Move all hash checking to hash_matches

[leouieda]

Including the exception raising controlled by new parameter strict.
This removes the need for re-calculating the new hash just so we can
include it in the error message. The hash calculation can be heavy
for large files so we don't want to do it twice. Also improve the error
message to include the hash algorithm and suggest what went wrong.

Working towards #152

Reminders:

• Run make format and make check to make sure the code follows the style guide.
• Add tests for new features or tests that would have caught the bug that you're fixing.
• Add new public functions/methods/classes to doc/api/index.rst and the base __init__.py file for the package.
• Write detailed docstrings for all functions/classes/methods. It often helps to design better code if you write the docstrings first.
• If adding new functionality, add an example to the docstring, gallery, and/or tutorials.
• Add your full name, affiliation, and ORCID (optional) to the AUTHORS.md file (if you haven't already) in case you'd like to be listed as an author on the Zenodo archive of the next release.

[danshapero]

I'm assuming that the strict = False case will get used for #152. If you wanted to be really obsessive about branch coverage then you could add an explicit test for that but I don't think it's necessary.

[leouieda]

@danshapero right, I forgot to add the extra tests to this PR. I'll add some explicit tests for this function (right now it gets tested as part of fetch).

Both cases are used within fetch: strict=False when checking for updates and stritc=True for verifying downloaded files. My plan for #152 was to allow known_hash=None which would always result in a True return value. That way bypassing the hash check would only require changing this one function.

[leouieda]

Thanks for the review, by the way 🙂 Much appreciated