Apply pyupgrade to theano.tensor

theano/tensor/blas.py

@@ -477,7 +477,7 @@ def _ldflags(ldflags_str, libs, flags, libs_dir, include_dir):
             assert t0 == "-"
         except Exception:
             raise ValueError(
-                'invalid token "%s" in ldflags_str: "%s"' % (t, ldflags_str)
+                'invalid token "{}" in ldflags_str: "{}"'.format(t, ldflags_str)
             )
         if libs_dir and t1 == "L":
             rval.append(t[2:])
@@ -800,7 +800,7 @@ class GemmRelated(Op):
 
     def build_gemm_call(self):
         if hasattr(self, "inplace"):
-            setup_z_Nz_Sz = "if(%%(params)s->inplace){%s}else{%s}" % (
+            setup_z_Nz_Sz = "if(%(params)s->inplace){{{}}}else{{{}}}".format(
                 self.setup_z_Nz_Sz_inplace,
                 self.setup_z_Nz_Sz_outplace,
             )
@@ -880,7 +880,7 @@ class Gemm(GemmRelated):
             inplace_str = "inplace"
         else:
             inplace_str = "no_inplace"
-        return "%s{%s}" % (self.__class__.__name__, inplace_str)
+        return "{}{{{}}}".format(self.__class__.__name__, inplace_str)
 
     def __setstate__(self, dct):
         self.__dict__.update(dct)
@@ -1671,7 +1671,7 @@ class Dot22(GemmRelated):
         if node.inputs[0].type.dtype.startswith("complex"):
             raise MethodNotDefined("%s.c_code" % self.__class__.__name__)
         if len(self.c_libraries()) <= 0:
-            return super(Dot22, self).c_code(node, name, (_x, _y), (_zout,), sub)
+            return super().c_code(node, name, (_x, _y), (_zout,), sub)
         full_code = self.build_gemm_call() % dict(locals(), **sub)
         return full_code
 
@@ -1941,7 +1941,7 @@ class Dot22Scalar(GemmRelated):
         if node.inputs[0].type.dtype.startswith("complex"):
             raise MethodNotDefined("%s.c_code" % self.__class__.__name__)
         if len(self.c_libraries()) <= 0:
-            return super(Dot22Scalar, self).c_code(node, name, (_x, _y), (_zout,), sub)
+            return super().c_code(node, name, (_x, _y), (_zout,), sub)
         full_code = self.build_gemm_call() % dict(locals(), **sub)
         return full_code
 

theano/tensor/blas_c.py

@@ -23,7 +23,7 @@ from theano.tensor.blas import (
 from theano.tensor.opt import in2out
 
 
-class BaseBLAS(object):
+class BaseBLAS:
     def c_libraries(self):
         return ldflags()
 
@@ -617,7 +617,7 @@ class CGemv(BaseBLAS, Gemv):
     )
 
     def __init__(self, inplace):
-        super(CGemv, self).__init__(inplace)
+        super().__init__(inplace)
 
     def c_code(self, node, name, inp, out, sub):
         y, alpha, A, x, beta = inp

theano/tensor/blas_headers.py

@@ -770,7 +770,7 @@ def blas_header_text():
                 "precision": "d",
             }
         if not common_code or not template_code:
-            raise IOError(
+            raise OSError(
                 "Unable to load NumPy implementation of BLAS functions from C source files."
             )
         blas_code += common_code

theano/tensor/elemwise_cgen.py

@@ -77,7 +77,9 @@ def make_checks(loop_orders, dtypes, sub):
             if index != "x":
                 # Initialize the variables associated to the jth loop
                 # jump = stride - adjust
-                jump = "(%s) - (%s)" % ("%(var)s_stride%(index)s" % locals(), adjust)
+                jump = "({}) - ({})".format(
+                    "%(var)s_stride%(index)s" % locals(), adjust
+                )
                 init += (
                     """
                 %(var)s_n%(index)s = PyArray_DIMS(%(var)s)[%(index)s];

theano/tensor/extra_ops.py

@@ -274,7 +274,7 @@ class CumOp(Op):
 
     def __init__(self, axis=None, mode="add"):
         if mode not in ("add", "mul"):
-            raise ValueError('%s: Unknown mode "%s"' % (type(self).__name__, mode))
+            raise ValueError('{}: Unknown mode "{}"'.format(type(self).__name__, mode))
         self.axis = axis
         self.mode = mode
 
