Better IDE/autocomplete support

[knkski]

@johnsca brought this up. It would be nice if we could figure out same way of adding better support for tooling such as autocomplete to this library. get_interfaces() returns a dictionary that doesn't allow for any real introspection of what's in it.

The difficulty is that we're tied to agreeing on a schema version at runtime, which cuts our options down by a lot, possibly down to no real options. Using regular Python classes like charmcraft libs doesn't help, as we could end up in a situation like this:

from dataclasses import dataclass

@dataclass
class IngressV1Schema:
    foo: str

@dataclass
class IngressV2Schema:
    bar: str

If get_interfaces() has to decide at runtime which class will be used, static tooling that IDEs use won't be much help anyways. In fact they'll probably pick one or the other and provide wrong autocomplete options.

[sed-i]

I wonder if it would be possible to always delegate to handlers per version to help static analyzers, for example:

def load_rel_data(...) -> Union[IngressV1Schema, IngressV2Schema]:
    pass

# ...

def handle_v1(rel_data: IngressV1Schema):
    # use v1 fields
    pass

def handle_v2(rel_data: IngressV2Schema):
    # use v2 fields
    pass

Or context dependent factory methods

value = RelDataParser.v1(rel_data).alerts.name 
        if version == "v1" 
        else RelDataParser.v2(rel_data).group[0].alerts.name