Missing 2.0 style ORM functions

[jonyscathe]

Describe the bug
After @seano314's update to support sqlalchemy 2.0 I have had success migrating a lot of my code and test code over to 2.0 style form following the migration guide.

The one thing I have struggled with is with the use of scalar()
According to the migration guide, simple queries for a single value should be changed from something like:

session.query(User).count()

to

session.scalar(
  select(func.count()).
  select_from(User)
)

# or

session.scalar(
  select(func.count(User.id))
)

Currently I don't believe that that new style method of using count() can be mocked using mock-alchemy as it currently stands.
And I'm 100% not sure that this is an easy fix with scalar() in 2.0 now being able to be the start and end of a chain rather than just the end of a chain.
If anyone has any good ideas on how to fix this I am happy to make a PR. But this isn't quite as simple

To Reproduce
Try mock data for something like:

session.scalar(
  select(func.count(User.id))
)

Expected behavior
For it to be possible to make a UnifiedAlchemyMagicMock with data set such that session.scalar() calls can be mocked correctly.

Screenshots
If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: [Ubuntu]
• Version [0.2.6]

Additional context
Another aspect of the new ORM that isn't yet supported is scalar_one though I believe that one is rather easy to fix as it is essentially the same as one.
And the other one is using scalars() as the beginning of a chain as in:

session.scalars(
  select(User)
).all()

However it is easy enough to get the same thing with:

session.execute(
  select(User)
).scalars().all()

So that one isn't super urgent.

As a workaround for now, it is possible to run scalar as something like:

session.execute(
 select(func.count(User.id))
).scalar()

But given that is not the suggested method in the sqlalchemy docs I'd like to be able to use session.scalar() directly