Changelog

This page summarizes historic changes in the library. Please also see the release page

0.3

0.3.4

  • check_model does not output unhashable warnings.

  • Added pydantic_sweep.convert to convert between json, yaml and Python code dumps of pydantic Models. This is also an entrypoint that can be executed as python -m pydantic_sweep.convert source target --model 'path.to.Model'

  • initialize now accepts any mixture of str, tuple, and dot-separated keys for default and constant values.

  • Added a model_replace method to create a copy of a pydantic model while overwriting certain values.

0.3.3

  • check_model now warns on non-hashable type hints (can be configured)

  • check_model now explicitly forbids the arbitrary_types_allowed pydantic setting.

0.3.2

  • Fix for Python 3.10

  • Add example folder

  • License documentation and examples under 0BSD

0.3.1

  • initialize accepts nested dictionaries for default/constant arguments.

0.3.0

  • field now checks that values are hashable by default. This can be disabled by setting check=False.

  • Added as_hashable utility to easily compare different configs and pydantic Models.

  • Added check_unique utility to check whether models are unique.

0.2

0.2.1

  • BaseModel gained a custom model_validator that guards against ambiguous model selection from union types in nested models.

  • Relaxed version requirements of pydantic and more-itertools

0.2.0

  • DefaultValue is checked against conflicting settings the same way as normal values

  • Passing DefaultValue to the default/constant argument of initialize works as expected.

0.1

  • Initial release