new logic for tests - bad and good example; hook checking nb outputs;

[AgnieszkaZaba]

No description provided.

[slayoo]

To make the hooks maintainable, it would be great to store in the repo some example files that demonstrate when the hooks should fail. Chat suggests the following pattern:

Could it work?

[Sfonxu]

To make the hooks maintainable, it would be great to store in the repo some example files that demonstrate when the hooks should fail. Chat suggests the following pattern: [...] Could it work?

I think it could, with maybe a more clear naming convention (like tests/passing/(hook_name).py, tests/failing/(hook_name).py )?

Same goes for example notebooks I suppose. And maybe a good idea would be to add a CI job to run the hooks on those files as well.

[AgnieszkaZaba]

Needed to actually make commit. We can think about running this pre-commits only on CI.

[slayoo]

like in the above suggestion?

[AgnieszkaZaba]

Yes, but then we have to remove this hook from config because the suggested code will not run on excluded files.

[AgnieszkaZaba]

• run pre-commit hooks on both paths (bad and good) with fail mark
• add test with edge cases (injected cells in test)