Adding 2 example tests for the pytest lesson.

[sjayellis]

@orbeckst Here are the additional tests for the pytest lesson.

Changes made in this Pull Request:

• Added a non-fixture test for charmm data intended to be referenced for an exercise.
• Added a test for an expected failure in the extract function.

[pep8speaks]

Hello @sjayellis! Thanks for opening this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file mdgeom/tests/test_info.py:

Line 58:1: W293 blank line contains whitespace
Line 60:1: W293 blank line contains whitespace
Line 65:1: E266 too many leading '#' for block comment

[orbeckst]

Lgtm — optional comments on precision of fixtures (and these comments may well equally apply to anything I originally added — just occurred to me while looking at yours that it may look nicer to choose ground truth number, although just choosing the numbers that you get when you run the code is more straightforward).

I’m going to merge and if you feel like changing anything we can do this in another PR.

[orbeckst]

You could probably set to 1.0 as this is the “true” Timestep. I think this is only due to float32 precision looking a bit funny.

If the tests don’t pass with 1.0 then I wouldn’t change because then you have to start a discussion about the importance of doing floating point comparisons correctly—-which is an important topic but perhaps not something you want to discuss at this moment.

[orbeckst]

Similar to other comment: could perhaps be set to 97.0

[sjayellis]

I did a quick test, setting dt to 1 and totaltime to 97 for charmm passes tests.
Similarly, setting the dt to 100 and totaltime to 900 for gmx passes.

I do think the cleaner numbers look better, but don't have a particular preference on which we use. I intended to have the everyone copy those text blocks instead of writting them, so it should be straightforward either way.

[orbeckst]

Thanks for checking!

Let's leave it with rough numbers for right now. If we want to prettify all reference numbers we could do it in a separate PR. I will go with your decision.