No module named 'aws_resource_validator.pydantic_models'

[jschnare]

This looks like an interesting project. Unfortunately I can't try the pydantic model examples because they aren't included in the package.

$ uv pip show aws_resource_validator -f
Name: aws-resource-validator
Version: 1.0.2
Location: /home/jschnare/.venv/lib/python3.12/site-packages
Requires: botocore, exrex, pydantic
Required-by:
Files:
  aws_resource_validator-1.0.2.dist-info/INSTALLER
  aws_resource_validator-1.0.2.dist-info/LICENSE
  aws_resource_validator-1.0.2.dist-info/METADATA
  aws_resource_validator-1.0.2.dist-info/RECORD
  aws_resource_validator-1.0.2.dist-info/REQUESTED
  aws_resource_validator-1.0.2.dist-info/WHEEL
  aws_resource_validator/class_definitions.py
  aws_resource_validator/models.py

It looks like pyproject.toml is explicitly filtering out anything else - https://github.com/CoreOxide/aws_resource_validator/blob/1.0.2/pyproject.toml#L12-L15

[dsully]

I just came here to say the same thing..

[CoreOxide]

Hey guys,
So the project was split into individual packages, where the convention is boto3-pydantic2-<service_name>.
i.e. : boto3-pydantic-wafv2

Do you think it's confusing?

P.S.
I now see that I did not commit the new docs. Sorry.

[dsully]

Yes, that's pretty confusing, especially without the docs. 🙃

[dsully]

The existing modules don't work either.

from boto3_pydantic_s3.s3_classes import GetObjectOutput

   from boto3_pydantic_s3.s3_classes import GetObjectOutput
  File "/Users/dsully/dev/work/.venv/lib/python3.13/site-packages/boto3_pydantic_s3/s3_classes.py", line 8, in <module>
    from aws_resource_validator.pydantic_models.s3.s3_constants import *
ModuleNotFoundError: No module named 'aws_resource_validator.pydantic_models'

s3_classes.py has:

from typing import Any, Dict, IO, List, Literal, Mapping, Optional, Sequence, Union, Callable
from datetime import datetime

from boto3.s3.transfer import TransferConfig
from botocore.client import BaseClient
from pydantic import BaseModel, Field
from botocore.response import StreamingBody
from aws_resource_validator.pydantic_models.s3.s3_constants import *
from ..base_validator_model import BaseValidatorModel, EventStream

The last two lines don't exist as modules. The last line has an extra ".", and the 2nd to last line should also be just "from .s3_constants import *`

[dsully]

If I hand patch that, then there is a typing error:

aws_resource_validator/pydantic_models/s3/s3_classes.py", line 148, in ClientGeneratePresignedPostRequest
    Conditions: Optional[Union[Any], Dict[str, List[Any]]] = None
                ~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/dsully/.local/share/uv/python/cpython-3.13.3-macos-aarch64-none/lib/python3.13/typing.py", line 432, in inner
    return func(*args, **kwds)
  File "/Users/dsully/.local/share/uv/python/cpython-3.13.3-macos-aarch64-none/lib/python3.13/typing.py", line 576, in __getitem__
    return self._getitem(self, parameters)
           ~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^
  File "/Users/dsully/.local/share/uv/python/cpython-3.13.3-macos-aarch64-none/lib/python3.13/typing.py", line 809, in Optional
    arg = _type_check(parameters, f"{self} requires a single type.")
  File "/Users/dsully/.local/share/uv/python/cpython-3.13.3-macos-aarch64-none/lib/python3.13/typing.py", line 202, in _type_check
    raise TypeError(f"{msg} Got {arg!r:.100}.")
TypeError: typing.Optional requires a single type. Got (typing.Any, typing.Dict[str, typing.List[typing.Any]]).

[CoreOxide]

I see.
It worked on my .venv because I already had all of the generated classes, but when installing just a single package into a clean .venv, it indeed has missing import.

Do you think all of the classes should be packed together?
It's just that most people need 1-2, maybe 10 classes max, and there are tens of services that are pretty exotic that I dobt anyone uses...

[dsully]

I'm good with separate packages. The generated code should be smoke tested for imports though. :)

[CoreOxide]

I'll fix the packaging logic. Thanks for opening a ticket!

[vvchistyakov]

Hey @CoreOxide,
Just in case, do you have plans for fixing it? =)

[CoreOxide]

@vvchistyakov plan? sure!
have time? not really :(
I'll try to get to it next weekend.

[CoreOxide]

Hi @jschnare, @dsully, @vvchistyakov,

We have release v2.0 with numerous fixes and improvement! Sorry for the huge delay.
You can now install individual packages like so:

pip install 'aws-resource-validator[s3]' 
pip install 'aws-resource-validator[lambda]' 

And you will get base requirements + only the classes you requested.

Imports are tested by on-the-fly generated tests:

from aws_resource_validator.pydantic_models.lambda_.lambda__classes import DeleteLayerVersionRequestTypeDef
delete_layer: DeleteLayerVersionRequestTypeDef = DeleteLayerVersionRequestTypeDef(LayerName='string', VersionNumber=123)

Please refer to the updated README.md for more information, or let me know if you have any questions or comments.