`elasticai.creator.function_utils`#

Module Contents#

Classes#

`FunctionDecorator`	Apply a callback to functions and either their own or custom names.
`RegisterDescriptor`	Automatically connect the `FunctionDecorator` to a callback and make it look like a method.
`KeyedFunctionDispatcher`	Functions can be registered with custom key or by using the function’s name as key. Upon `call` the registry will use the function specified by the `dispatch_key_fn` and the argument passed to `call` to process the argument and return the result.

Data#

`Tin`
`Tout`
`FN`

API#

elasticai.creator.function_utils.Tin#: ‘TypeVar(…)’

elasticai.creator.function_utils.Tout#: ‘TypeVar(…)’

elasticai.creator.function_utils.FN#: ‘TypeVar(…)’

class elasticai.creator.function_utils.FunctionDecorator(callback: Callable[[str, elasticai.creator.function_utils.FN], elasticai.creator.function_utils.Tout])[source]#

Bases: typing.Generic[elasticai.creator.function_utils.FN, elasticai.creator.function_utils.Tout]

Apply a callback to functions and either their own or custom names.

Param:: callback: will be called as callback(name, fn)

Important

if you want to use this as a decorator, do not forget to return the wrapped function from your callback.

Examples:

>>> registry = dict()
>>> def register_fn(name, fn):
...  registry[name] = fn
...  return fn
...
>>> register = FunctionDecorator(register_fn)
>>> def my_fn(x):
...   print(x)
...
>>> register(my_fn)
>>> registry["my_fn"]("hello, world!")
"hello, world!"

another example could look like this:

registry = dict()

def register_fn(name, fn):
    registry[name] = fn
    return fn

register = FunctionDecorator(register_fn)

@register("other_name")
@register
def my_fn(x):
    print(x)

This will add my_fn to the registry using the name "my_fn" and the name "other_name".

Initialization

__call__(arg: elasticai.creator.function_utils.FN | str, arg2: elasticai.creator.function_utils.FN | None = None, /) → elasticai.creator.function_utils.Tout | Callable[[elasticai.creator.function_utils.FN], elasticai.creator.function_utils.Tout][source]#

class elasticai.creator.function_utils.RegisterDescriptor[source]#

Bases: typing.Generic[elasticai.creator.function_utils.Tin, elasticai.creator.function_utils.Tout]

Automatically connect the FunctionDecorator to a callback and make it look like a method.

The owning instance needs to define a callback that has the name f"_{name}_callback", where name is the name of the field assigned to this descriptor.

For an example see the KeyedFunctionDispatcher below.

__set_name__(instance, name)[source]#

__get__(instance, owner=None) → elasticai.creator.function_utils.FunctionDecorator[Callable[[elasticai.creator.function_utils.Tin], elasticai.creator.function_utils.Tout], Callable[[elasticai.creator.function_utils.Tin], elasticai.creator.function_utils.Tout]][source]#

class elasticai.creator.function_utils.KeyedFunctionDispatcher(dispatch_key_fn: Callable[[elasticai.creator.function_utils.Tin], str])[source]#