Remove unused import for SQLAlchemy 2 compatibility

[WilliamGentry]

This PR removes an unused import that was preventing the databricks-sql-python module from being compatible with SQLAlchemy 2.

In our company's project, we have fully migrated to SQLAlchemy 2.0.4. However, the current databricks-sql-python module obviously requires SQLAlchemy >=1.3.0. Despite this, I found that by simply removing a single unused import, the module could be effectively used in our SQLAlchemy 2.0.4 project.

I ran the unit tests before and after removing the import. In both cases, the results were identical with 92 tests passed and 2 skipped. This makes sense since the imported processors object was never used anywhere.

While the ideal solution would be to update the entire project to be compatible with SQLAlchemy 2.0, this smaller change allows users to integrate databricks-sql-python into their SQLAlchemy 2 projects without any immediate breaking issues, if they're willing to use the connector out-of-spec. This serves as a temporary workaround for those users until a full migration to SQLAlchemy 2 can be achieved (I'm happy to help with that)

[susodapop]

Thanks for doing the research on this! Of course this is an easy merge. Did you have any issues installing with SQLAlchemy==2.x given that we're pinned to ^1.x right now?

I thought that if we merge this it will still fail on install because of the dependency specification in pyproject.toml.

[susodapop]

Change LGTM just wondering if we should update pyproject.toml for this as well. Even experimentally as a .dev release?

[WilliamGentry]

Like you inferred, installing does fail for us. We're largely experimenting right now so we're working off my fork where I bumped SQLAlchemy to ^2.0 and Pandas to ^2.0. An official or .dev release with loosened requirements would be great.

For the record all unit tests pass when run with the upgraded deps (as long as processors import removed) and so far connect library works fine with upgraded deps for our workload (mostly basic SELECT query construction).

[susodapop]

Sorry to leave this hanging so long. I confirmed the tests pass and am preparing to merge. I'll cut you a dev release with the SQLAlchemy version relaxed to include sqlalchemy>=2.0 -- I suspect we have more changes needed to properly support sqla v2 but this will be a start.

[susodapop]

I've forked main as of a few moments ago into a branch called sqla2. There is one commit on that branch right now which relaxes the SQLAlchemy requirement to allow version 2 and sets the project version to 2.7.1.dev3, which you can now install from Pypi with pip install databricks-sql-connector==2.7.1.dev3.

I note that most of our e2e tests fail because Engine.execute() is deprecated in sqla2. But this should give you something to experiment with 😃