959 Add more tests for various load models Part 1

[JacobDomagala]

Part 1

List of Load Models to be tested:

• LinearModel
• MultiplePhases
• NaivePersistence
• PersistenceMedianLastN

Following will be covered in Part 2:

• Norm
• CommOverhead
• SelectSubphases

[codecov]

Codecov Report

Merging #1001 into develop will increase coverage by 0.08%.
The diff coverage is 86.32%.

@@             Coverage Diff             @@
##           develop    #1001      +/-   ##
===========================================
+ Coverage    77.54%   77.62%   +0.08%     
===========================================
  Files          656      666      +10     
  Lines        25390    25554     +164     
===========================================
+ Hits         19688    19836     +148     
- Misses        5702     5718      +16     

Impacted Files	Coverage Δ
...nit/collection/test_model_multiple_phases.nompi.cc	83.33% <83.33%> (ø)
...t/collection/test_model_naive_persistence.nompi.cc	85.71% <85.71%> (ø)
...s/unit/collection/test_model_linear_model.nompi.cc	87.09% <87.09%> (ø)
...tion/test_model_persistence_median_last_n.nompi.cc	87.09% <87.09%> (ø)
.../vt/vrt/collection/balance/model/composed_model.cc	88.88% <100.00%> (+11.11%)	⬆️
...lection/balance/model/persistence_median_last_n.cc	94.11% <100.00%> (ø)
...vt/vrt/collection/balance/model/multiple_phases.cc	88.88% <0.00%> (ø)
...llection/balance/model/persistence_median_last_n.h	100.00% <0.00%> (ø)
... and 7 more

[PhilMiller]

@JacobDomagala could we finish this PR and merge it, and have you open and address issues for the remaining model classes separately? I have some somewhat urgent changes to make on #995 that I'm building on top of this branch.

The only thing I'd ask to have added here before we go ahead with the merging is the fix for the defect you spotted in the Norm model of not actually looping over subphases.

[lifflander]

Here is an overview of what got changed by this pull request:

Clones added
============
- tests/unit/collection/test_model_multiple_phases.nompi.cc  6
- tests/unit/collection/test_model_persistence_median_last_n.nompi.cc  8
- tests/unit/collection/test_model_linear_model.nompi.cc  8
- tests/unit/collection/test_model_naive_persistence.nompi.cc  4
         

See the complete overview on Codacy

[PhilMiller]

This raises an interesting question of whether the linear model should be changed to clamp its result to be >= 0, since negative loads don't really make sense. We can consider that later, though,

[lifflander]

Looks good to me.