Better handling of bjobs-related errors

[mbhall88]

Changed

• When log file is not found, we now return failed instead of running. Returning running was effectively causing an infinite loop.
• Unknown line status now also returns failed instead of raising an error. Raising errors without returning a status causes problems with the snakemake master process.

I have also created a Makefile to make a few tasks easier/more reproducible. One thing I uncovered in the process was that we were calculating code coverage very incorrectly. We were testing the test code coverage. I have changed this, but it will cause a big drop in project coverage, but the important files are well covered.

[codecov-io]

Codecov Report

Merging #34 into master will decrease coverage by 11.10%. The diff coverage is 81.81%.

@@             Coverage Diff             @@
##           master      #34       +/-   ##
===========================================
- Coverage   99.43%   88.32%   -11.11%     
===========================================
  Files           5        6        +1     
  Lines         881      454      -427     
  Branches        0       44       +44     
===========================================
- Hits          876      401      -475     
- Misses          5       46       +41     
- Partials        0        7        +7     

Flag	Coverage Δ
#unittests	88.32% <81.81%> (-11.11%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
{{cookiecutter.profile_name}}/OSLayer.py	79.06% <55.55%> (ø)
{{cookiecutter.profile_name}}/CookieCutter.py	72.00% <66.66%> (ø)
{{cookiecutter.profile_name}}/lsf_status.py	84.80% <95.23%> (ø)
tests/test_lsf_config.py
tests/test_lsf_submit.py
tests/test_memory_units.py
{{cookiecutter.profile_name}}/lsf_config.py	100.00% <0.00%> (ø)
{{cookiecutter.profile_name}}/lsf_submit.py	88.15% <0.00%> (ø)
{{cookiecutter.profile_name}}/memory_units.py	100.00% <0.00%> (ø)
... and 2 more