Several tests fail with Pydantic 1.10.15 (latest 1.x)

[musicinmybrain]

Description

Pydantic 1.10.15 aliases the pydantic namespace as pydantic.v1 as a forward-compatibility feature. This seems to interact badly with several

Steps to reproduce

$ gh repo clone aws/serverless-application-model
$ cd serverless-application-model
$ python3 -m venv _e
$ . _e/bin/activate
(_e) $ pip install -e .[dev]
(_e) $ pip install pydantic==1.10.15
(_e) $ make test

Observed result

=========================================================================================== short test summary info ============================================================================================
FAILED tests/model/test_resource_validator.py::TestResourceValidatorFailure::test_connector_with_empty_properties - AttributeError: module 'pydantic.v1' has no attribute 'error_wrappers'
FAILED tests/model/test_resource_validator.py::TestResourceValidatorFailure::test_connector_with_invalid_permission - AttributeError: module 'pydantic.v1' has no attribute 'error_wrappers'
FAILED tests/model/test_resource_validator.py::TestResourceValidatorFailure::test_connector_with_invalid_permission_type - AttributeError: module 'pydantic.v1' has no attribute 'error_wrappers'
FAILED tests/model/test_resource_validator.py::TestResourceValidatorFailure::test_connector_without_source - AttributeError: module 'pydantic.v1' has no attribute 'error_wrappers'
FAILED tests/translator/test_translator.py::test_transform_invalid_document[error_graphqlapi] - AttributeError: module 'pydantic.v1' has no attribute 'error_wrappers'
============================================================================= 5 failed, 3967 passed, 5 rerun in 104.48s (0:01:44) =============================================================================

Expected result

(_e) $ pip install pydantic==1.10.14
(_e) $ make test

======================================================================================= 3972 passed in 103.63s (0:01:43) =======================================================================================

Additional environment details

1. OS: Fedora 39 x86_64
2. If using the SAM CLI, sam --version: N/A
3. AWS region: N/A

[musicinmybrain]

In Pydantic 1.10.14, from pydantic import v1 as pydantic raises an ImportError in

serverless-application-model/samtranslator/compat.py

Lines 1 to 6 in 44562dc

	try:
	from pydantic import v1 as pydantic
	except ImportError:
	# Unfortunately mypy cannot handle this try/expect pattern, and "type: ignore"
	# is the simplest work-around. See: https://github.com/python/mypy/issues/1153
	import pydantic # type: ignore

and import pydantic is used.

In Pydantic 1.10.14, from pydantic import v1 as pydantic succeeds, but pydantic.v1.error_wrappers is not present, only pydantic.error_wrappers; therefore, samtranslator.compat.pydantic.error_wrappers is not present.

Related: https://github.com/pydantic/pydantic/blob/5476a758c8ac59887dbfa3aa1c3481d0a0e20837/pydantic/v1.py#L5-L6

[musicinmybrain]

I’ll try to offer a PR for this later today.

[musicinmybrain]

I’ll try to offer a PR for this later today.

I am actually not sure what the correct approach is here. Using pydantic.ValidationError in the except …: isn’t right (on Pydantic v2, it is pydantic_core._pydantic_core.ValidationError, which has no direct inheritance relationship with pydantic.v1.error_wrappers.ValidationError).

It seems like it might be a bug that pydantic.v1.error_wrappers isn’t present in Pydantic 1.10.15, and that this might be the same as pydantic/pydantic#9357, but I don’t have enough Pydantic experience to state that confidently.

[shikha372]

Was able to reproduce this issue on local

Issue opened with pydantic team for import error
pydantic/pydantic#9625

[shikha372]

Thanks @musicinmybrain for highlighting this issue. We've opened a bug at pydantic end to check this, meanwhile added this to list of expected exception at SAM end in PR#3611 . Let us know if you still have any questions regarding this.

[github-actions]

This issue is now closed. Comments on closed issues are hard for our team to see.
If you need more assistance, please either tag a team member or open a new issue that references this one.