[CT-1981] Issue when passing --select and --model to a command together

[aranke]

Did some manual testing by locally overwriting a command that accepts the new select (--select, --model) alias command with print(ctx.obj["flags"].SELECT) and found some issues when passing --select and --model to a command together.

The options are not treated as mutually exclusive via MultiOption and the resulting flags.SELECT value is clobbered by the value from the last alias option:

❯ python3 core/dbt/cli/main.py docs generate --select model_a model_b  --model model_c
('model_c',)

❯ python3 core/dbt/cli/main.py docs generate --model model_c --select model_a model_b
('model_a', 'model_b')

This was unexpected behaviour to me, but for better or for worse this is also how main's parsing currently works (printing args.select and exiting on main to test):

❯ dbt docs generate --select model_a model_b --model model_c
['model_c']
❯ dbt docs generate --select model_a model_b --model model_c
['model_c']
❯ dbt docs generate --model model_c --select model_a model_b 
['model_a', 'model_b']

We should probably open an issue to document this behaviour and investigate a fix that gets us --select/--model option mutual exclusivity if possible. But given that it's the current behaviour on main I don't see it as pressing to fix as part of this PR.

Originally posted by @MichelleArk in #6787 (comment)

[iknox-fa]

@JerCo can we close this as a won't fix due to the fact that we want to deprecate --model?

[aranke]

Tagging the right Jerco, @jtcohen6.

[iknox-fa]

sigh I have done this more times than I'd like to admit.
Thanks @aranke