Model signatures contain reserved words.

[strue36]

Checks

• I added a descriptive title to this issue
• I have searched (google, github) for similar issues and couldn't find anything
• I have read and followed the docs and still think this is a bug

Bug

When creating a model with a field that has an alias that is a python reserved word, the signature uses the reserved word even when allow_population_by_field_name is true.

I expect the signature to use the field name in this case. I came across this issue when generating tests using hypothesis. This uses the model signature and was creating invalid test code as it was using the reserved words in the signature.

pydantic version: 1.9.0
pydantic compiled: False
python version: 3.9.0 (tags/v3.9.0:9cf6752, Oct  5 2020, 15:34:40) [MSC v.1927 64 bit (AMD64)]
platform: Windows-10-10.0.19041-SP0
optional deps. installed: ['devtools', 'dotenv', 'email-validator', 'typing-extensions']

from pydantic import BaseModel, Field
from inspect import signature

class Foo(BaseModel):
    from_: str = Field(..., alias='from')

    class Config:
        allow_population_by_field_name = True

str(signature(Foo))

'(*, from: str) -> None'

Suggestion

In utils.generate_model_signature() we have these two checks.

elif not param_name.isidentifier():
    if allow_names and field_name.isidentifier():

I propose replacing these with calls to the following (new) function.

def is_valid_identifier(identifier: str) -> bool:
    """
    Checks that a string is a valid identifier and not a reserved word.
    :param identifier: The identifier to test.
    :return: True if the identifier is valid.
    """
    return identifier.isidentifier() and not keyword.iskeyword(identifier)

I believe that this behaviour is closer to what a user would expect in the common case that they are generating models from a schema containing python reserved words.