Vectorial model rewrite and update associated tests and documentation

[sjoset]

This is a rewrite of the vectorial model to utilize numpy where possible, assemble results neatly, and bring the results better in line with the fortran version.

All tests in sbpy/activity/gas/tests/test_core.py passed and flake8 was clear.

[codecov]

Codecov Report

Merging #336 (599d9da) into main (462f453) will increase coverage by 1.18%.
The diff coverage is 98.33%.

❗ Current head 599d9da differs from pull request most recent head b679364. Consider uploading reports for the commit b679364 to get more accurate results

@@            Coverage Diff             @@
##             main     #336      +/-   ##
==========================================
+ Coverage   75.65%   76.84%   +1.18%     
==========================================
  Files          78       78              
  Lines        6795     6978     +183     
==========================================
+ Hits         5141     5362     +221     
+ Misses       1654     1616      -38     

Impacted Files	Coverage Δ
sbpy/activity/gas/core.py	97.54% <97.63%> (+6.92%)	⬆️
sbpy/activity/gas/tests/test_core.py	100.00% <100.00%> (+3.92%)	⬆️

... and 4 files with indirect coverage changes

[mkelley]

@sjoset There are some lines of code not being covered by the tests. See the coverage report for details: https://app.codecov.io/gh/NASA-Planetary-Science/sbpy/pull/336/blob/sbpy/activity/gas/core.py Most of them are trivial branches, but the q_t parameter and the binned_production method are especially important. Never mind the scipy import tests and line 497 in the GasComa class, I'll get those separately.

[mkelley]

That failing test will be addressed with #374

[mkelley]

Could you make an entry for this update in CHANGES.rst under New Features, including the improvements and the new binned_production constructor? Also, add any API changes to the relevant section (e.g., angular_substeps was removed from VectorialModel).

Finally, remove the warning in the documentation (gas.rst).

Do you have or do you plan to have a citeable reference for the comparison to the FORTRAN version?

[mkelley]

Please provide a docstring for this and the other VM* data classes.

[mkelley]

Are they meant to be used by users? If not, could you prefix them with "_" to indicate they are private to the submodule? Then the docstrings can just be one-line descriptions.

[mkelley]

Thanks for the fixes and updates!