[CT-375] Unique_key as list tests reformat

[McKnight-42]

Is there an existing feature request for this?

• I have searched the existing issues

Describe the Feature

Next steps for work for #4858

• Rewriting these tests to use the new pytest framework
• Make them truly cross-database compatible (I think by turning expected models into seeds)
• Add them to the inheritable adapter tests, and turn them on for Redshift + Snowflake + BigQuery + Spark. Delete the copy--pasted versions present in those repos :)

Describe alternatives you've considered

No response

Who will this benefit?

Tests will be much cleaner, and a lot less copy paste between repos

Are you interested in contributing this feature?

No response

Anything else?

No response

[jtcohen6]

Anticipated process for this work:

1. Rewrite the test, added to this repo by update of macro for postgres/redshift use of unique_key as a list #4858, using new pytest framework
2. Rewrite the logic to be database-agnostic, so the exact same code can run everywhere
   • Less-database-specific error message — handled in update of macro for postgres/redshift use of unique_key as a list #4858
   • Using seeds instead of SQL literals to load expected fixtures, to avoid database-specific data type casting
3. Move these tests into the "adapter testing zone" within this repo
4. For each plugin: Inherit the test cases from the core repo, and delete the copy-pasted code that's currently in that plugin.
   • For Snowflake, we need to test with both incremental strategies: merge (default) + delete+insert

[jtcohen6]

@leahwicz As discussed yesterday, we're thinking that the group doing this work might split and do it in parallel with another "adapter zone" test conversion.

My vote would be for test_docs_generate, which has proven especially obnoxious for us in the adapter plugins, requiring frequent copy-paste updates. I think we still want some testing for those capabilities in plugins, since catalog contents can vary. A nice exercise in pytest fixture overrides :)