Maintain pydantic.v1 model compatibility while using pydantic v2 #393
Conversation
voidnologo
changed the title
from pydantic.v1 import .... With the new changes to support pydantic V2 we are wanting to upgrade Spectree to latest and begin writing new serializers, however, we are not ready to update all of our current models to pydantic be pydantic v2 BaseModels.
`from pydantic.v1 import ...`. With the new changes to support pydantic V2 we are wanting to upgrade Spectree to latest and begin writing new serializers, however, we are not ready to update all of our current models to pydantic be pydantic v2 BaseModels. Currently if a ValidationError is raised, the falcon plug in does not catch it, since ValidationError in the plugin is coming from pydantic, not pydantic.v1. Since you are aliasing the pydantic.v1.ValidationError as InternalValidationError, catching it as well allows us to continue to support our v1 models while we migrate to v2 models. Looking at the codebase, this looks like a similar issue in the other plugins. Let me know if you would like me to submit this update for them as well. I'm not sure where you would like a test for this in the project.
voidnologo
force-pushed
the
pydanticV1_compatability
branch
from
9eafcd8 to
c741462
Compare
The test will install the current code: So it should be the latest commit in this PR. One thing to note is that you may need to run |
|
Thanks. That helps. Still trying to figure out the test. Patching the code this way locally in our project fixes the issues; just need to figure out how to write a test here. Thanks for being so responsive. |
|
|
||
| @pytest.mark.skipif(not PYDANTIC2, reason="only matters if using both model types") | ||
| def test_falcon_validate_both_v1_and_v2_validation_errors(client): | ||
| class CompatibilityView: |
There was a problem hiding this comment.
Inlined this view because adding it to the test app then causes issue with the generated spec tests.
We have pydantic V2 installed in our environemnt, but have been using
from pydantic.v1 import .... With the new changes to support pydantic V2 we are wanting to upgrade Spectree to latest and begin writing new serializers, however, we are not ready to update all of our current models to be pydantic v2 BaseModels.Currently if a
ValidationErroris raised, the falcon plug in does not catch it, sinceValidationErrorin the plugin is coming frompydantic, notpydantic.v1.Since you are aliasing the
pydantic.v1.ValidationError as InternalValidationError, catching it as well allows us to continue to support our v1 models while we migrate to v2 models.Looking at the codebase, this looks like a similar issue in the other plugins. Let me know if you would like me to submit this update for them as well. I'm not sure where you would like a test for this in the project.