@@ -287,7 +287,7 @@ class CumOp(Op):
         if self.axis is None:
             out_type = theano.tensor.vector(dtype=x.dtype)  # Flatten
         elif self.axis >= x.ndim or self.axis < -x.ndim:
-            raise ValueError("axis(={0}) out of bounds".format(self.axis))
+            raise ValueError("axis(={}) out of bounds".format(self.axis))
 
         return theano.Apply(self, [x], [out_type])
 
@@ -327,7 +327,9 @@ class CumOp(Op):
             return [cumsum((fx * gi)[reverse_slicing], self.axis)[reverse_slicing] / x]
         else:
             raise NotImplementedError(
-                '%s: unknown gradient for mode "%s"' % (type(self).__name__, self.mode)
+                '{}: unknown gradient for mode "{}"'.format(
+                    type(self).__name__, self.mode
+                )
             )
 
     def infer_shape(self, node, shapes):
@@ -391,7 +393,7 @@ class CumOp(Op):
         return (8,)
 
     def __str__(self):
-        return "%s{%s, %s}" % (self.__class__.__name__, self.axis, self.mode)
+        return "{}{{{}, {}}}".format(self.__class__.__name__, self.axis, self.mode)
 
 
 def cumsum(x, axis=None):

theano/tensor/io.py

@@ -58,7 +58,7 @@ class LoadFromDisk(Op):
         out[0][0] = result
 
     def __str__(self):
-        return "Load{dtype: %s, broadcastable: %s, mmep: %s}" % (
+        return "Load{{dtype: {}, broadcastable: {}, mmep: {}}}".format(
             self.dtype,
             self.broadcastable,
             self.mmap_mode,

theano/tensor/nnet/bn.py

@@ -28,7 +28,7 @@ class BNComposite(Composite):
         o = add(mul(true_div(sub(x, mean), std), gamma), beta)
         inputs = [x, mean, std, gamma, beta]
         outputs = [o]
-        super(BNComposite, self).__init__(inputs, outputs)
+        super().__init__(inputs, outputs)
 
     def grad(self, inps, grads):
         x, mean, std, gamma, beta = inps

theano/tensor/nnet/conv.py

@@ -488,7 +488,7 @@ class ConvOp(OpenMPOp):
             )
 
         # Init the openmp attribute
-        super(ConvOp, self).__init__(openmp=openmp)
+        super().__init__(openmp=openmp)
         if not all_shape or self.openmp:
             # Only this version is parallelized
             unroll_patch = True
@@ -687,7 +687,7 @@ class ConvOp(OpenMPOp):
         return True
 
     def __setstate__(self, d):
-        super(ConvOp, self).__setstate__(d)
+        super().__setstate__(d)
         self.direction_hint = d.get("direction_hint", None)
         self._rehash()
 
@@ -1197,7 +1197,7 @@ using namespace std;
         if theano.gof.cmodule.gcc_version() in ["4.3.0"] and self.kshp == (1, 1):
             ret += ["-O2"]
         # Add the -fopenmp flags
-        ret += super(ConvOp, self).c_compile_args()
+        ret += super().c_compile_args()
 
         return ret
 

theano/tensor/nnet/corr.py

 import logging
 import os
 
-from six import integer_types
-
 import theano
 from theano import gof
 from theano.gof.graph import Apply
@@ -86,8 +84,8 @@ class BaseCorrMM(gof.OpenMPOp):
         unshared=False,
         openmp=None,
     ):
-        super(BaseCorrMM, self).__init__(openmp=openmp)
-        if isinstance(border_mode, integer_types):
+        super().__init__(openmp=openmp)
+        if isinstance(border_mode, int):
             if border_mode < 0:
                 raise ValueError(
                     "invalid border_mode {}, which must be a "
@@ -175,7 +173,7 @@ class BaseCorrMM(gof.OpenMPOp):
     padW_r = property(lambda self: self.pad[1][1])
 
     def __str__(self):
-        return "%s{%s, %s, %s, %s %s}" % (
+        return "{}{{{}, {}, {}, {} {}}}".format(
             self.__class__.__name__,
             self.border_mode,
             str(self.subsample),
@@ -210,7 +208,7 @@ class BaseCorrMM(gof.OpenMPOp):
 
     def c_compile_args(self):
         compile_args = ldflags(libs=False, flags=True)
-        compile_args += super(BaseCorrMM, self).c_compile_args()
+        compile_args += super().c_compile_args()
         return compile_args
 
     def c_lib_dirs(self):
@@ -221,7 +219,7 @@ class BaseCorrMM(gof.OpenMPOp):
 
     def c_headers(self):
         headers = ["<stdio.h>"]
-        headers += super(BaseCorrMM, self).c_headers()
+        headers += super().c_headers()
         return headers
 
