Provide a mode parameter in validate_python and model_validate

[DanielRosenwasser]

Initial Checks

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

Description

Given the following:

from dataclasses import dataclass
import pydantic
import json

@dataclass
class Blah:
  a: str

adapted_model = pydantic.TypeAdapter(Blah) 

The following two calls will differ in behavior:

# works! ✅
adapted.model.validate_json('{"a": "hello!"}', strict=True)

# fails! ❌
adapted_model.validate_python(json.loads('{"a": "hello!"}'), strict=True)

This is because validate_python really expects an instance of a Blah instead of a dict, whereas validate_json loosens the rules here.

This is a problem if the data has already been parsed from JSON, or comes from a source other than JSON (e.g. YAML).

A suggestion for exposing something like this already exists over at pydantic-core (pydantic/pydantic-core#712).

This issue is really more meant to track exposing the same logic in the appropriate methods Pydantic ships.

Affected Components

• Compatibility between releases
• 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
• 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]

@DanielRosenwasser,

Thanks for the well thought out feature request. Do you feel as though the feature request you have here is different enough from #8086 that we should leave both issues open?

If so, happy to leave this open. We could add this to the v2.8.0 milestone 👍

[DanielRosenwasser]

My hunch is that since that one is more concerned with serialization, and this one is concerned with validation (plus considering some of the involved issues discussed at pydantic/pydantic-core#712 and pydantic/pydantic-core#1226), it might be better to keep both of these open.

I'll defer to your expertise and familiarity here - I'm sure the core maintainers are way more familiar with what needs to be done to close out both issues at the same time.

[sydney-runkle]

Let's leave this open for now!