perf: cache TypeAdapters

[vvanglro]

Simple test example:

from datetime import date, datetime
from timeit import timeit

from pydantic import TypeAdapter


def parse_obj(annot, value):
    ta = TypeAdapter(annot)

    result1 = timeit(lambda: ta.validate_python(value), number=10000)
    result2 = timeit(lambda: TypeAdapter(annot).validate_python(value), number=10000)

    print(f'{annot.__name__:<8} {result2 / result1:.2f}')


if __name__ == '__main__':
    for t, v in (
        (int, "0"),
        (float, '0.5'),
        (str, b"123"),
        (bool, 'true'),
        (date, '2024-01-01'),
        (datetime, '2024-01-01T00:00:00'),
    ):
        parse_obj(t, v)

output:

int      131.77
float    107.20
str      109.12
bool     128.88
date     283.73
datetime 272.90

Refer to: https://docs.pydantic.dev/latest/concepts/performance/#typeadapter-instantiated-once

[vvanglro]

Sorry I'm late. Ci should be working now.

Reference about lru_cache and pyright:
pydantic/pydantic#5766
microsoft/pyright#5123

[RobertCraigie]

nit: I think this conveys the intent of this code more clearly (and you can add the lru_cache import at the top of the file), thoughts?

Suggested change

	from pydantic import TypeAdapter as PyTypeAdapter
	def TypeAdapter(type_: type[_T]) -> PyTypeAdapter[_T]:
	return PyTypeAdapter(type_)
	if not TYPE_CHECKING:
	TypeAdapter = lru_cache(TypeAdapter)
	if TYPE_CHECKING:
	from pydantic import TypeAdapter
	else:
	from pydantic import TypeAdapter as _TypeAdapter
	TypeAdapter = lru_cache(TypeAdapter)

[vvanglro]

I agree, It has been modified.

[RobertCraigie]

Thanks for this and sorry for the delayed review!