     def c_code_cache_version(self):
@@ -710,7 +708,7 @@ class CorrMM(BaseCorrMM):
     def c_code(self, node, nodename, inp, out_, sub):
         bottom, weights = inp
         (top,) = out_
-        return super(CorrMM, self).c_code_helper(bottom, weights, top, sub)
+        return super().c_code_helper(bottom, weights, top, sub)
 
     def grad(self, inp, grads):
         bottom, weights = inp
@@ -835,9 +833,7 @@ class CorrMM_gradWeights(BaseCorrMM):
         bottom, top = inp[:2]
         height, width = inp[2:] or (None, None)
         (weights,) = out_
-        return super(CorrMM_gradWeights, self).c_code_helper(
-            bottom, weights, top, sub, height, width
-        )
+        return super().c_code_helper(bottom, weights, top, sub, height, width)
 
     def grad(self, inp, grads):
         bottom, top = inp[:2]
@@ -969,9 +965,7 @@ class CorrMM_gradInputs(BaseCorrMM):
         weights, top = inp[:2]
         height, width = inp[2:] or (None, None)
         (bottom,) = out_
-        return super(CorrMM_gradInputs, self).c_code_helper(
-            bottom, weights, top, sub, height, width
-        )
+        return super().c_code_helper(bottom, weights, top, sub, height, width)
 
     def grad(self, inp, grads):
         weights, top = inp[:2]

theano/tensor/nnet/corr3d.py

 import logging
 import os
 
-from six import integer_types
-
 import theano
 from theano import gof
 from theano.gof.graph import Apply
@@ -77,8 +75,8 @@ class BaseCorr3dMM(gof.OpenMPOp):
         openmp=None,
         num_groups=1,
     ):
-        super(BaseCorr3dMM, self).__init__(openmp=openmp)
-        if isinstance(border_mode, integer_types):
+        super().__init__(openmp=openmp)
+        if isinstance(border_mode, int):
             if border_mode < 0:
                 raise ValueError(
                     "invalid border_mode {}, which must be a "
@@ -159,7 +157,7 @@ class BaseCorr3dMM(gof.OpenMPOp):
     padD = property(lambda self: self.pad[2])
 
     def __str__(self):
-        return "%s{%s, %s, %s, %s}" % (
+        return "{}{{{}, {}, {}, {}}}".format(
             self.__class__.__name__,
             self.border_mode,
             str(self.subsample),
@@ -193,7 +191,7 @@ class BaseCorr3dMM(gof.OpenMPOp):
 
     def c_compile_args(self):
         compile_args = ldflags(libs=False, flags=True)
-        compile_args += super(BaseCorr3dMM, self).c_compile_args()
+        compile_args += super().c_compile_args()
         return compile_args
 
     def c_lib_dirs(self):
@@ -204,7 +202,7 @@ class BaseCorr3dMM(gof.OpenMPOp):
 
     def c_headers(self):
         headers = ["<stdio.h>"]
-        headers += super(BaseCorr3dMM, self).c_headers()
+        headers += super().c_headers()
         return headers
 
     def c_code_cache_version(self):
@@ -650,7 +648,7 @@ class Corr3dMM(BaseCorr3dMM):
     def c_code(self, node, nodename, inp, out_, sub):
         bottom, weights = inp
         (top,) = out_
-        return super(Corr3dMM, self).c_code_helper(bottom, weights, top, sub)
+        return super().c_code_helper(bottom, weights, top, sub)
 
     def grad(self, inp, grads):
         bottom, weights = inp
@@ -764,9 +762,7 @@ class Corr3dMMGradWeights(BaseCorr3dMM):
         bottom, top = inp[:2]
         height, width, depth = inp[2:] or (None, None, None)
         (weights,) = out_
-        return super(Corr3dMMGradWeights, self).c_code_helper(
-            bottom, weights, top, sub, height, width, depth
-        )
+        return super().c_code_helper(bottom, weights, top, sub, height, width, depth)
 
     def grad(self, inp, grads):
         bottom, top = inp[:2]
@@ -900,9 +896,7 @@ class Corr3dMMGradInputs(BaseCorr3dMM):
         weights, top = inp[:2]
         height, width, depth = inp[2:] or (None, None, None)
         (bottom,) = out_
-        return super(Corr3dMMGradInputs, self).c_code_helper(
-            bottom, weights, top, sub, height, width, depth
-        )
+        return super().c_code_helper(bottom, weights, top, sub, height, width, depth)
 
     def grad(self, inp, grads):
         weights, top = inp[:2]

