def numba_funcify_CAReduce(op, node, **kwargs):
axes = op.axis
if axes is None:
axes = list(range(node.inputs[0].ndim))
if hasattr(op, "acc_dtype") and op.acc_dtype is not None:
acc_dtype = op.acc_dtype
else:
acc_dtype = node.outputs[0].type.dtype
np_acc_dtype = np.dtype(acc_dtype)
scalar_op_identity = op.scalar_op.identity
if np_acc_dtype.kind == "i" and not np.isfinite(scalar_op_identity):
if np.isposinf(scalar_op_identity):
scalar_op_identity = np.iinfo(np_acc_dtype).max
else:
scalar_op_identity = np.iinfo(np_acc_dtype).min
# Make sure it has the correct dtype
scalar_op_identity = np.array(scalar_op_identity, dtype=np_acc_dtype)
input_name = get_name_for_object(node.inputs[0])
ndim = node.inputs[0].ndim
careduce_py_fn = create_multiaxis_reducer(
op.scalar_op,
scalar_op_identity,
axes,
ndim,
np.dtype(node.outputs[0].type.dtype),
input_name=input_name,
)
careduce_fn = jit_compile_reducer(node, careduce_py_fn)
return careduce_fn
def numba_funcify_ScalarOp(op, node, **kwargs):
# TODO: Do we need to cache these functions so that we don't end up
# compiling the same Numba function over and over again?
scalar_func_name = op.nfunc_spec[0]
if scalar_func_name.startswith("scipy."):
func_package = scipy
scalar_func_name = scalar_func_name.split(".", 1)[-1]
else:
func_package = np
if "." in scalar_func_name:
scalar_func = reduce(getattr, [scipy] + scalar_func_name.split("."))
else:
scalar_func = getattr(func_package, scalar_func_name)
input_names = ", ".join([v.auto_name for v in node.inputs])
global_env = {"scalar_func": scalar_func}
scalar_op_fn_name = get_name_for_object(scalar_func)
scalar_op_src = f"""
def {scalar_op_fn_name}({input_names}):
return scalar_func({input_names})
"""
scalar_op_fn = compile_function_src(scalar_op_src, scalar_op_fn_name,
global_env)
return numba.njit(scalar_op_fn)
def numba_funcify_CAReduce(op, node, **kwargs):
axes = op.axis
if axes is None:
axes = list(range(node.inputs[0].ndim))
if hasattr(op, "acc_dtype") and op.acc_dtype is not None:
acc_dtype = op.acc_dtype
else:
acc_dtype = node.outputs[0].type.dtype
np_acc_dtype = np.dtype(acc_dtype)
scalar_op_identity = np.asarray(op.scalar_op.identity, dtype=np_acc_dtype)
input_name = get_name_for_object(node.inputs[0])
ndim = node.inputs[0].ndim
careduce_py_fn = create_multiaxis_reducer(
op.scalar_op,
scalar_op_identity,
axes,
ndim,
np.dtype(node.outputs[0].type.dtype),
input_name=input_name,
)
careduce_fn = jit_compile_reducer(node, careduce_py_fn)
return careduce_fn
def test_fgraph_to_python_names():
import inspect
x = scalar("1x")
y = scalar("_")
z = scalar()
q = scalar("def")
r = NoneConst
out_fg = FunctionGraph([x, y, z, q, r], [x, y, z, q, r], clone=False)
out_jx = fgraph_to_python(out_fg, to_python)
sig = inspect.signature(out_jx)
assert (x.auto_name, "_", z.auto_name, q.auto_name,
r.name) == tuple(sig.parameters.keys())
assert (1, 2, 3, 4, 5) == out_jx(1, 2, 3, 4, 5)
obj = object()
assert get_name_for_object(obj) == type(obj).__name__
def numba_funcify_CAReduce(op, node, **kwargs):
axes = op.axis
if axes is None:
axes = list(range(node.inputs[0].ndim))
if hasattr(op, "acc_dtype") and op.acc_dtype is not None:
acc_dtype = op.acc_dtype
else:
acc_dtype = node.outputs[0].type.dtype
np_acc_dtype = np.dtype(acc_dtype)
scalar_op_identity = np.asarray(op.scalar_op.identity, dtype=np_acc_dtype)
acc_dtype = numba.np.numpy_support.from_dtype(np_acc_dtype)
scalar_nfunc_spec = op.scalar_op.nfunc_spec
# We construct a dummy `Apply` that has the minimum required number of
# inputs for the scalar `Op`. Without this, we would get a scalar function
# with too few arguments.
dummy_node = Apply(
op,
[tensor(acc_dtype, [False]) for i in range(scalar_nfunc_spec[1])],
[tensor(acc_dtype, [False]) for o in range(scalar_nfunc_spec[2])],
)
elemwise_fn = numba_funcify_Elemwise(op,
dummy_node,
use_signature=True,
**kwargs)
input_name = get_name_for_object(node.inputs[0])
ndim = node.inputs[0].ndim
careduce_fn = create_multiaxis_reducer(elemwise_fn,
scalar_op_identity,
axes,
ndim,
acc_dtype,
input_name=input_name)
return numba.njit(careduce_fn)
def numba_funcify_ScalarOp(op, node, **kwargs):
# TODO: Do we need to cache these functions so that we don't end up
# compiling the same Numba function over and over again?
scalar_func_name = op.nfunc_spec[0]
if scalar_func_name.startswith("scipy."):
func_package = scipy
scalar_func_name = scalar_func_name.split(".", 1)[-1]
else:
func_package = np
if "." in scalar_func_name:
scalar_func = reduce(getattr, [scipy] + scalar_func_name.split("."))
