Composition verification for external sources

[dmontgomeryNREL]

This PR adds a test to the EB_ODEQty regression test, namely composition-test-2d.inp, which verifies the implementation of user-defined external sources for density $\rho$ and species mass fractions $Y_m$. The initial mass fractions are set to $Y_{CO2} = 0.15$, $Y_{N2} = 0.25$, and $Y_{AR} = 0.6$.

After $t_s$ seconds, CO2 is added at a rate $S_{ext,\rho Y_{CO2}}$, which also causes the density to increase at a rate of $S_{ext,\rho} = S_{ext,\rho Y_{CO2}}$.

The results can be plotted against the analytical solutions in plot-comps.py (see figure below) and tested in the CI via test.py, where errors must be less than a specified maximum error threshold. This was done by adding add_test_rt to CMakeLists, which runs any python script called test.py in the specified case directory.

Additionally, this PR fixes a minor bug in problem_modify_ext_sources, where time is passed in via a_timestamp_old instead of a_timestamp_new due to the explicit treatment of the external source terms.

[dmontgomeryNREL]

Documenting the problem with ubuntu-24.04 EB-ON CI test:
The new test.py test for EB_ODEQty/composition-test-2d.inp passes during the macOS-latest EB-ON test, but fails during the ubuntu-24.04 EB-ON test. As shown in commit #c7a863a, the macOS and ubuntu tests run the associated PeleLMeX executable with four processors. The ubuntu test appears to be running the PeleLMeX executable four times while the macOS test runs the test in parallel. The temporals/tempExtremas data for each test are included in the images below.

Data read in during macOS test:

Data read in during the ubuntu test (4 processors each adding to tempExtremas):

[jrood-nrel]

Let's see if MPI works better if we step ubuntu back #446 until I figure what happened to MPI in the latest image.

[jrood-nrel]

So we can use ubuntu-24.04 which is also now ubuntu-latest on everything except for when we run with MPI. Looks like we need ubuntu-22.04 to run with MPI. I'm trying to figure out what happened to MPI on the new image.

[dmontgomeryNREL]

@jrood-nrel, any idea why there are expected tests for CPU-CMake ubuntu-24.04 in the testing queue?

[jrood-nrel]

Those have to be updated after this is merged. It is because the required tests were renamed.

[baperry2]

Looks good, just the minor change for ignore-unused.

[baperry2]

This ignore unused and the one below shouldn't be needed. ignore-Unused should only be needed when certain variables are excluded by compile time macros, not runtime conditionals

[dmontgomeryNREL]

Thanks for clarifying. I made the update and merged the most recent branch.