Leftover cosmic ray transport and diffusion changes

[felker]

Follow up from #550 (mostly stuff for @yanfeij to generally weigh in on that PR)

• I still think that it is unfortunate that there is no consistency in variable names between the two CR algorithms. E.g. Dpara/Dperp vs. sigma_diff(0:2, ...), initial condition set via pcrdiff->ecr(k,j,i) vs. pcr->u_cr(4,k,j,i), even though they are in different modules and have different array dimensions as @tomidakn emphasizes
• https://github.com/PrincetonUniversity/athena/wiki/Cosmic-Ray-Transport only describes cr/vmax and EnrollOpacityFunction(). Need to describe some other parameters and functions in the class, especially those related to CR streaming and source functions. Also would be good to include the counterpart of the "equations solved" copied from Jiang and Oh (2018), assuming that they are different from the CR diffusion equations with the inclusion of the velocity-related terms. This is related to making the names consistent between CR diffusion and transport modules. Some items from cr.hpp header:

  Real vlim;
  Real max_opacity;
  AthenaArray<Real> sigma_adv;  // KGF: show an example of how this is set in the enroll functions 
...
  // Function in problem generators to update opacity
...
  void EnrollStreamingFunction(CRStreamingFunc MyStreamingFunction);

  void EnrollUserCRSource(CRSrcTermFunc my_func);

[felker]

delaying until after the 24.0 release