Python incompatibility with pydantic 1.8

[adaamz]

Describe the bug
Looks like datamodel-code-generator >=0.14.0 is not compatible with pydantic 1.8.

To Reproduce

pip install datamodel-code-generator==0.22.0 pydantic==1.8.2
datamodel-codegen --input example.yml --input-file-type=openapi --output example.py --output-model-type="pydantic.BaseModel"

Example schema:

openapi: "3.0.0"
info:
  version: 1.0.0
  title: Example
paths:
  asdf:
    parameters:
      - $ref: "#/components/schemas/asdf"

components:
  schemas:
    asdf:
      type: object
      required:
        - example_id
      properties:
        example_id:
          type: string

Used commandline:

pip install datamodel-code-generator==0.22.0 pydantic==1.8.2
datamodel-codegen --input example.yml --input-file-type=openapi --output example.py --output-model-type="pydantic.BaseModel"

Expected behavior
No error or impossible to install with old version of Pydantic.

Current behavior

Traceback (most recent call last):
  File "/home/adam/.pyenv/versions/3.9.13/bin/datamodel-codegen", line 5, in <module>
    from datamodel_code_generator.__main__ import main
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/site-packages/datamodel_code_generator/__init__.py", line 35, in <module>
    from datamodel_code_generator.parser.base import Parser
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/site-packages/datamodel_code_generator/parser/base.py", line 30, in <module>
    from datamodel_code_generator.model import pydantic as pydantic_model
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/site-packages/datamodel_code_generator/model/__init__.py", line 5, in <module>
    from ..types import DataTypeManager as DataTypeManagerABC
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/site-packages/datamodel_code_generator/types.py", line 54, in <module>
    from datamodel_code_generator.reference import Reference, _BaseModel
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/site-packages/datamodel_code_generator/reference.py", line 50, in <module>
    class _BaseModel(BaseModel):
  File "pydantic/main.py", line 282, in pydantic.main.ModelMetaclass.__new__
  File "pydantic/typing.py", line 287, in pydantic.typing.resolve_annotations
    For use of globalns and localns see the docstring for get_type_hints().
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/typing.py", line 292, in _eval_type
    return t._evaluate(globalns, localns, recursive_guard)
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/typing.py", line 553, in _evaluate
    type_ = _type_check(
  File "/home/adam/.pyenv/versions/3.9.13/lib/python3.9/typing.py", line 158, in _type_check
    raise TypeError(f"{arg} is not valid as type argument")
TypeError: typing.ClassVar[typing.Set[str]] is not valid as type argument

Version:

• OS: linux mint
• Python version: 3.9.13
• datamodel-code-generator version: 0.14.0 - 0.22.0

[koxudaxi]

Thank you for creating an issue.
I think this is not a bug in datamodel-code-generator.
The bug was fixed in pydantic==1.9.0
pydantic/pydantic#3401

We should remove from __future__ import annotations if a model has typing.ClassVar.
Also, we need to change lazy loading types to string.

[koxudaxi]

I got another bug.
https://github.com/koxudaxi/datamodel-code-generator/actions/runs/6427909476/job/17454257906
pydantic/pydantic#4011

It's diffcult to support old pydantic verions :(