theano/tensor/nnet/nnet.py

@@ -2206,7 +2206,7 @@ class Prepend_scalar_constant_to_each_row(Op):
         self.val = val
 
     def __str__(self):
-        return "%s{%s}" % (self.__class__.__name__, self.val)
+        return "{}{{{}}}".format(self.__class__.__name__, self.val)
 
     def make_node(self, mat):
         # check type of input

theano/tensor/nnet/sigm.py

@@ -91,7 +91,7 @@ class ScalarSigmoid(scalar.UnaryScalarOp):
             raise NotImplementedError("only floatingpoint is implemented")
 
     def c_code_cache_version(self):
-        v = super(ScalarSigmoid, self).c_code_cache_version()
+        v = super().c_code_cache_version()
         if v:
             return (2,) + v
         else:
@@ -404,7 +404,7 @@ class ScalarSoftplus(scalar.UnaryScalarOp):
             raise NotImplementedError("only floatingpoint is implemented")
 
     def c_code_cache_version(self):
-        v = super(ScalarSoftplus, self).c_code_cache_version()
+        v = super().c_code_cache_version()
         if v:
             return (2,) + v
         else:

theano/tensor/opt.py

@@ -13,7 +13,7 @@ from collections import defaultdict
 from functools import reduce
 
 import numpy as np
-from six import StringIO, integer_types
+from six import StringIO
 
 import theano
 import theano.scalar.basic as ts
@@ -515,7 +515,8 @@ class InplaceElemwiseOptimizer(Optimizer):
 
     def print_summary(self, stream=sys.stdout, level=0, depth=-1):
         print(
-            "%s%s (%s)" % ((" " * level), self.__class__.__name__, self.op), file=stream
+            "{}{} ({})".format((" " * level), self.__class__.__name__, self.op),
+            file=stream,
         )
         return inplace_elemwise_optimizer
 
@@ -996,7 +997,7 @@ class MakeVectorPrinter:
 tt.pprint.assign(MakeVector, MakeVectorPrinter())
 
 
-class ShapeFeature(object):
+class ShapeFeature:
     """Graph optimizer for removing all calls to shape().
 
     This optimizer replaces all Shapes and Subtensors of Shapes with
@@ -1222,12 +1223,10 @@ class ShapeFeature(object):
             # don't make the optimizer merge a zillion ones together
             # by always returning the same object to represent 1
             return self.lscalar_one
-        if type(s_i) is float and int(s_i) == s_i:
+        if isinstance(s_i, float) and int(s_i) == s_i:
             s_i = int(s_i)
-        if (
-            type(s_i) in integer_types
-            or isinstance(s_i, np.integer)
-            or (isinstance(s_i, np.ndarray) and s_i.ndim == 0)
+        if isinstance(s_i, (np.integer, int)) or (
+            isinstance(s_i, np.ndarray) and s_i.ndim == 0
         ):
             # this shape is a constant
             if s_i < 0:
@@ -1241,7 +1240,7 @@ class ShapeFeature(object):
                 # message.
                 raise AssertionError(msg)
             return tt.constant(s_i, dtype="int64")
-        if type(s_i) in (tuple, list):
+        if isinstance(s_i, (tuple, list)):
             # this dimension is the same as many of the inputs
             # which tells us that if one of the inputs is known,
             # the others all become known.
@@ -1384,11 +1383,11 @@ class ShapeFeature(object):
                 #  - Shape_i(i)(other_r);
                 #  - Shape_i(i)(r).
                 merged_shape.append(r_shape[i])
-            elif isinstance(r_shape[i], (Constant, integer_types)):
+            elif isinstance(r_shape[i], (Constant, int)):
                 # We do this to call less often ancestors and make
                 # sure we have the simplest shape possible.
                 merged_shape.append(r_shape[i])
-            elif isinstance(other_shape[i], (Constant, integer_types)):
+            elif isinstance(other_shape[i], (Constant, int)):
                 # We do this to call less often ancestors and make
                 # sure we have the simplest shape possible.
                 merged_shape.append(other_shape[i])
@@ -2225,7 +2224,7 @@ def local_subtensor_remove_broadcastable_index(node):
         elif isinstance(elem, slice):
             if elem != slice(None):
                 return
-        elif isinstance(elem, (integer_types, np.integer)):
+        elif isinstance(elem, (int, np.integer)):
             if elem in [0, -1] and node.inputs[0].broadcastable[dim]:
                 remove_dim.append(dim)
         else:
@@ -2277,7 +2276,7 @@ def local_subtensor_make_vector(node):
     else:
         return
 
