Remove apparently-unnecessary model_rebuild

[dmontagu]

Partially addresses the performance issues noted in #5351

[dmontagu]

I'll note here that this doesn't have as dramatic of an effect on overall pytest runtime as it does on hot-loop generic model creation.

I ran pytest tests/test_generics.py --durations 20 on this branch and on the pre-refactor commit af73124e, and it seems to attribute most of the difference in runtime to the "setup" of the tests:

# Pre-refactor:
0.46s call     tests/test_generics.py::test_generics_work_with_many_parametrized_base_models
0.10s call     tests/test_generics.py::test_caches_get_cleaned_up
0.02s setup    tests/test_generics.py::test_generic_config
0.02s setup    tests/test_generics.py::test_classvar
0.01s setup    tests/test_generics.py::test_enum_generic
0.01s call     tests/test_generics.py::test_caches_get_cleaned_up_with_aliased_parametrized_bases
0.01s call     tests/test_generics.py::test_generic_name
0.01s call     tests/test_generics.py::test_generic_recursive_models_complicated
0.01s call     tests/test_generics.py::test_generic_recursive_models_triple
0.01s setup    tests/test_generics.py::test_generic_recursive_models_with_a_concrete_parameter
0.01s setup    tests/test_generics.py::test_generic_recursive_models_in_container
0.01s setup    tests/test_generics.py::test_generic_with_user_defined_generic_field
0.01s setup    tests/test_generics.py::test_construct_other_generic_model_with_validation
0.01s setup    tests/test_generics.py::test_multilevel_generic_binding
0.01s setup    tests/test_generics.py::test_generic_subclass
0.01s setup    tests/test_generics.py::test_generic_model_as_parameter_to_generic_type_alias
0.01s setup    tests/test_generics.py::test_generic_recursive_models_complicated
0.01s setup    tests/test_generics.py::test_generic_recursive_models_triple
0.01s setup    tests/test_generics.py::test_multi_inheritance_generic_defaults
0.01s setup    tests/test_generics.py::test_double_typevar_substitution

# This branch:
0.56s call     tests/test_generics.py::test_generics_work_with_many_parametrized_base_models
0.09s call     tests/test_generics.py::test_caches_get_cleaned_up
0.03s setup    tests/test_generics.py::test_generic_enum
0.03s setup    tests/test_generics.py::test_generic_annotated
0.03s setup    tests/test_generics.py::test_parent_field_parametrization
0.03s setup    tests/test_generics.py::test_nested
0.03s setup    tests/test_generics.py::test_optional_value
0.03s setup    tests/test_generics.py::test_construct_generic_model_with_validation
0.03s setup    tests/test_generics.py::test_multilevel_generic_binding
0.03s setup    tests/test_generics.py::test_generic_recursive_models
0.03s setup    tests/test_generics.py::test_generic_recursion_contextvar
0.03s setup    tests/test_generics.py::test_generic_enums
0.03s setup    tests/test_generics.py::test_partial_specification_instantiation
0.03s setup    tests/test_generics.py::test_generic_subclass_with_extra_type
0.03s setup    tests/test_generics.py::test_custom_generic_naming
0.03s setup    tests/test_generics.py::test_schema_is_valid
0.03s setup    tests/test_generics.py::test_construct_other_generic_model_with_validation
0.03s setup    tests/test_generics.py::test_generic_recursive_models_triple
0.03s setup    tests/test_generics.py::test_generic_recursive_models_in_container
0.03s setup    tests/test_generics.py::test_double_typevar_substitution

I also checked and there doesn't seem to be any meaningful changes to fixtures in test_generics.py or conftest.py, so I'm not sure where the extra setup time is coming from, or how to profile it.

I also checked, adding time.sleep(0.02) as a fixture brings the runtimes in line, so it at least seems consistent with something slow happening in a fixture or similar, but it would still be nice to confirm/resolve that.

[adriangb]

Some profiling for the record

from typing import Generic, List, Type, TypeVar

from pydantic import BaseModel

T = TypeVar('T')
C = TypeVar('C')


class A(BaseModel, Generic[T, C]):
    x: T
    y: C


class B(A[int, C], BaseModel, Generic[C]):
    pass


models: List[Type[BaseModel]] = []
for i in range(1_000):
    models.append(type(f'M{i}', (BaseModel,), {}))

for m in models:
    B[m]

A pyinstrument report:
report.html.zip

JSON data for https://www.speedscope.app/:
CPU_profile_for_test.speedscope.json.zip

[dmontagu]

It turns out the garbage collection fixture was the one that had increased in runtime from ~10ms to ~30ms. Perhaps worth looking into more, but removing from most tests, the runtime becomes much more similar pre- and post-refactor.

[dmontagu]

@samuelcolvin I think unless this change seems suspicious to you we should merge this, and if the garbage collection stuff concerns you at all let's just create a separate issue for that.

Given the massive simplifications I realized were possible in #5361 I just think that the improvements to model creation eliminated the need for some of the hacks that I had to use to get it working in the first place.

[samuelcolvin]

Amazing, thank you.