Undocumented
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
No class docstring; 1/1 method documented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Class |
|
Undocumented |
Function | check |
Undocumented |
Function | floordiv |
Undocumented |
Function | recursionlimit |
Undocumented |
Function | test |
Undocumented |
Function | test |
This is a thorough test attempting to cover important promotion paths and ensuring that arrays and scalars stay as aligned as possible. However, if it creates troubles, it should maybe just be removed. |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
Undocumented |
Function | test |
This test covers scalar subclass deferral. Note that this is exceedingly complicated, especially since it tends to fall back to the array paths and these additionally add the "array priority" mechanism. |
Variable | complex |
Undocumented |
Variable | floating |
Undocumented |
Variable | objecty |
Undocumented |
Variable | ops |
Undocumented |
Variable | reasonable |
Undocumented |
Variable | types |
Undocumented |
Function | _signs |
Undocumented |
@given(
def test_array_scalar_ufunc_dtypes(op, dt1, dt2): (source) ¶
Undocumented
@settings(
@given(
def test_array_scalar_ufunc_equivalence(op, arr1, arr2): (source) ¶
This is a thorough test attempting to cover important promotion paths and ensuring that arrays and scalars stay as aligned as possible. However, if it creates troubles, it should maybe just be removed.
@pytest.mark.parametrize(
def test_clongdouble_inf_loop(op, val): (source) ¶
Undocumented
def test_int_float_promotion_truediv(fscalar): (source) ¶
Undocumented
@pytest.mark.parametrize(
def test_longdouble_inf_loop(op, val): (source) ¶
Undocumented
def test_operator_object_left(o, op, type_): (source) ¶
Undocumented
def test_operator_object_right(o, op, type_): (source) ¶
Undocumented
def test_operator_scalars(op, type1, type2): (source) ¶
Undocumented
@pytest.mark.parametrize(
@np._no_nep50_warning()
def test_pyscalar_subclasses(subtype, __op__, __rop__, op, cmp): (source) ¶
Undocumented
@pytest.mark.parametrize(
def test_scalar_integer_operation_divbyzero(dtype, operation): (source) ¶
Undocumented
@pytest.mark.parametrize(
def test_scalar_integer_operation_overflow(dtype, operation): (source) ¶
Undocumented
@pytest.mark.parametrize(
def test_scalar_signed_integer_overflow(dtype, operation): (source) ¶
Undocumented
def test_scalar_unsigned_integer_overflow(dtype): (source) ¶
Undocumented
@pytest.mark.parametrize(
def test_subclass_deferral(sctype, __op__, __rop__, op, cmp): (source) ¶
This test covers scalar subclass deferral. Note that this is exceedingly complicated, especially since it tends to fall back to the array paths and these additionally add the "array priority" mechanism.
The behaviour was modified subtly in 1.22 (to make it closer to how Python scalars work). Due to its complexity and the fact that subclassing NumPy scalars is probably a bad idea to begin with. There is probably room for adjustments here.