cosapp.utils.helpers

Various small helper functions.

Functions

check_arg(arg, argname, dtype[, value_ok, ...])

Utility function for argument type and value validation.

get_typename(dtype[, multiformat])

is_number(value)

Test if a value is a Number or a 0d numerical array

is_numerical(value)

Test if a value is numerical based on its type.

partition(iterable, predicate)

Partition a collection into two lists, using filter function predicate.
cosapp.utils.helpers.check_arg(arg: Any, argname: str, dtype: Type | Iterable[Type], value_ok: Callable[[Any], bool] = None, stack_shift: int = 0)[source]

Utility function for argument type and value validation. Raises a TypeError exception if type(arg) is not in type list given by ‘dtype’. Raises a ValueError exception if value_ok(arg) is False, where ‘value_ok’ is a boolean function defining a validity criterion.

For example: >>> check_arg(-0.12, ‘my_var’, float) does not raise any exception, as -0.12 is a float

>>> check_arg(-0.12, 'my_var', (int, str))
raises TypeError, as first argument is neither an int, not a str

>>> check_arg(-0.12, 'my_var', float, value_ok = lambda x: x > 0)
raises ValueError, as first argument is not strictly positive
cosapp.utils.helpers.get_typename(dtype: Type | Tuple[Type], multiformat='({})') → str[source]
cosapp.utils.helpers.is_number(value: Any) → bool[source]

Test if a value is a Number or a 0d numerical array

Currently type considered as numerical are:

  • int; but not bool

  • float; including numpy.inf and numpy.nan

  • complex

  • numpy.ndarray of dtype numpy.number and ndim == 0

Parameters:

value (Any) – Value to test

Returns:

Is the value numerical type?

Return type:

bool

cosapp.utils.helpers.is_numerical(value: Any) → bool[source]

Test if a value is numerical based on its type.

Currently type considered as numerical are:

  • int; but not bool

  • float; including numpy.inf and numpy.nan

  • complex

  • numpy.ndarray of dtype numpy.number

  • Collection convertible in a numpy array of dtype derived from numpy.number

Parameters:

value (Any) – Value to test

Returns:

Is the value numerical type?

Return type:

bool

cosapp.utils.helpers.partition(iterable: Iterable[Any], predicate: Callable[[Any], bool])[source]

Partition a collection into two lists, using filter function predicate.

Parameters:

  • iterable: Iterable[Any]

    Iterable collection of elements confronted to predicate.

  • predicate: Callable[[Any], bool]

    Boolean function used to partition the collection.

Returns:

yays, nays: tuple[list, list]

Lists containing the elements for which predicate is True (yays) and False (nays).