-    if isinstance(idx, (integer_types, np.integer)):
+    if isinstance(idx, (int, np.integer)):
         # We don't need to copy over any stack traces here
         return [x.owner.inputs[idx]]
     elif isinstance(idx, Variable):
@@ -3014,7 +3013,7 @@ def local_useless_subtensor(node):
 
             length_pos = shape_of[node.inputs[0]][pos]
 
-            if isinstance(idx.stop, (integer_types, np.integer)):
+            if isinstance(idx.stop, (int, np.integer)):
                 length_pos_data = sys.maxsize
                 try:
                     length_pos_data = get_scalar_constant_value(
@@ -3281,12 +3280,10 @@ def merge_two_slices(slice1, len1, slice2, len2):
             n_val = sl1.stop - 1 - sl2 * sl1.step
             if config.warn.subtensor_merge_bug:
                 warnings.warning(
-                    (
-                        "Your current code is fine, but Theano versions "
-                        "prior to 0.5rc2 might have given an incorrect result. "
-                        "To disable this warning, set the Theano flag "
-                        "warn.subtensor_merge_bug to False."
-                    )
+                    "Your current code is fine, but Theano versions "
+                    "prior to 0.5rc2 might have given an incorrect result. "
+                    "To disable this warning, set the Theano flag "
+                    "warn.subtensor_merge_bug to False."
                 )
             # we need to pick either n_val or p_val and then follow same
             # steps as above for covering the index error cases
@@ -5467,7 +5464,7 @@ class Canonizer(LocalOptimizer):
         return getattr(
             self,
             "name",
-            "Canonizer(%s, %s, %s)" % (self.main, self.inverse, self.reciprocal),
+            "Canonizer({}, {}, {})".format(self.main, self.inverse, self.reciprocal),
         )
 
 
@@ -6125,17 +6122,15 @@ def local_reduce_join(node):
         if len(reduce_axis) != 1 or 0 not in reduce_axis:
             if theano.config.warn.reduce_join:
                 warnings.warning(
-                    (
-                        "Your current code is fine, but Theano versions "
-                        "prior to 0.7 (or this development version Sept 2014) "
-                        "might have given an incorrect result for this code. "
-                        "To disable this warning, set the Theano flag "
-                        "warn.reduce_join to False. The problem was an "
-                        "optimization, that modified the pattern "
-                        '"Reduce{scalar.op}(Join(axis=0, a, b), axis=0)", '
-                        "did not check the reduction axis. So if the "
-                        "reduction axis was not 0, you got a wrong answer."
-                    )
+                    "Your current code is fine, but Theano versions "
+                    "prior to 0.7 (or this development version Sept 2014) "
+                    "might have given an incorrect result for this code. "
+                    "To disable this warning, set the Theano flag "
+                    "warn.reduce_join to False. The problem was an "
+                    "optimization, that modified the pattern "
+                    '"Reduce{scalar.op}(Join(axis=0, a, b), axis=0)", '
+                    "did not check the reduction axis. So if the "
+                    "reduction axis was not 0, you got a wrong answer."
                 )
             return
 
@@ -7049,8 +7044,8 @@ def constant_folding(node):
         # The op asks not to be constant folded.
         return False
 
-    storage_map = dict([(i, [i.data]) for i in node.inputs])
-    compute_map = dict([(i, [True]) for i in node.inputs])
+    storage_map = {i: [i.data] for i in node.inputs}
+    compute_map = {i: [True] for i in node.inputs}
     for o in node.outputs:
         storage_map[o] = [None]
         compute_map[o] = [False]
@@ -7578,7 +7573,7 @@ def local_elemwise_fusion_op(op_class, max_input_fct=lambda node: 32, maker=None
             if (
                 i.owner
                 and isinstance(i.owner.op, op_class)
-                and len(set([n for n, idx in i.clients])) == 1
+                and len({n for n, idx in i.clients}) == 1
                 and
                 # Do not merge elemwise that don't have the same
                 # broadcastable pattern to don't redo duplicate
@@ -7789,7 +7784,6 @@ class FusionOptimizer(Optimizer):
                             nb_replacement += 1
                         except InconsistencyError:
                             nb_inconsistency_replace += 1
-                            pass
             nb_iter += 1
 
         if fgraph.profile:

