Adding third unit-test run where Numba jitting is disabled

[aleaverfay]

We are seeing poor test coverage for files with significant numba jitting. This third pass of the unit tests should (in theory?) examine code coverage with jitting disabled via the environment variable NUMBA_DISABLE_JIT. I'll let build kite tell me what it thinks.

[codecov[bot]]

Codecov Report

Merging #218 into master will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #218   +/-   ##
=======================================
  Coverage   93.74%   93.74%           
=======================================
  Files         228      228           
  Lines       10697    10697           
=======================================
  Hits        10028    10028           
  Misses        669      669

Flag	Coverage Δ
#_shrug_Testing_CPU	91.5% <ø> (ø)	:arrow_up:
#_shrug_Testing_CUDA	93.72% <ø> (ø)	:arrow_up:

---

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 31b3188...4ffa2ad. Read the comment docs.

[codecov[bot]]

Codecov Report

Merging #218 into master will increase coverage by 2.03%. The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #218      +/-   ##
==========================================
+ Coverage   93.74%   95.78%   +2.03%     
==========================================
  Files         228      228              
  Lines       10697    10711      +14     
==========================================
+ Hits        10028    10259     +231     
+ Misses        669      452     -217

Flag	Coverage Δ
#_shrug_Testing_CPU	93.53% <100%> (+2.03%)	:arrow_up:
#_shrug_Testing_CPU_w_o_jit	91.97% <65.38%> (?)
#_shrug_Testing_CUDA	93.73% <100%> (ø)	:arrow_up:

Impacted Files	Coverage Δ
tmol/tests/score/ljlk/test_numba_lj.py	100% <100%> (ø)	:arrow_up:
tmol/tests/score/ljlk/test_numba_lk_isotropic.py	100% <100%> (ø)	:arrow_up:
tmol/tests/score/ljlk/test_script_modules.py	100% <100%> (ø)	:arrow_up:
tmol/tests/numba.py	46.42% <100%> (+11.64%)	:arrow_up:
tmol/tests/utility/test_args.py	96.29% <0%> (+1.23%)	:arrow_up:
tmol/score/lk_ball/numba/lk_ball.py	97.06% <0%> (+13.86%)	:arrow_up:
tmol/kinematics/scan_ordering.py	99.09% <0%> (+58.18%)	:arrow_up:
tmol/score/cartbonded/identification.py	100% <0%> (+64.38%)	:arrow_up:
.../numeric/interpolation/cubic_hermite_polynomial.py	100% <0%> (+71.42%)	:arrow_up:

---

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 31b3188...416af8c. Read the comment docs.

[aleaverfay]

This fixes the code-coverage deficit in numba-jitted files, e.g., cartbonded/identification.py:

https://codecov.io/gh/uw-ipd/tmol/pull/218/src/tmol/score/cartbonded/identification.py