v1.0.0

In v1.0.0 we god rid of both the attrs and marshmallow dependency and use only pydantic models which makes constructing and serializing models way more consistent. Also this allows for easy integration with FastAPI backends.

The historic and hard to explain-dependency injection setup remains for now but still this is a great step forward.
The updated tests in each of the breaking change-commits are a good indicator on how you need to adapt model dumps and loads in v1.

What's Changed

• Switch CategorizedKeyExtract from attrs+marshmallow to pydantic (breaking) by @hf-kklein in #640
• Switch ContentEvaluationResult from attrs+marshmallow to pydantic (breaking) by @hf-kklein in #677
• chore: fix deprecation warning about datetime.utcnow() by @hf-kklein in #678
• Switch models from mapping_results module from attrs+marshmallow to pydantic (breaking) by @hf-kklein in #679
• chore: fix linter warning in json schema generation by @hf-kklein in #681
• chore(tests): fix wrong AsyncMock setup by @hf-kklein in #682
• switch module evaluation_results from attrs+marshmallow to pydantic (breaking) by @hf-kklein in #680
• chore(deps): drop attrs dependency by @hf-kklein in #683
• chore(deps): drop unmaintained marshmallow_enum dependency by @hf-kklein in #684
• chore(deps): bump lark from 1.3.0 to 1.3.1 by @dependabot[bot] in #686
• Drop (Concise)TreeSchemas, use pydantic TypeAdapter instead; Remove marshmallow dependency; Drop Python 3.9 support (breaking) by @hf-kklein in #685
• chore: Update Read the Docs-Requirements by @hf-kklein in #687
• fix(docs): update minimal working example jupyter notebook; fix build errors there by @hf-kklein in #688

Full Changelog: v0.16.4...v1.0.0