using pytest command instead of bash

[shimwell]

Proposed changes

Following on from the improvments to the CI that @RemDelaporteMathurin made I am keen to rewrite on of the actions.

Previously this action didn't fail even when the tests failed.

I think is is due to running the tests via a bash script.

So this PR runs the tests directly with a pytest command

Types of changes

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Code refactoring
• Documentation Update (if none of the other choices apply)
• New tests

Checklist

• Pep8 applied
• Unit tests pass locally with my changes
• I have added tests that prove my fix is effective or that my feature works
• I have added necessary documentation (if appropriate)

[shimwell]

Interestingly the improved GitHub actions CI appears to have caught an error that appears to pass on CircleCI

[codecov]

Codecov Report

Merging #72 (746863f) into develop (6dc9ef2) will increase coverage by 0.18%.
The diff coverage is 100.00%.

@@             Coverage Diff             @@
##           develop      #72      +/-   ##
===========================================
+ Coverage    97.72%   97.90%   +0.18%     
===========================================
  Files           75       75              
  Lines         5096     4870     -226     
===========================================
- Hits          4980     4768     -212     
+ Misses         116      102      -14     

Impacted Files	Coverage Δ
paramak/__init__.py	100.00% <ø> (ø)
paramak/parametric_reactors/ball_reactor.py	98.91% <ø> (-0.01%)	⬇️
...parametric_reactors/center_column_study_reactor.py	100.00% <ø> (ø)
...aramak/parametric_reactors/eu_demo_2015_reactor.py	100.00% <ø> (ø)
paramak/parametric_reactors/submersion_reactor.py	98.63% <ø> (-0.01%)	⬇️
paramak/utils.py	97.30% <ø> (+4.55%)	⬆️
paramak/reactor.py	91.95% <100.00%> (-0.95%)	⬇️
paramak/shape.py	96.44% <100.00%> (-1.10%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6dc9ef2...746863f. Read the comment docs.

[RemDelaporteMathurin]

@shimwell the commit e87c91f adds the action provided by Codecov for GA which successfully uploaded to codecov, fixing #23 .
However, the tests still don't pass 😄

[shimwell]

Yep it appears to be failing on the first use of Cubit.

There is a problem combining use of pymoab and cubit in the same script as they both point to different hdf5 libraries

Here is a description of that problem https://forum.coreform.com/t/conflicting-hdf5-libraries/1237/3?u=jshimwell

So there are four difficulties we have had with containerising cubit,

• getting the lic file into the right folder /root/.config/Coreform/licenses/
• having the correct filename name (must be cubit-learn.lic)
• getting the contents of the lic file correct . should be a date Fri May 28 2021
• avoiding the pymoab hdf5 conflict by not using pymoab and cubit in the same script

I guess the problem in the tests is the last one but tricky to check as no output is printed (even with verbose options turned on)

[RemDelaporteMathurin]

And it's not failing with Circle

[shimwell]

It also passes when I download the docker and run locally.

[shimwell]

Ok I think I've got a solution for all of this. It is a bit radical but let me know what you think

I shall write it up in an issue #75