Output Files and Placement for find-orientations

[donald-e-boyce]

I am proposing some changes to placement of output files in find-orientations
because the current placement seems pretty awkward.

find-orientations produces several output files, and only one (grains.out)
goes into the analysis directory. The rest go into the current working
directory with long names based on the name of the analysis directory and
possibly the active material. Then if you run find-orientations several
times on different data, you get a big mess in the current (working) directory.
By contrast, fit-grains writes everything to the analysis directory.

I propose changing the names and locations of the output files for
find-orientations so that everything goes into the analysis directory
with a simpler name. Here is an example of input:

analysis_name: cstudy
working_dir: .

material:
  definitions: materials.h5
  active: ruby

If you run find-orientations with this input, it wiil create the following
files:

• cstudy_ruby_eta-ome_maps.npz
• accepted_orientations_cstudy_ruby.dat
• find-orientations_cstudy_ruby.log
• scored_orientations_cstudy_ruby.npz
• cstudy/grains.out

So it creates the output directory cstudy, but it only puts one file there
and creates four other files in the current directory with
cstudy as part of their names.

I would like to change it so that the output files would all be in the
cstudy directory with names:

• eta-ome_maps_ruby.npz
• accepted_orientations_ruby.dat
• find-orientations_ruby.log
• scored_orientations_ruby.npz
• grains.out

I would also like to make two more small changes:

1. Also write grains.npz.
2. Allow the analysis_name input to be a path (containing "/" characters).

Overall these changes would be pretty easy to implement, and I don't them
as being disruptive to people's workflows.

[psavery]

I completely agree with all of. Thanks for the idea, Don!