Bumped pydantic to v2

[SantanaTiago]

• Bump pydantic to version ^2.5.3 >=2,<3
• Added missing None in Optional fields
• Replace deprecated methods
• Add pydantic.v1 compatibility in pydantic.tools and validateFunction decorator
• Replace validateFunction decorator with v2 validate_call decorator

[RedRoserade]

Thanks for the contribution!

Could you please check that there are no instances of deprecated methods such as the following?:

• .parse_obj() / .parse_raw()
• .dict() / .json()

If run the tests you should get a warnings report at the end, telling you what's deprecated.

[SantanaTiago]

@RedRoserade Many thanks for the inputs. I've indeed run the tests and all pass except one with the following error

ERROR    mognet.worker.worker:worker.py:360 Could not call task function <function add_with_model at 0x7f5f70ebe170> because of a validation error
Traceback (most recent call last):
  File "/project-mognet/mognet/worker/worker.py", line 344, in _run_request
    model = validated.init_model_instance(context, *req.args, **req.kwargs)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/decorator.py", line 130, in init_model_instance
    return self.model(**values)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/main.py", line 341, in __init__
    raise validation_error
pydantic.v1.error_wrappers.ValidationError: 1 validation error for AddWithModel
model
  instance of AddModel expected (type=type_error.arbitrary_type; expected_arbitrary_type=AddModel)

This is related to pydantic v1 ValidatedFunction, a v1 decorator class that's used here for validation (and we using config with arbitrary_types_allowed).
Using pydantic.v1 to keep the current code/logic does not fully work and we have the error, and pydantic v2 decorator don't have a class with same/similar methods in order to keep this logic (as far as I know). Am I wrong and indeed exists a similar v2 version of this which we can apply? Or indeed this logic need to be reworked with new pydantic v2 methods (if so, I appreciate some guidelines)?

[RedRoserade]

@RedRoserade Many thanks for the inputs. I've indeed run the tests and all pass except one with the following error

ERROR    mognet.worker.worker:worker.py:360 Could not call task function <function add_with_model at 0x7f5f70ebe170> because of a validation error
Traceback (most recent call last):
  File "/project-mognet/mognet/worker/worker.py", line 344, in _run_request
    model = validated.init_model_instance(context, *req.args, **req.kwargs)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/decorator.py", line 130, in init_model_instance
    return self.model(**values)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/main.py", line 341, in __init__
    raise validation_error
pydantic.v1.error_wrappers.ValidationError: 1 validation error for AddWithModel
model
  instance of AddModel expected (type=type_error.arbitrary_type; expected_arbitrary_type=AddModel)

This is related to pydantic v1 ValidatedFunction, a v1 decorator class that's used here for validation (and we using config with arbitrary_types_allowed). Using pydantic.v1 to keep the current code/logic does not fully work and we have the error, and pydantic v2 decorator don't have a class with same/similar methods in order to keep this logic (as far as I know). Am I wrong and indeed exists a similar v2 version of this which we can apply? Or indeed this logic need to be reworked with new pydantic v2 methods (if so, I appreciate some guidelines)?

Thanks for having a look, I know that Pydantic v2 is pickier when it comes to these situations. I assume this is the failing test?

Because I find it odd that it considers AddModel to be an "arbitrary type" (it shouldn't, it's a Pydantic model):

instance of AddModel expected (type=type_error.arbitrary_type; expected_arbitrary_type=AddModel)

I have a feeling it's because the decorator is not using model_validate(_json), because I don't think it will automatically convert the dict to an AddModel.

I will try to have a look later today.

[SantanaTiago]

@RedRoserade Many thanks for the inputs. I've indeed run the tests and all pass except one with the following error

ERROR    mognet.worker.worker:worker.py:360 Could not call task function <function add_with_model at 0x7f5f70ebe170> because of a validation error
Traceback (most recent call last):
  File "/project-mognet/mognet/worker/worker.py", line 344, in _run_request
    model = validated.init_model_instance(context, *req.args, **req.kwargs)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/decorator.py", line 130, in init_model_instance
    return self.model(**values)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/main.py", line 341, in __init__
    raise validation_error
pydantic.v1.error_wrappers.ValidationError: 1 validation error for AddWithModel
model
  instance of AddModel expected (type=type_error.arbitrary_type; expected_arbitrary_type=AddModel)

This is related to pydantic v1 ValidatedFunction, a v1 decorator class that's used here for validation (and we using config with arbitrary_types_allowed). Using pydantic.v1 to keep the current code/logic does not fully work and we have the error, and pydantic v2 decorator don't have a class with same/similar methods in order to keep this logic (as far as I know). Am I wrong and indeed exists a similar v2 version of this which we can apply? Or indeed this logic need to be reworked with new pydantic v2 methods (if so, I appreciate some guidelines)?

Thanks for having a look, I know that Pydantic v2 is pickier when it comes to these situations. I assume this is the failing test?

Because I find it odd that it considers AddModel to be an "arbitrary type" (it shouldn't, it's a Pydantic model):

instance of AddModel expected (type=type_error.arbitrary_type; expected_arbitrary_type=AddModel)

I have a feeling it's because the decorator is not using model_validate(_json), because I don't think it will automatically convert the dict to an AddModel.

I will try to have a look later today.

Many thanks for the inputs. I've tried to make it work and tried migrate to v2 decorator validate_call methods but no success. Did you manage to have a look?

[SantanaTiago]

@RedRoserade Many thanks for the inputs. I've indeed run the tests and all pass except one with the following error

ERROR    mognet.worker.worker:worker.py:360 Could not call task function <function add_with_model at 0x7f5f70ebe170> because of a validation error
Traceback (most recent call last):
  File "/project-mognet/mognet/worker/worker.py", line 344, in _run_request
    model = validated.init_model_instance(context, *req.args, **req.kwargs)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/decorator.py", line 130, in init_model_instance
    return self.model(**values)
  File "/project-mognet/venv/lib/python3.10/site-packages/pydantic/v1/main.py", line 341, in __init__
    raise validation_error
pydantic.v1.error_wrappers.ValidationError: 1 validation error for AddWithModel
model
  instance of AddModel expected (type=type_error.arbitrary_type; expected_arbitrary_type=AddModel)

This is related to pydantic v1 ValidatedFunction, a v1 decorator class that's used here for validation (and we using config with arbitrary_types_allowed). Using pydantic.v1 to keep the current code/logic does not fully work and we have the error, and pydantic v2 decorator don't have a class with same/similar methods in order to keep this logic (as far as I know). Am I wrong and indeed exists a similar v2 version of this which we can apply? Or indeed this logic need to be reworked with new pydantic v2 methods (if so, I appreciate some guidelines)?

Thanks for having a look, I know that Pydantic v2 is pickier when it comes to these situations. I assume this is the failing test?
Because I find it odd that it considers AddModel to be an "arbitrary type" (it shouldn't, it's a Pydantic model):

instance of AddModel expected (type=type_error.arbitrary_type; expected_arbitrary_type=AddModel)

I have a feeling it's because the decorator is not using model_validate(_json), because I don't think it will automatically convert the dict to an AddModel.
I will try to have a look later today.

Many thanks for the inputs. I've tried to make it work and tried migrate to v2 decorator validate_call methods but no success. Did you manage to have a look?

We manage to make it work with v2 decorator validate_call

[RedRoserade]

Approved! Thanks for taking care of this.

Please make sure to squash the commits before merging.

[RedRoserade]

Also, please make sure to do a major version bump, because the changes here are backwards-incompatible (in terms of dependencies).