feat: Initial support for apiv2

[jgadling]

This is a checkpoint: Basic relationships and filters work, but I haven't had a chance to test all of the filter operators and update test infra yet.

[andy-sweet]

Generally looks good, and I'm happy that the codegen module didn't need to be changed much.

After setting up an v2 backend, I was able to run some quick manual tests, so agreed this seems to be kinda working. Lots of the list relationships give empty lists, but that could just be due to the local test infra fakes.

With the new version of the backend, can we specify comments/descriptions on the item/list relationships? That was not possible with hasura, so I just had to autogenerate those descriptions with sometimes awkward results (e.g. The alignment this per section alignment parameters is a part of).

[andy-sweet]

If these are always expected to be identical now, maybe this should just be a tuple of gql/model names instead of a mapping?

[andy-sweet]

Nit: I find this chained call pretty hard to read. Consider unwrapping it with something like.

Suggested change

	of_type = get_named_type(
	get_named_type(field_type.fields["edges"].type).fields["node"].type
	)
	edges_type = get_named_type(field_type.fields["edges"].type)
	of_type = get_named_type(edges_type.fields["node"].type)

[uermel]

This seems to cause an exception in strcase, as self.__name is a list, not a string.

to reproduce:

import cryoet_data_portal as cdp

client = cdp.Client()
point_annos = cdp.AnnotationFile.find(
    client,
    [
        cdp.AnnotationFile.tomogram_voxel_spacing.run_id == 1000,
        cdp.AnnotationFile.annotation_shape.shape_type._in(["Point", "OrientedPoint"]),  # noqa
    ],
)

TypeError: expected string or bytes-like object

[uermel]

Fix seems to be as easy as

return [strcase.to_lower_camel(n) for n in self.__name]