Inquiry - What is the algorithmic complexity of the `hash_files` function & what number of apps / migration files are computationally feasible

[TaufiqSyed]

First off - fantastic project - feel free to close this issue.

I wanted to check regarding this, since I saw on line 224-225 in django-migrations-ci/django_migrations_ci/django.py

    for n in range(depth + 1):
        combinations = itertools.combinations_with_replacement(checksums, n)

I'm sure this would work seamlessly with most scenarios but I wanted to ask, what would be the Big O complexity of the entire calculation. And as it is exponential, what values is this feasible for, I wanted to try running this on a cloud pipeline.

Thanks a lot.

[iurisilvio]

It is exponential with number of apps you have with migrations. Hashes are md5.

If it doesn't hit cache occasionally, it'll just execute the slow migration once, so there is no big reason to be too aggressive with depth. It is a trade off between one slow run once in a while or being a bit slower every run. I'd go with n=1 or 2 and it should be fine in most cases.

Just try it in your codebase and adjust.