Pyfields

**default value factories**

- `default_factory` callables now receive one argument: the object instance. Fixes [6](https://github.com/smarie/python-pyfields/issues/6)
- New decorator `<field>.default_factory` to define a default value factory. Fixed [27](https://github.com/smarie/python-pyfields/issues/27)
- New `copy_value`, `copy_field` and `copy_attr` helper functions to create default value factories. Fixed [26](https://github.com/smarie/python-pyfields/issues/26)

**support for slots**

- `field` now automatically detects when a native field is attached to a class with slots and no `__dict__` is present. In that case, the native field is replaced with a descriptor field. Fixed [20](https://github.com/smarie/python-pyfields/issues/20).

See [documentation page](https://smarie.github.io/python-pyfields/changelog/) for details.

**fields**

- `field()` method to easily define class fields without necessarily defining a `__init__`.

- "native" fields are created by default, or if `native=True` is set. A `NativeField` is a non-data descriptor that replaces itself automatically with a native python attribute after the first read, to get the same performance level on later access.

- "descriptor" fields are created when type or value validation is required, or if `native=False` is set. A `DescriptorField` uses the standard python descriptor protocol so that type and value can be validated on all future access without messing with the `__setattr__` method.

- support for `type_hint` declaration to declare the type of a field. If `validate_type` provided, the descriptor will *not* be replaced with a native field, and the type will be checked on every value modification. A `TypeError` will be raised if type does not comply. Type hints are correctly defined so that IDEs can pick them. Fixes [10](https://github.com/smarie/python-pyfields/issues/10)

- support for `validators` relying on `valid8`. Validators can receive `(val)`, `(obj, val)` or `(obj, field, val)` to support validation based on several fields. The only requirement is to return `True` or `None` in case of success. Fixes [3](https://github.com/smarie/python-pyfields/issues/3)

**init**

- `make_init` method to create an entire `__init__` method with control of which fields are injected, and with possibility to blend a post-init callback in. Fixes [14](https://github.com/smarie/python-pyfields/issues/14).

- `init_fields` decorator to auto-init fields before your `__init__` method.

- `inject_fields` decorator to easily inject `fields` in an init method and perform the assignment precisely when users want (for easy debugging). Fixes [13](https://github.com/smarie/python-pyfields/issues/13)

**misc**

- `__weakref__` added in all relevant classes. Fixes [21](https://github.com/smarie/python-pyfields/issues/21)

- Now using stubs [17](https://github.com/smarie/python-pyfields/issues/17)

- Fixed bug [11](https://github.com/smarie/python-pyfields/issues/11).

- Fixed `ValueError` with mini-lambda < 2.2. Fixed [22](https://github.com/smarie/python-pyfields/issues/22)

- Because of a [limitation in PyCharm type hints](https://youtrack.jetbrains.com/issue/PY-38151) we had to remove support for class-level field access. This created [#12](https://github.com/smarie/python-pyfields/issues/12) which will be fixed as soon as PyCharm issue is fixed.


See [documentation page](https://smarie.github.io/python-pyfields/changelog/) for details.