Bug: potential naming conflict in generated protocols

[Karrenbelt]

If a performative in a protocol specification has the same name as custom type, then the tests/performatives.py, which contains helper models for hypothesis testing, will overwrite the existing definition within that scope. Example snippet:

from packages.zarathustra.protocols.asset_bridging.custom_types import (
    BridgeResult,
)


class BridgeResult(BaseModel):
    """Model for the `BRIDGE_RESULT` initial speech act performative."""

    result: BridgeResult


class RequestStatus(BaseModel):
    """Model for the `REQUEST_STATUS` initial speech act performative."""

    result: BridgeResult 

[Karrenbelt]

while we're at it, usage of model.__fields__ might be replaced with model_fields to reduce the overload of warnings produced.

The `__fields__` attribute is deprecated, use `model_fields` instead. Deprecated in Pydantic V2.0 to be removed in V3.0. See Pydantic V2 Migration Guide at https://errors.pydantic.dev/2.11/migration/

used here

auto_dev/auto_dev/data/templates/protocols/test_dialogues.jinja

Lines 24 to 27 in c2abb54

	def shallow_dump(model: BaseModel) -> dict:
	"""Shallow dump pydantic model."""
	return {name: getattr(model, name) for name in model.__fields__}

auto_dev/auto_dev/data/templates/protocols/test_messages.jinja

Lines 28 to 31 in c2abb54

	def shallow_dump(model: BaseModel) -> dict:
	"""Shallow dump pydantic model."""
	return {name: getattr(model, name) for name in model.__fields__}