feat: allow invalid output fields to be dropped instead of raising

[haneug]

Closes Issues

Closes None

Description

• Adds an optional keyword strict=True to Output.__init__() and to Output.parse() which allows to drop the fields where Pydantic validation fail when it is set to False. This way we do not switch the current behavior but have the option for a less strict output generation.

---

Release Notes

Added

• Output.parse(), Output.parse_property(), Output.parse_gbw(), and Output.__init__() now accept a strict parameter (default True). When set to False, output fields that fail Pydantic validation are silently set to None and a UserWarning is emitted instead of raising a ValidationError (feat: allow invalid output fields to be dropped instead of raising #248).

[timmyte]

Thanks for adding this feature. That's a very good idea.

It would be great if you could generally document your code more extensively, as it does introduce new concepts, which are not trivial.

What is not clear to me: You are saying, that in non-strict mode invalid keys are set to None? But what if their default value is not None?