Enable `mccabe` in `ruff`

[purva-thakre]

If I add C90 to the ruff tools: https://docs.astral.sh/ruff/rules/#mccabe-c90

toqito/channel_ops/partial_channel.py:12:5: C901 `partial_channel` is too complex (12 > 10)
toqito/channels/partial_trace.py:11:5: C901 `partial_trace` is too complex (15 > 10)
toqito/channels/partial_transpose.py:11:5: C901 `partial_transpose` is too complex (12 > 10)
toqito/helper/channel_dim.py:8:5: C901 `channel_dim` is too complex (13 > 10)
toqito/helper/npa_hierarchy.py:63:5: C901 `_gen_words` is too complex (13 > 10)
toqito/helper/npa_hierarchy.py:125:5: C901 `npa_constraints` is too complex (19 > 10)
toqito/helper/tests/test_npa_hierarchy.py:12:5: C901 `cglmp_inequality` is too complex (12 > 10)
toqito/matrices/gell_mann.py:7:5: C901 `gell_mann` is too complex (11 > 10)
toqito/matrix_ops/tensor.py:6:5: C901 `tensor` is too complex (14 > 10)
toqito/matrix_props/sk_norm.py:18:5: C901 `sk_operator_norm` is too complex (27 > 10)
toqito/nonlocal_games/extended_nonlocal_game.py:123:9: C901 `nonsignaling_value` is too complex (25 > 10)
toqito/nonlocal_games/extended_nonlocal_game.py:294:9: C901 `__optimize_alice` is too complex (11 > 10)
toqito/nonlocal_games/nonlocal_game.py:363:9: C901 `__optimize_alice` is too complex (11 > 10)
toqito/nonlocal_games/nonlocal_game.py:473:9: C901 `nonsignaling_value` is too complex (25 > 10)
toqito/perms/permute_systems.py:12:5: C901 `permute_systems` is too complex (16 > 10)
toqito/state_props/is_product.py:59:5: C901 `_is_product` is too complex (11 > 10)
toqito/state_props/is_separable.py:20:5: C901 `is_separable` is too complex (32 > 10)

Adding the errors here to understand what is too complex later

[vprusso]

Interesting. Yes, the is_separable one makes sense as there are quite a few if conditions that need to be checked before the symmetric extension approach is invoked. Good to know!

[Shivansh20128]

Yes. Here is the the reference for what too complex means. I checked the is_product function, and there are many if conditions there as well, so much so that at line 104, it enters the 3rd nested if condition. Maybe that is where the problem is. We could modify it to something simpler, or maybe we could find a way to change the threshold of McCabe complexity. This maybe relevant to that.

[vprusso]

Thanks for the references, @Shivansh20128 .

I think the McCabe complexity measure is a nice one, but similar to test coverage, I suppose I want to be careful that we aren't "chasing the wrong metric". It may very well be the case that some functions will simply have a high McCabe metric, and any refactoring to reduce that would have a net negative impact on the clarity of the code. It might be nice to hook this in as a diagnostic, but I'm not sure if we want to "live and die" by the numbers :). Does that make sense?

[Shivansh20128]

Yes, suree! I just thought that since this issue is open maybe this is something you want to change.