Basic type conversion: convert & friends

The core of pane is its data conversion routines. The simplest to use is convert:

>>> import typing as t
>>> import pane
>>> pane.convert(-5., float)  # simple scalar type
-5.0
>>> pane.convert([1., 'mystring', 5], (float, str, int))  # tuple-like type
(1.0, 'mystring', 5)
>>> pane.convert({'x': 5.0, 'y': 'mystring'}, {'x': float, 'y': str})  # struct-like type
{'x': 5.0, 'y': 'mystring'}
>>> pane.convert(5, t.Union[str, int, None])  # untagged union
5

convert actually performs two separate actions. First, it calls into_data on a value. This attempts to convert the value into a "data interchange type", which is a dialect of types supported by all converters.

After into_data, convert calls from_data, which attempts to convert the value into the desired type.

Currently, the data interchange types are: - Scalar types: str, bytes, bool, int, float, complex, and None - Sequences: list, tuple and t.Sequence - Mappings: dict and t.Mapping

All implementations of Converter.into_data must output a data interchange type, and all implementations of Converter must handle any data interchange type (even if 'handle' just means returning an error). Data interchange types may be added in major releases, so Converter implementations should be made robust to new types.