Selecting a subquery with a GROUP BY clause produces invalid SQL

[yomatters]

Describe the bug
Starting in version 0.24.0, selecting a subquery containing a GROUP BY clause produces invalid SQL. The field names in the GROUP BY clause are given aliases, which causes a SQL error when the query is executed.

To Reproduce
This example script illustrates the problem:

from tortoise import fields, run_async, Tortoise
from tortoise.expressions import Subquery
from tortoise.models import Model


class A(Model):
    id = fields.IntField(primary_key=True)
    name = fields.TextField()


class B(Model):
    id = fields.IntField(primary_key=True)
    name = fields.TextField()


async def main():
    await Tortoise.init(db_url="sqlite://:memory:", modules={"models": ["__main__"]})
    await Tortoise.generate_schemas()

    query = A.annotate(b_name=Subquery(B.all().group_by("name").order_by("name").limit(1).values("name")))
    print(query.sql())

run_async(main())

Starting in version 0.24.0, the output of the script contains an invalid field alias b__name in the GROUP BY clause of the subquery:

SELECT "id","name",(SELECT "name" "name" FROM "b" GROUP BY "name" "b__name" ORDER BY "name" ASC LIMIT ?) "b_name" FROM "a"

Expected behavior
Versions 0.23.0 and earlier produce a valid GROUP BY clause without field aliases:

SELECT "id","name",(SELECT "name" "name" FROM "b" GROUP BY "name" ORDER BY "name" ASC LIMIT ?) "b_name" FROM "a"

[PrabhbirJ]

I've submitted a PR that fixes this issue: #2001

The problem was in the _resolve_group_bys method which was incorrectly adding field aliases to GROUP BY clauses. Removed the .as_() call to generate clean SQL.

Tested with the reproduction case from this issue and added a regression test.