Add warning for AgMIP probabilty distributions and return nothing

[lrennels]

Issue: The get_probdists_gtap_df function does not take the damage function as an argument, so it returns samples from a probability distributionconstructed from a distribution of lowDF, midDF, and highDF for all damage function settings.

Update: The function will now return those samples only if the damage function is one of lowDF, midDF, and highDF, and otherwise return nothing since we do not currently have a method to produce distributions for that damage function. In addition, this PR adds the parameter gtap_name to the component so that any model using the component can easily recover the name of the damage function used to specific the component.

@djsmithumn and @bryanparthum for awareness

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 0% with 11 lines in your changes are missing coverage. Please review.

Project coverage is 78.46%. Comparing base (177e8e7) to head (76ed7e4).

Files	Patch %	Lines
src/core/mcs.jl	0.00%	11 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #30      +/-   ##
==========================================
- Coverage   82.25%   78.46%   -3.80%     
==========================================
  Files           6        6              
  Lines          62       65       +3     
==========================================
  Hits           51       51              
- Misses         11       14       +3     

Flag	Coverage Δ
unittests	78.46% <0.00%> (-3.80%)	⬇️

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

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[bryanparthum]

Nice addition @lrennels! Will this alter the overall size of the mcs? i.e., will the random parameters of the other (non-ag) components receive different values for no other reason than changing the Ag damage function setting? Or does the passing of nothing maintain the ordering for other mcs parameter draws?

[lrennels]

@bryanparthum ah that's a good point, I assume your concern being if you wanted to apples to apples compare a Monte Carlo Simulation with agMIP or a non-agMIP and that in this case the component parameter draws would differ? You're right this would break that consistency, as you would have fewer calls ... one way I could fix that would be to make the function act the same each time but then just return either a Vector or nothing.

[bryanparthum]

You got it (assuming this is still sensitive to the n in the mcs). I think passing a vector of length n of nothings would be okay, no? But you obviously know what the most consistent way to code it would be. If I were (am?) a user looking at the mcs parameters, a vector of nothing for the Ag component would be intuitive.

[lrennels]

Really good catch, a vector of nothing would be fine but won't fix the issue, since that still wouldn't call rand. Instead see the commit I just added, which basically makes sure the get_prob_dists function behaves the same way re. sampling and then just chooses what to return.