Customize Pydantic Error Type (AKA subclassing PydanticCustomError)

[TiansuYu]

Initial Checks

• [X] I have searched Google & GitHub for similar requests and couldn't find anything
• [X] I have read and followed the docs and still think this feature is missing

Description

I would like to subclass the pydantic_core.PydanticCustomError such that each time a specific field_validator fails, it can raise a MySpecificError instead of the normal pydantic ones.

Currently, PydanticCustomerError is marked as @final and as such, subclassing will lead to TypeError: type 'pydantic_core._pydantic_core.PydanticCustomError' is not an acceptable base type.

Futhermore, I have specific needs to do downstream error handling based on MyTypeAError(PydanticCustomError) and MyTypeBError(PydanticCustomError). Currently, this is not possible.

I know workarounds would be wrap around the model creation step Model(field_a=xxx, field_b=xxx) with a catch block, but still, I do not want to involve my lib user to accept this complicated user experience.

Please let me know:

• If this is a feature could be added into Pydantic;
• If there are any workarounds for this (e.g. add some additional logic into BaseModel.__init__).

Affected Components

• [ ] Compatibility between releases
• [X] Data validation/parsing
• [ ] Data serialization - .model_dump() and .model_dump_json()
• [ ] JSON Schema
• [ ] Dataclasses
• [ ] Model Config
• [ ] Field Types - adding or changing a particular data type
• [ ] Function validation decorator
• [ ] Generic Models
• [X] Other Model behaviour - model_construct(), pickling, private attributes, ORM mode
• [ ] Plugins and integration with other tools - mypy, FastAPI, python-devtools, Hypothesis, VS Code, PyCharm, etc.

[sydney-runkle]

@TiansuYu,

This seems reasonable - let's move this to pydantic-core - hopefully we could just remove the @final decorator.

[sydney-runkle]

Related to https://github.com/pydantic/pydantic/issues/9686, can probably be done at the same time.