correct matpower parsing

[jd-lara]

No description provided.

[codecov]

Codecov Report

Merging #836 (f6a2f2d) into master (2c33f27) will increase coverage by 0.03%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #836      +/-   ##
==========================================
+ Coverage   88.91%   88.95%   +0.03%     
==========================================
  Files         143      143              
  Lines        6929     6706     -223     
==========================================
- Hits         6161     5965     -196     
+ Misses        768      741      -27     

Flag	Coverage Δ
unittests	88.95% <ø> (+0.03%)	⬆️

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

Impacted Files	Coverage Δ
src/models/static_models.jl	25.00% <0.00%> (-15.00%)	⬇️
src/models/generated/MultiStartCost.jl	23.07% <0.00%> (-10.26%)	⬇️
src/models/SalientPoleExponential.jl	64.28% <0.00%> (-2.39%)	⬇️
src/utils/network_calculations/common.jl	18.58% <0.00%> (-1.91%)	⬇️
src/models/HybridSystem.jl	69.73% <0.00%> (-1.52%)	⬇️
src/parsers/common.jl	80.95% <0.00%> (-1.41%)	⬇️
src/models/generated/Transfer.jl	76.47% <0.00%> (-1.31%)	⬇️
src/models/generated/PSSSimple.jl	80.00% <0.00%> (-1.25%)	⬇️
src/models/components.jl	64.28% <0.00%> (-1.24%)	⬇️
...rc/utils/network_calculations/ptdf_calculations.jl	64.91% <0.00%> (-1.19%)	⬇️
... and 35 more

[jd-lara]

@rodrigomha this PR is addressing an issue in the power flow from Matpower and for some reason is breaking some tests from PSID. It seems we parse some of these cases for matpower files.

[jd-lara]

@daniel-thom we are logging the errors from the test datasets now. Is it reasonable to disable the test about error logs?

[daniel-thom]

@daniel-thom we are logging the errors from the test datasets now. Is it reasonable to disable the test about error logs?

I'll make a case for keeping this "feature." It makes developers keep their tests tidy. If you take it away then developers can let error messages spew across the console. When a new developer runs the tests and sees messages in red scrolling across the screen they might think something is wrong. They have to really pay attention to the final result.

This is currently the case in the PowerSimulations dev branch. This "feature" got disabled and then someone let an error message go unsuppressed. A 76-line error message spans the screen every time you run the tests. I find this messy.

It's usually not a big deal to suppress expected error messages with the test_logs macro. If that is too onerous, you can also set the console log level to Logging.Error.level + 1.

That's my case (recommendation). As in the past for this question, I'm OK being out-voted.

[jd-lara]

@daniel-thom I managed to fix most of the tests throwing errors due to bad input data in 37cb8a3 let me know if this is a good solution.

@claytonpbarrows merging this will deviate us from PowerModels while @ccoffrin fixes lanl-ansi/PowerModels.jl#798