theano/tensor/raw_random.py

@@ -6,7 +6,6 @@ from copy import copy
 from functools import reduce
 
 import numpy as np
-from six import string_types
 
 import theano
 from theano import gof, tensor
@@ -149,7 +148,7 @@ class RandomFunction(gof.Op):
             state = dct
         fn, outtype, inplace, ndim_added = state
         self.fn = fn
-        if isinstance(fn, string_types):
+        if isinstance(fn, str):
             self.exec_fn = getattr(np.random.RandomState, fn)
         else:
             self.exec_fn = fn
@@ -353,7 +352,7 @@ def _infer_ndim_bcast(ndim, shape, *args):
             else:
                 if s >= 0:
                     pre_v_shape.append(tensor.as_tensor_variable(s))
-                    bcast.append((s == 1))
+                    bcast.append(s == 1)
                 elif s == -1:
                     n_a_i = 0
                     for a in args:
@@ -370,11 +369,9 @@ def _infer_ndim_bcast(ndim, shape, *args):
                     else:
                         if n_a_i == 0:
                             raise ValueError(
-                                (
-                                    "Auto-shape of -1 must overlap"
-                                    "with the shape of one of the broadcastable"
-                                    "inputs"
-                                )
+                                "Auto-shape of -1 must overlap"
+                                "with the shape of one of the broadcastable"
+                                "inputs"
                             )
                         else:
                             pre_v_shape.append(tensor.as_tensor_variable(1))
@@ -393,7 +390,7 @@ def _infer_ndim_bcast(ndim, shape, *args):
         # but we need to know ndim
         if not args:
             raise TypeError(
-                ("_infer_ndim_bcast cannot infer shape without" " either shape or args")
+                "_infer_ndim_bcast cannot infer shape without" " either shape or args"
             )
         template = reduce(lambda a, b: a + b, args)
         v_shape = template.shape
@@ -957,7 +954,7 @@ optdb.register(
 )
 
 
-class RandomStreamsBase(object):
+class RandomStreamsBase:
     def binomial(self, size=None, n=1, p=0.5, ndim=None, dtype="int64", prob=None):
         """
         Sample n times with probability of success p for each trial and

theano/tensor/shared_randomstreams.py

@@ -59,7 +59,7 @@ class RandomStreams(raw_random.RandomStreamsBase):
         return list(self.state_updates)
 
     def __init__(self, seed=None):
-        super(RandomStreams, self).__init__()
+        super().__init__()
         # A list of pairs of the form (input_r, output_r).  This will be
         # over-ridden by the module instance to contain stream generators.
         self.state_updates = []

theano/tensor/sharedvar.py

 import traceback
 
 import numpy as np
-from six import integer_types
 
 import theano.tensor.basic
 from theano.compile import SharedVariable, shared_constructor
@@ -95,7 +94,7 @@ def scalar_constructor(
     if target != "cpu":
         raise TypeError("not for cpu")
 
-    if not isinstance(value, (np.number, float, integer_types, complex)):
+    if not isinstance(value, (np.number, float, int, complex)):
         raise TypeError()
     try:
         dtype = value.dtype

theano/tensor/signal/pool.py

@@ -470,7 +470,7 @@ class Pool(OpenMPOp):
         return rval
 
     def __init__(self, ignore_border=False, mode="max", ndim=2, openmp=None):
-        super(Pool, self).__init__(openmp=openmp)
+        super().__init__(openmp=openmp)
         self.ndim = ndim
         self.ignore_border = ignore_border
         if mode == "max_deterministic":
@@ -649,7 +649,7 @@ class Pool(OpenMPOp):
 
     def c_headers(self):
         headers = ["<algorithm>"]
-        headers += super(Pool, self).c_headers()
+        headers += super().c_headers()
         return headers
 
     def c_code(self, node, name, inp, out, sub):
@@ -1121,7 +1121,7 @@ class PoolGrad(OpenMPOp):
                 " 'average_inc_pad' and 'average_exc_pad'. Got %s" % mode
             )
         self.mode = mode
-        super(PoolGrad, self).__init__(openmp=openmp)
+        super().__init__(openmp=openmp)
 
     def prepare_node(self, node, storage_map, compute_map, impl):
         if len(node.inputs) < 5:  # 5 for AveragePoolGrad, 6 for MaxPoolGrad
@@ -1836,7 +1836,7 @@ class DownsampleFactorMaxGradGrad(OpenMPOp):
         self.ndim = ndim
         self.ignore_border = ignore_border
         self.mode = mode
-        super(DownsampleFactorMaxGradGrad, self).__init__(openmp=openmp)
+        super().__init__(openmp=openmp)
         assert self.mode == "max"
 
     def make_node(self, x, maxout, gz, ws, stride=None, pad=None):
@@ -2172,7 +2172,7 @@ class MaxPoolRop(OpenMPOp):
     )
 
