Release 0.7.3

• Fix issues with newer version of pydantic and pyright (typing failed) by @ddanier
• Add support for partially hydrated models, thanks to @mathieu-gillot (see #36)

Release 0.7.0

• More types are not checked for equality and will not trigger model_has_changed to be true
  • datetime/date/time/timedelta
  • list/set/tuple and also dict
  • pydantic.BaseModel instances, as pydantic supports comparing those
• You may not override checking for equality more easily using the two methods _model_value_is_comparable_type and _model_value_is_actually_unchanged. If any type is missing for you, feel free to extend the class 😉

Release 0.6.7

• Fix issue with pylance, now "exporting" the library contents correctly
• Officially supporting Python 3.12 (tox tests on Python 3.12, too)

Release 0.6.6

• Add support for new pydantic 2.7 parameters (context and serialize_as_any) to model_dump and model_dump_json

Release 0.6.5

• Fix issue #27

Release 0.6.3

• Fix typing issue with __deepcopy__

Release 0.6.2

• Fix #26: Default values did not initialise changedetection state

Release 0.6.0

With version 0.6.0 we added three new features:

• Changes to scalar values will now only be seen as a change if the new value is actually different. See #17 for details. Thanks for @dannosaur for the report the old behaviour does not match user expectation.
• There are now utility methods to restore original field values or a copy of the original model state. See #11 for details.
• You may now use change markers to mark objects as changed without a field actually being changes. This can be useful when for example propagating changes state from child objects to its parents. See #10 for details.
• The README was updated to reflect those new features. Also some example code was added to chow how set_changed(...) works.

Release 0.5.1

Update pydantic version to also support 2.1 (and further 2.x releases).

Release 0.5.0

Add support for pydantic 2.0 while keeping support for 1.9 and 1.10. 🥳