feat(python): Add [de]serialization tests

[amckinney]

Problem

Users have encountered a variety of runtime issues related to our Pydantic V1 <-> V2 compatibility shim which appear whenever the user [de]serializes their Pydantic models with the SDK.

Solution

We can add simple round-trip [de]serialization tests for each of the generated models using examples (which are auto-generated if not user-specified).

Code

The tests can use a simple utility like the following:

T = TypeVar('T', bound=BaseModel)

def assert_json_serialization(value: T) -> T:
    """
    Assert that a Pydantic model instance can be serialized to JSON and back.
    
    Args:
        instance: The model instance to [de]serialize.
    
    Raises:
        pytest.fail: If serialization or deserialization fails
    """
    model_type = type(value)
    type_name = model_type.__name__

    try:
        serialized = jsonable_encoder(value)
    except Exception as e:
        pytest.fail(f"Failed to serialize {type_name} to JSON: {e}")

    try:
        deserialized = parse_obj_as(model_type, serialized)
    except Exception as e:
        pytest.fail(f"Failed to deserialize {type_name} from JSON: {e}")

    assert value == deserialized, (
        f"JSON round-trip failed for {type_name}.\n"
        f"\tOriginal: {value}\n"
        f"\tDeserialized: {deserialized}"
    )

    # Return the deserialized instance for additional assertions if needed.
    return deserialized

With this, separate test files can be generated like the following:

from seed.types.types import Node, Tree
from tests.utils.assert_json_serialization import assert_json_serialization

def test_node_serialization():
    value = Node(
        name="root",
        nodes=[
            Node(
                name="left",
            ),
            Node(
                name="right",
            ),
        ],
        trees=[
            Tree(
                nodes=[
                    Node(
                        name="left",
                    ),
                    Node(
                        name="right",
                    ),
                ],
            )
        ],
    )
    assert_json_serialization(value)

CI Matrix

With this, we can generate a lightweight GitHub workflow that runs all of the tests using the matrix of Python and Pydantic versions like so:

name: Pydantic Compatibility

on: [push]
jobs:
  compat:
    runs-on: ubuntu-20.04
    strategy:
      fail-fast: false
      matrix:
        python-version: ['3.8', '3.9', '3.10', '3.11', '3.12']
        pydantic-version: ['1.10.20', '2.10.6']
    steps:
      - uses: actions/checkout@v4
      - name: Set up Python ${{ matrix.python-version }}
        uses: actions/setup-python@v4
        with:
          python-version: ${{ matrix.python-version }}
      - name: Bootstrap poetry
        run: |
          curl -sSL https://install.python-poetry.org | python - -y --version 1.5.1
      - name: Install dependencies
        run: |
          poetry install
          poetry add pydantic==${{ matrix.pydantic-version }}
      - name: Run tests
        run: |
          poetry run pytest -rP .