else:
scalar_func = getattr(func_package, scalar_func_name)
scalar_op_fn_name = get_name_for_object(scalar_func)
unique_names = unique_name_generator([scalar_op_fn_name, "scalar_func"],
suffix_sep="_")
input_names = ", ".join(
[unique_names(v, force_unique=True) for v in node.inputs])
global_env = {"scalar_func": scalar_func}
scalar_op_src = f"""
def {scalar_op_fn_name}({input_names}):
return scalar_func({input_names})
"""
scalar_op_fn = compile_function_src(scalar_op_src, scalar_op_fn_name, {
**globals(),
**global_env
})
signature = create_numba_signature(node, force_scalar=True)
return numba_basic.numba_njit(
signature, inline="always",
fastmath=config.numba__fastmath)(scalar_op_fn)
def numba_funcify_ScalarOp(op, node, **kwargs):
# TODO: Do we need to cache these functions so that we don't end up
# compiling the same Numba function over and over again?
scalar_func_name = op.nfunc_spec[0]
if scalar_func_name.startswith("scipy."):
func_package = scipy
scalar_func_name = scalar_func_name.split(".", 1)[-1]
else:
func_package = np
if "." in scalar_func_name:
scalar_func = reduce(getattr, [scipy] + scalar_func_name.split("."))
else:
scalar_func = getattr(func_package, scalar_func_name)
scalar_op_fn_name = get_name_for_object(scalar_func)
unique_names = unique_name_generator([scalar_op_fn_name, "scalar_func"],
suffix_sep="_")
global_env = {"scalar_func": scalar_func}
input_tmp_dtypes = None
if func_package == scipy and hasattr(scalar_func, "types"):
# The `numba-scipy` bindings don't provide implementations for all
# inputs types, so we need to convert the inputs to floats and back.
inp_dtype_kinds = tuple(
np.dtype(inp.type.dtype).kind for inp in node.inputs)
accepted_inp_kinds = tuple(
sig_type.split("->")[0] for sig_type in scalar_func.types)
if not any(
all(dk == ik for dk, ik in zip(inp_dtype_kinds, ok_kinds))
for ok_kinds in accepted_inp_kinds):
# They're usually ordered from lower-to-higher precision, so
# we pick the last acceptable input types
#
# XXX: We should pick the first acceptable float/int types in
# reverse, excluding all the incompatible ones (e.g. `"0"`).
# The assumption is that this is only used by `numba-scipy`-exposed
# functions, although it's possible for this to be triggered by
# something else from the `scipy` package
input_tmp_dtypes = tuple(
np.dtype(k) for k in accepted_inp_kinds[-1])
if input_tmp_dtypes is None:
unique_names = unique_name_generator(
[scalar_op_fn_name, "scalar_func"], suffix_sep="_")
input_names = ", ".join(
[unique_names(v, force_unique=True) for v in node.inputs])
scalar_op_src = f"""
def {scalar_op_fn_name}({input_names}):
return scalar_func({input_names})
"""
else:
global_env["direct_cast"] = numba_basic.direct_cast
global_env["output_dtype"] = np.dtype(node.outputs[0].type.dtype)
input_tmp_dtype_names = {
f"inp_tmp_dtype_{i}": i_dtype
for i, i_dtype in enumerate(input_tmp_dtypes)
}
global_env.update(input_tmp_dtype_names)
unique_names = unique_name_generator(
[scalar_op_fn_name, "scalar_func"] + list(global_env.keys()),
suffix_sep="_")
input_names = [unique_names(v, force_unique=True) for v in node.inputs]
converted_call_args = ", ".join([
f"direct_cast({i_name}, {i_tmp_dtype_name})"
for i_name, i_tmp_dtype_name in zip(input_names,
input_tmp_dtype_names.keys())
])
scalar_op_src = f"""
def {scalar_op_fn_name}({', '.join(input_names)}):
return direct_cast(scalar_func({converted_call_args}), output_dtype)
"""
scalar_op_fn = compile_function_src(scalar_op_src, scalar_op_fn_name, {
**globals(),
**global_env
})
signature = create_numba_signature(node, force_scalar=True)
return numba_basic.numba_njit(
signature, inline="always",
fastmath=config.numba__fastmath)(scalar_op_fn)
