✨ Support field max_length

[proever]

Currently, when max_length is specified in a Field, it has no real effect on the corresponding SQL schema (#126, #746). As pointed out by @chris-beedie, this is because the configured max_length is stored as an annotated_types.MaxLen, rather than a PydanticMetadata object, and is therefore ignored in get_field_metadata.

This PR adds a simple extra check for that specific case. I have confirmed it works using PostgreSQL 16.2.

I also wanted to write a test for it, along the following lines:

def test_field_max_length_is_enforced(clear_sqlmodel):
    class Hero(SQLModel, table=True):
        id: Optional[int] = Field(default=None, primary_key=True)
        name: str = Field(max_length=5)

    engine = create_engine("sqlite://")

    SQLModel.metadata.create_all(engine)

    with pytest.raises(IntegrityError):
        with Session(engine) as session:
            hero_1 = Hero(name="Deadpond")

            session.add(hero_1)
            session.commit()
            session.refresh(hero_1)

Unfortunately, this doesn't work, as sqlite does not actually enforce something like VARCHAR(5).

One solution would be to use a testcontainer (see here) to run a PostgreSQL instance and test against that, but it would significantly increase the complexity & dependencies of the unit tests.

[mbwhite]

FYI _ stumbled across this trying to debug SQLModel using DB2. The create table SQL ends up using VARCHAR(None) that DB2 rejects; I tried the max_length that didn't work hence, came here.

Not sure if that helps but DB2 doesn't like unbounded varchars

[revoltn]

A vote for me as well. This is, for many database centric applications, a significant issue if the max_length for a string does not create a corresponding text(max_length) or varchar(max_length) etc in the table.

I sthere any way this can be escalated, or would you accept an updated pull request from me if I was able to resolve it?