upgrade to sqlalchemy 2.0

[sanderegg]

and install sqlalchemy like so:
_base.in
sqlalchemy[postgresql_asyncpg]
_test.in
sqlalchemy[postgresql_asyncpg, mypy]

Includes async features through asyncpg engine.

Here the sqlaclhemy 2.0 migration guide: https://docs.sqlalchemy.org/en/20/changelog/migration_20.html

• SQLALCHEMY_WARN_20 enabled in all [unit/int] tests
• add the following in all the setup.cfg files as soon as the warnings are all gone to prevent going back to pre-2.0 style

  [tool:pytest]
  addopts = -W error::sqlalchemy.exc.SAWarning

TIPs:

sa.select([clusters.c.id, clusters.c.name]) # old style with array
sa.select(clusters.c.id, clusters.c.name) # new style with args

Tasks

• ✨Personalized resources: remove projects to projects nodes table 🗃️ #4384
• ⬆️ ♻️ Maintenance: sqlachemy2 compatibility upgrades #4330
• enable the future flag on sqlalchemy

[pcrespov]

• Since ✨ Personalized resource limits: add API to change node resources ⚠️🗃️ #4374, testing postgres-database will now turn into errors anything related to pre-sql 2.0 code.

[sanderegg]

@sanderegg add progress list of packages/services

[pcrespov]

Quoting @sanderegg .

So the current idea is to remove ALL the SQLAlchemy2.0 warnings from the tests (I do hope this should remove >90% of the problems since we do not have full coverage maybe some stuff could slip by)