     def __init__(self, ignore_border=False, mode="max", ndim=2, openmp=None):
-        super(MaxPoolRop, self).__init__(openmp=openmp)
+        super().__init__(openmp=openmp)
         self.ndim = ndim
         self.ignore_border = ignore_border
         self.mode = mode
@@ -2276,7 +2276,7 @@ class MaxPoolRop(OpenMPOp):
 
     def c_headers(self):
         headers = ["<algorithm>"]
-        headers += super(MaxPoolRop, self).c_headers()
+        headers += super().c_headers()
         return headers
 
     def c_code(self, node, name, inp, out, sub):

theano/tensor/sort.py

@@ -33,7 +33,7 @@ class SortOp(Op):
         self.order = order
 
     def __str__(self):
-        return self.__class__.__name__ + "{%s, %s}" % (self.kind, str(self.order))
+        return self.__class__.__name__ + "{{{}, {}}}".format(self.kind, str(self.order))
 
     def make_node(self, input, axis=-1):
         input = theano.tensor.as_tensor_variable(input)
@@ -168,7 +168,7 @@ class ArgSortOp(Op):
         self.order = order
 
     def __str__(self):
-        return self.__class__.__name__ + "{%s, %s}" % (self.kind, str(self.order))
+        return self.__class__.__name__ + "{{{}, {}}}".format(self.kind, str(self.order))
 
     def make_node(self, input, axis=-1):
         input = theano.tensor.as_tensor_variable(input)

theano/tensor/subtensor.py

@@ -5,7 +5,6 @@ from itertools import chain, groupby
 from textwrap import dedent
 
 import numpy as np
-from six import integer_types
 
 import theano
 from theano import config, gof
@@ -36,8 +35,6 @@ class AdvancedIndexingError(TypeError):
 
     """
 
-    pass
-
 
 def as_index_constant(a):
     """Convert Python literals to Theano constants--when possible--in Subtensor arguments.
@@ -52,7 +49,7 @@ def as_index_constant(a):
             as_index_constant(a.stop),
             as_index_constant(a.step),
         )
-    elif isinstance(a, (integer_types, np.integer)):
+    elif isinstance(a, (int, np.integer)):
         return scal.ScalarConstant(scal.int64, a)
     elif not isinstance(a, theano.tensor.Variable):
         return theano.tensor.as_tensor(a)
@@ -537,7 +534,7 @@ class Subtensor(Op):
                 slice_c = None
 
             return slice(slice_a, slice_b, slice_c)
-        elif isinstance(entry, (integer_types, np.integer)):
+        elif isinstance(entry, (int, np.integer)):
             # Disallow the use of python scalars in idx_list
             raise TypeError(
                 "Python scalar in idx_list." "Please report this error to theano-dev."
@@ -662,7 +659,7 @@ class Subtensor(Op):
                         if start is None:
                             start = 0
                         if p.stop is None or (
-                            isinstance(p.stop, (integer_types, np.integer, np.ndarray))
+                            isinstance(p.stop, (int, np.integer, np.ndarray))
                             and p.stop > start
                         ):
                             broadcastable.append(True)
@@ -778,7 +775,7 @@ class Subtensor(Op):
                 indices.append(self.str_from_slice(entry))
             else:
                 indices.append(str(entry))
-        return "%s{%s}" % (self.__class__.__name__, ", ".join(indices))
+        return "{}{{{}}}".format(self.__class__.__name__, ", ".join(indices))
 
     @staticmethod
     def default_helper_c_code_args():
@@ -844,7 +841,7 @@ class Subtensor(Op):
             return pos[1]
 
         def init_entry(entry, depth=0):
-            if isinstance(entry, (np.integer, integer_types)):
+            if isinstance(entry, (np.integer, int)):
                 init_cmds.append("subtensor_spec[%i] = %i;" % (spec_pos(), entry))
                 inc_spec_pos(1)
                 if depth == 0:
@@ -1144,7 +1141,7 @@ class SubtensorPrinter:
                 pstate.precedence = -1000
 
                 for entry in idxs:
-                    if isinstance(entry, integer_types):
+                    if isinstance(entry, int):
                         sidxs.append(str(entry))
                     elif isinstance(entry, scal.Scalar):
                         sidxs.append(pstate.pprinter.process(inputs.pop()))
@@ -1164,7 +1161,7 @@ class SubtensorPrinter:
                         else:
                             msg3 = ":%s" % entry.step
 
-                        sidxs.append("%s:%s%s" % (msg1, msg2, msg3))
+                        sidxs.append("{}:{}{}".format(msg1, msg2, msg3))
             finally:
                 pstate.precedence = old_precedence
 
@@ -1173,7 +1170,7 @@ class SubtensorPrinter:
                 sub = pstate.pprinter.process(input, pstate)
             finally:
                 pstate.precedence = old_precedence
-            return "%s[%s]" % (sub, ", ".join(sidxs))
+            return "{}[{}]".format(sub, ", ".join(sidxs))
         else:
             raise TypeError("Can only print Subtensor.")
 
@@ -1464,7 +1461,7 @@ class IncSubtensor(Op):
             msg += "Inc"
         else:
             msg += "Set"
-        return "%s{%s;%s}" % (self.__class__.__name__, msg, ", ".join(indices))
+        return "{}{{{};{}}}".format(self.__class__.__name__, msg, ", ".join(indices))
 
     def make_node(self, x, y, *inputs):
         """
@@ -2430,7 +2427,7 @@ class AdvancedIncSubtensor(Op):
             raise NotImplementedError("In place computation is not" " implemented")
 
     def __str__(self):
-        return "%s{%s, %s}" % (
+        return "{}{{{}, {}}}".format(
             self.__class__.__name__,
             "inplace=" + str(self.inplace),
             " set_instead_of_inc=" + str(self.set_instead_of_inc),

theano/tensor/type.py

@@ -283,7 +283,9 @@ class TensorType(Type):
             }[self.dtype]
         except KeyError:
             raise TypeError(
-                "Unsupported dtype for %s: %s" % (self.__class__.__name__, self.dtype)
+                "Unsupported dtype for {}: {}".format(
+                    self.__class__.__name__, self.dtype
+                )
             )
 
     def to_scalar_type(self):
@@ -391,7 +393,7 @@ class TensorType(Type):
                     bcast = str(b)
                 else:
                     bcast = "%iD" % len(b)
-            return "TensorType(%s, %s)" % (str(self.dtype), bcast)
+            return "TensorType({}, {})".format(self.dtype, bcast)
 
     def __repr__(self):
         return str(self)

theano/tensor/type_other.py

@@ -94,7 +94,7 @@ class SliceConstant(Constant):
         return (SliceConstant, self.data.start, self.data.stop, self.data.step)
 
     def __str__(self):
-        return "%s{%s, %s, %s}" % (
+        return "{}{{{}, {}, {}}}".format(
             self.__class__.__name__,
             self.data.start,
             self.data.stop,

theano/tensor/var.py

@@ -4,7 +4,6 @@ import warnings
 from collections.abc import Iterable
 
 import numpy as np
-from six import integer_types
 
 import theano
 from theano import config
@@ -15,7 +14,7 @@ from theano.tensor.type import TensorType
 from theano.tensor.utils import hash_from_ndarray
 
 
-class _tensor_py_operators(object):
+class _tensor_py_operators:
     def __abs__(self):
         return theano.tensor.basic.abs_(self)
 
@@ -297,7 +296,7 @@ class _tensor_py_operators(object):
         """
 
         if ndim is not None:
-            if not isinstance(ndim, integer_types):
+            if not isinstance(ndim, int):
                 raise ValueError(
                     "Expected ndim to be an integer, is " + str(type(ndim))
                 )
@@ -640,11 +639,9 @@ class _tensor_py_operators(object):
         except TypeError:
             # This prevents accidental iteration via sum(self)
             raise TypeError(
-                (
-                    "TensorType does not support iteration. "
-                    "Maybe you are using builtins.sum instead of "
-                    "theano.tensor.sum? (Maybe .max?)"
-                )
+                "TensorType does not support iteration. "
+                "Maybe you are using builtins.sum instead of "
+                "theano.tensor.sum? (Maybe .max?)"
             )
 
     ndim = property(lambda self: self.type.ndim)
@@ -839,7 +836,7 @@ class TensorVariable(_tensor_py_operators, Variable):
     """
 
     def __init__(self, type, owner=None, index=None, name=None):
-        super(TensorVariable, self).__init__(type, owner=owner, index=index, name=name)
+        super().__init__(type, owner=owner, index=index, name=name)
         if config.warn_float64 != "ignore" and type.dtype == "float64":
             msg = (
                 "You are creating a TensorVariable "
@@ -997,7 +994,7 @@ class TensorConstant(_tensor_py_operators, Constant):
 
     def __str__(self):
         if self.tag.unique_value is not None:
-            name = "%s of %s" % (str(self.data.shape), str(self.tag.unique_value))
+            name = "{} of {}".format(str(self.data.shape), str(self.tag.unique_value))
         else:
             name = "%s" % self.data
         if len(name) > 20: