = logging.getLogger(_logger_name)
 _logger.setLevel(logging.INFO)
@@ -47,7 +49,7 @@ class NaiveAlgo(object):
             if code:
                 raise SupportCodeError(scalar_op)
         except gof.utils.MethodNotDefined:
-                pass
+            pass
         self.scalar_op = scalar_op
         self.sync = sync
         self.inplace_pattern = inplace_pattern

theano/sandbox/linalg/ops.py

@@ -696,9 +696,11 @@ def spectral_radius_bound(X, log2_exponent):
     if X.type.ndim != 2:
         raise TypeError('spectral_radius_bound requires a matrix argument', X)
     if not isinstance(log2_exponent, int):
-        raise TypeError('spectral_radius_bound requires a integer exponent', log2_exponent)
+        raise TypeError('spectral_radius_bound requires an integer exponent',
+                        log2_exponent)
     if log2_exponent <= 0:
-        raise ValueError('spectral_radius_bound requires a strictly positive exponent', log2_exponent)
+        raise ValueError('spectral_radius_bound requires a strictly positive '
+                         'exponent', log2_exponent)
     XX = X
     for i in xrange(log2_exponent):
         XX = tensor.dot(XX, XX)

theano/sandbox/neighbours.py

@@ -11,8 +11,10 @@ if cuda_available:
 
 
 class BadOldCode(Exception):
-    """ We create a specific Exception to be sure it don't get caught
-    by mistake"""
+    """
+    We create a specific Exception to be sure it does not get caught by
+    mistake.
+    """
     pass
 
 

theano/scan_module/tests/test_scan.py

@@ -2600,10 +2600,11 @@ class T_Scan(unittest.TestCase):
         assert numpy.allclose(theano_y , v_y[-4:])
 
     def test_opt_order(self):
-        """  Verify that scan optimizations are applied before blas
-           optimizations.
-           This is needed as otherwise, the dot won't become a dot22
-           so it will be slower and won't get transferred to the gpu.
+        """
+        Verify that scan optimizations are applied before blas
+        optimizations.
+        This is needed as otherwise, the dot won't become a dot22
+        so it will be slower and won't get transferred to the gpu.
         """
         x = theano.tensor.matrix('x')
         A = theano.tensor.matrix('A')

theano/tensor/tensor_grad.py

@@ -182,7 +182,7 @@ def Lop(f, wrt, eval_points, consider_constant=None, warn_type=False,
 
     inputs = gof.graph.inputs(f)
     gmap = gradient.grad_sources_inputs(
-            zip(f,eval_points),
+            zip(f, eval_points),
             list(inputs) + list(consider_constant),
             warn_type=warn_type)
 
@@ -337,7 +337,7 @@ def grad(cost, wrt, g_cost=None, consider_constant=None, warn_type=False,
 
     if len(ret) == 1 and not (using_list or using_tuple):
         # `wrt` was a single Variable, so we return a single Variable too.
-            return ret[0]
+        return ret[0]
     else:
         # Ensure we preserve the original type of `wrt`.
         if using_tuple: