Add support for ruff

.pre-commit-config.yaml

@@ -33,6 +33,8 @@ repos:
     rev: 6.0.0
     hooks:
       - id: flake8
+        additional_dependencies:
+          - flake8-comprehensions
   - repo: https://github.com/pycqa/isort
     rev: 5.12.0
     hooks:

pyproject.toml

@@ -172,6 +172,34 @@ skip_gitignore = true
 skip = "pytensor/version.py"
 skip_glob = "**/*.pyx"
 
+[tool.ruff]
+select=["C","E","F","W"]
+ignore=["E501","E741","C408","C901"]
+exclude = [
+    "doc/",
+    "pytensor/_version.py",
+    "bin/pytensor_cache.py",
+]
+
+[tool.ruff.per-file-ignores]
+# TODO: Get rid of these:
+    "**/__init__.py"=["F401","E402","F403"]
+    "pytensor/tensor/linalg.py"=["F401","F403"]
+    "pytensor/scalar/basic_scipy.py"=["E402","F403","F401"]
+    "pytensor/graph/toolbox.py"=["E402","F403","F401"]
+    "pytensor/link/jax/jax_dispatch.py"=["E402","F403","F401"]
+    "pytensor/link/jax/jax_linker.py"=["E402","F403","F401"]
+    "pytensor/sparse/sandbox/sp2.py"=["F401"]
+    "tests/tensor/test_math_scipy.py"=["E402"]
+    "tests/sparse/test_basic.py"=["E402"]
+    "tests/sparse/test_opt.py"=["E402"]
+    "tests/sparse/test_sp2.py"=["E402"]
+    "tests/sparse/test_utils.py"=["E402","F401"]
+    "tests/sparse/sandbox/test_sp.py"=["E402","F401"]
+    "tests/scalar/test_basic_sympy.py"=["E402"]
+    "pytensor/graph/rewriting/unify.py"=["F811"]
+
+
 [tool.mypy]
 ignore_missing_imports = true
 no_implicit_optional = true

pytensor/compile/builders.py

@@ -969,9 +969,7 @@ def inline_ofg_expansion(fgraph, node):
         return False
     if not op.is_inline:
         return False
-    return clone_replace(
-        op.inner_outputs, {u: v for u, v in zip(op.inner_inputs, node.inputs)}
-    )
+    return clone_replace(op.inner_outputs, dict(zip(op.inner_inputs, node.inputs)))
 
 
 # We want to run this before the first merge optimizer

pytensor/gradient.py

@@ -504,7 +504,7 @@ def grad(
     if not isinstance(wrt, Sequence):
         _wrt: List[Variable] = [wrt]
     else:
-        _wrt = [x for x in wrt]
+        _wrt = list(wrt)
 
     outputs = []
     if cost is not None:
@@ -791,8 +791,8 @@ def subgraph_grad(wrt, end, start=None, cost=None, details=False):
 
     pgrads = dict(zip(params, grads))
     # separate wrt from end grads:
-    wrt_grads = list(pgrads[k] for k in wrt)
-    end_grads = list(pgrads[k] for k in end)
+    wrt_grads = [pgrads[k] for k in wrt]
+    end_grads = [pgrads[k] for k in end]
 
     if details:
         return wrt_grads, end_grads, start_grads, cost_grads

pytensor/graph/basic.py

@@ -1629,7 +1629,7 @@ def as_string(
                 multi.add(op2)
             else:
                 seen.add(input.owner)
-    multi_list = [x for x in multi]
+    multi_list = list(multi)
     done: Set = set()
 
     def multi_index(x):

pytensor/graph/replace.py

@@ -142,7 +142,7 @@ def toposort_key(fg: FunctionGraph, ts, pair):
                 raise ValueError(f"{key} is not a part of graph")
 
     sorted_replacements = sorted(
-        tuple(fg_replace.items()),
+        fg_replace.items(),
         # sort based on the fg toposort, if a variable has no owner, it goes first
         key=partial(toposort_key, fg, toposort),
         reverse=True,

pytensor/graph/rewriting/basic.py

@@ -2575,8 +2575,8 @@ def print_profile(cls, stream, prof, level=0):
         for i in range(len(loop_timing)):
             loop_times = ""
             if loop_process_count[i]:
-                d = list(
-                    reversed(sorted(loop_process_count[i].items(), key=lambda a: a[1]))
+                d = sorted(
+                    loop_process_count[i].items(), key=lambda a: a[1], reverse=True
                 )
                 loop_times = " ".join([str((str(k), v)) for k, v in d[:5]])
                 if len(d) > 5:

pytensor/link/c/basic.py

@@ -633,11 +633,11 @@ def fetch_variables(self):
 
         # The orphans field is listified to ensure a consistent order.
         # list(fgraph.orphans.difference(self.outputs))
-        self.orphans = list(
+        self.orphans = [
             r
             for r in self.variables
             if isinstance(r, AtomicVariable) and r not in self.inputs
-        )
+        ]
         # C type constants (pytensor.scalar.ScalarType). They don't request an object
         self.consts = []
         # Move c type from orphans (pytensor.scalar.ScalarType) to self.consts

pytensor/link/c/params_type.py

@@ -810,7 +810,7 @@ def c_support_code(self, **kwargs):
             struct_extract_method=struct_extract_method,
         )
 
-        return list(sorted(list(c_support_code_set))) + [final_struct_code]
+        return sorted(c_support_code_set) + [final_struct_code]
 
     def c_code_cache_version(self):
         return ((3,), tuple(t.c_code_cache_version() for t in self.types))

pytensor/link/jax/dispatch/elemwise.py

@@ -41,7 +41,7 @@ def careduce(x):
         elif scalar_op_name:
             scalar_fn_name = scalar_op_name
 
-        to_reduce = reversed(sorted(axis))
+        to_reduce = sorted(axis, reverse=True)
 
         if to_reduce:
             # In this case, we need to use the `jax.lax` function (if there

pytensor/link/numba/dispatch/elemwise.py

@@ -361,7 +361,7 @@ def careduce_maximum(input):
 
     careduce_fn_name = f"careduce_{scalar_op}"
     global_env = {}
-    to_reduce = reversed(sorted(axes))
+    to_reduce = sorted(axes, reverse=True)
     careduce_lines_src = []
     var_name = input_name
 

pytensor/printing.py

@@ -796,7 +796,7 @@ def grad(self, input, output_gradients):
         return output_gradients
 
     def R_op(self, inputs, eval_points):
-        return [x for x in eval_points]
+        return list(eval_points)
 
     def __setstate__(self, dct):
         dct.setdefault("global_fn", _print_fn)

pytensor/scan/basic.py

@@ -492,7 +492,7 @@ def wrap_into_list(x):
     # wrap sequences in a dictionary if they are not already dictionaries
     for i in range(n_seqs):
         if not isinstance(seqs[i], dict):
-            seqs[i] = dict([("input", seqs[i]), ("taps", [0])])
+            seqs[i] = {"input": seqs[i], "taps": [0]}
         elif seqs[i].get("taps", None) is not None:
             seqs[i]["taps"] = wrap_into_list(seqs[i]["taps"])
         elif seqs[i].get("taps", None) is None:
@@ -504,7 +504,7 @@ def wrap_into_list(x):
         if outs_info[i] is not None:
             if not isinstance(outs_info[i], dict):
                 # by default any output has a tap value of -1
-                outs_info[i] = dict([("initial", outs_info[i]), ("taps", [-1])])
+                outs_info[i] = {"initial": outs_info[i], "taps": [-1]}
             elif (
                 outs_info[i].get("initial", None) is None
                 and outs_info[i].get("taps", None) is not None

pytensor/scan/op.py

@@ -1718,12 +1718,9 @@ def perform(self, node, inputs, output_storage, params=None):
             arg.shape[0]
             for arg in inputs[self.seqs_arg_offset : self.shared_arg_offset]
         ]
-        store_steps += [
-            arg
-            for arg in inputs[
-                self.nit_sot_arg_offset : self.nit_sot_arg_offset + info.n_nit_sot
-            ]
-        ]
+        store_steps += list(
+            inputs[self.nit_sot_arg_offset : self.nit_sot_arg_offset + info.n_nit_sot]
+        )
 
         # 2.1 Create storage space for outputs
         for idx in range(self.n_outs):
@@ -2270,7 +2267,7 @@ def infer_shape(self, fgraph, node, input_shapes):
         )
 
         offset = 1 + info.n_seqs
-        scan_outs = [x for x in input_shapes[offset : offset + n_outs]]
+        scan_outs = list(input_shapes[offset : offset + n_outs])
         offset += n_outs
         outs_shape_n = info.n_mit_mot_outs + info.n_mit_sot + info.n_sit_sot
         for x in range(info.n_nit_sot):
@@ -2301,7 +2298,7 @@ def infer_shape(self, fgraph, node, input_shapes):
                         shp.append(v_shp_i[0])
                 scan_outs.append(tuple(shp))
 
-        scan_outs += [x for x in input_shapes[offset : offset + info.n_shared_outs]]
+        scan_outs += list(input_shapes[offset : offset + info.n_shared_outs])
         # if we are dealing with a repeat-until, then we do not know the
         # leading dimension so we replace it for every entry with Shape_i
         if info.as_while:

pytensor/scan/rewriting.py

@@ -388,7 +388,7 @@ def add_to_replace(y):
             if out in local_fgraph_outs_set:
                 x = node.outputs[local_fgraph_outs_map[out]]
                 y = replace_with_out[idx]
-                y_shape = [shp for shp in y.shape]
+                y_shape = list(y.shape)
                 replace_with[x] = at.alloc(y, node.inputs[0], *y_shape)
 
         # We need to add one extra dimension to the outputs

pytensor/tensor/random/op.py

@@ -283,7 +283,7 @@ def infer_shape(self, fgraph, node, input_shapes):
 
         shape = self._infer_shape(size, dist_params, param_shapes=param_shapes)
 
-        return [None, [s for s in shape]]
+        return [None, list(shape)]
 
     def __call__(self, *args, size=None, name=None, rng=None, dtype=None, **kwargs):
         res = super().__call__(rng, size, dtype, *args, **kwargs)

pytensor/tensor/rewriting/math.py

@@ -1555,11 +1555,11 @@ def local_sum_prod_div_dimshuffle(fgraph, node):
                     )
 
                 if len(compatible_dims) > 0:
-                    optimized_dimshuffle_order = list(
+                    optimized_dimshuffle_order = [
                         ax
                         for i, ax in enumerate(dimshuffle_order)
                         if (i not in axis) or (ax != "x")
-                    )
+                    ]
 
                     # Removing leading 'x' (since it will be done automatically)
                     while (
@@ -1644,7 +1644,7 @@ def local_op_of_op(fgraph, node):
                     return [op_type(None, dtype=out_dtype)(node_inps.owner.inputs[0])]
 
                 # figure out which axes were in the original sum
-                newaxis = list(tuple(node_inps.owner.op.axis))
+                newaxis = list(node_inps.owner.op.axis)
                 for i in node.op.axis:
                     new_i = i
                     for ii in node_inps.owner.op.axis:

pytensor/tensor/shape.py

@@ -810,7 +810,7 @@ def shape_padleft(t, n_ones=1):
     """
     _t = at.as_tensor_variable(t)
 
-    pattern = ["x"] * n_ones + [i for i in range(_t.type.ndim)]
+    pattern = ["x"] * n_ones + list(range(_t.type.ndim))
     return _t.dimshuffle(pattern)
 
 
@@ -826,7 +826,7 @@ def shape_padright(t, n_ones=1):
     """
     _t = at.as_tensor_variable(t)
 
-    pattern = [i for i in range(_t.type.ndim)] + ["x"] * n_ones
+    pattern = list(range(_t.type.ndim)) + ["x"] * n_ones
     return _t.dimshuffle(pattern)
 
 
@@ -861,7 +861,7 @@ def shape_padaxis(t, axis):
     if axis < 0:
         axis += ndim
 
-    pattern = [i for i in range(_t.type.ndim)]
+    pattern = list(range(_t.type.ndim))
     pattern.insert(axis, "x")
     return _t.dimshuffle(pattern)
 

pytensor/tensor/subtensor.py

@@ -2604,7 +2604,7 @@ def infer_shape(self, fgraph, node, ishapes):
             ishapes[0], index_shapes, indices_are_shapes=True
         )
         assert node.outputs[0].ndim == len(res_shape)
-        return [[s for s in res_shape]]
+        return [list(res_shape)]
 
     def perform(self, node, inputs, out_):
         (out,) = out_

setup.cfg

@@ -1,7 +1,6 @@
 [flake8]
 select = C,E,F,W
-ignore = E203,E231,E501,E741,W503,W504,C901
-max-line-length = 88
+ignore = E203,E231,E501,E741,W503,W504,C408,C901
 per-file-ignores =
     **/__init__.py:F401,E402,F403
     pytensor/tensor/linalg.py:F401,F403

tests/graph/test_features.py

@@ -73,7 +73,7 @@ def inputs():
 
         assert hasattr(g, "get_nodes")
         for type, num in ((add, 3), (sigmoid, 3), (dot, 2)):
-            if len([t for t in g.get_nodes(type)]) != num:
+            if len(list(g.get_nodes(type))) != num:
                 raise Exception("Expected: %i times %s" % (num, type))
         new_e0 = add(y, z)
         assert e0.owner in g.get_nodes(dot)
@@ -82,7 +82,7 @@ def inputs():
         assert e0.owner not in g.get_nodes(dot)
         assert new_e0.owner in g.get_nodes(add)
         for type, num in ((add, 4), (sigmoid, 3), (dot, 1)):
-            if len([t for t in g.get_nodes(type)]) != num:
+            if len(list(g.get_nodes(type))) != num:
                 raise Exception("Expected: %i times %s" % (num, type))
 
 

tests/graph/test_op.py

@@ -87,7 +87,7 @@ def test_sanity_0(self):
         r1, r2 = MyType(1)(), MyType(2)()
         node = MyOp.make_node(r1, r2)
         # Are the inputs what I provided?
-        assert [x for x in node.inputs] == [r1, r2]
+        assert list(node.inputs) == [r1, r2]
         # Are the outputs what I expect?
         assert [x.type for x in node.outputs] == [MyType(3)]
         assert node.outputs[0].owner is node and node.outputs[0].index == 0

tests/tensor/rewriting/test_elemwise.py

@@ -1123,7 +1123,7 @@ def test_add_mul_fusion_inplace(self):
         out = dot(x, y) + x + y + z
 
         f = function([x, y, z], out, mode=self.mode)
-        topo = [n for n in f.maker.fgraph.toposort()]
+        topo = list(f.maker.fgraph.toposort())
         assert len(topo) == 2
         assert topo[-1].op.inplace_pattern
 

tests/tensor/rewriting/test_math.py

@@ -3994,9 +3994,9 @@ def test_is_1pexp(self):
         exp_op = exp
         assert is_1pexp(1 + exp_op(x), False) == (False, x)
         assert is_1pexp(exp_op(x) + 1, False) == (False, x)
-        for neg_, exp_arg in map(
-            lambda x: is_1pexp(x, only_process_constants=False),
-            [(1 + exp_op(-x)), (exp_op(-x) + 1)],
+        for neg_, exp_arg in (
+            is_1pexp(x, only_process_constants=False)
+            for x in [(1 + exp_op(-x)), (exp_op(-x) + 1)]
         ):
             assert not neg_ and is_same_graph(exp_arg, -x)
         assert is_1pexp(1 - exp_op(x), False) is None

tests/tensor/rewriting/test_subtensor.py

@@ -2004,7 +2004,7 @@ def test_local_subtensor_SpecifyShape_lift(x, s, idx, x_val, s_val):
     y_val_fn = function(
         [x] + list(s), y, on_unused_input="ignore", mode=no_rewrites_mode
     )
-    y_val = y_val_fn(*([x_val] + [s_ for s_ in s_val]))
+    y_val = y_val_fn(*([x_val] + list(s_val)))
 
     # This optimization should appear in the canonicalizations
     y_opt = rewrite_graph(y, clone=False)
@@ -2017,7 +2017,7 @@ def test_local_subtensor_SpecifyShape_lift(x, s, idx, x_val, s_val):
         assert isinstance(y_opt.owner.op, SpecifyShape)
 
     y_opt_fn = function([x] + list(s), y_opt, on_unused_input="ignore")
-    y_opt_val = y_opt_fn(*([x_val] + [s_ for s_ in s_val]))
+    y_opt_val = y_opt_fn(*([x_val] + list(s_val)))
 
     assert np.allclose(y_val, y_opt_val)
 

tests/tensor/test_blas.py

@@ -2589,10 +2589,10 @@ def test_ger(self):
     op=batched_dot,
     expected=(
         lambda xs, ys: np.asarray(
-            list(
+            [
                 x * y if x.ndim == 0 or y.ndim == 0 else np.dot(x, y)
                 for x, y in zip(xs, ys)
-            ),
+            ],
             dtype=aes.upcast(xs.dtype, ys.dtype),
         )
     ),
@@ -2694,7 +2694,7 @@ def check_first_dim(inverted):
         assert x.strides[0] == direction * np.dtype(config.floatX).itemsize
         assert not (x.flags["C_CONTIGUOUS"] or x.flags["F_CONTIGUOUS"])
         result = f(x, w)
-        ref_result = np.asarray(list(np.dot(u, v) for u, v in zip(x, w)))
+        ref_result = np.asarray([np.dot(u, v) for u, v in zip(x, w)])
         utt.assert_allclose(ref_result, result)
 
     for inverted in (0, 1):

tests/tensor/test_complex.py

@@ -15,9 +15,7 @@ def test_basic(self):
         x = zvector()
         rng = np.random.default_rng(23)
         xval = np.asarray(
-            list(
-                complex(rng.standard_normal(), rng.standard_normal()) for i in range(10)
-            )
+            [complex(rng.standard_normal(), rng.standard_normal()) for i in range(10)]
         )
         assert np.all(xval.real == pytensor.function([x], real(x))(xval))
         assert np.all(xval.imag == pytensor.function([x], imag(x))(xval))