Enable interpolation_setup to download precomputed weights for LLC90 or LLC270

[seamanticscience]

I added a method to interpolation_setup, which allows down/loading of precomputed interpolation weights based on the GridSpec (i.e. either 0.5° for ID=:LLC90 or 0.25° for ID=:LLC270).

γ=GridSpec(ID=:LLC270)
 gcmgrid 
  class        = LatLonCap
  path         = /Users/jml1/.julia/artifacts/1bc43784dfbcb787cc4e8018ab61d5c3cdcd5ea3/GRID_LLC270-1.0.0/
  fSize        = [(270, 810), (270, 810), (270, 270), (810, 270), (810, 270)]
  ioSize       = [270 3510]
  ioPrec       = Float32

λ=interpolation_setup(γ)
(lat = [-89.875 -89.625 … 89.625 89.875; -89.875 -89.625 … 89.625 89.875; … ; -89.875 -89.625 … 89.625 89.875; -89.875 -89.625 … 89.625 89.875], f = [4 4 4 4; 4 4 4 4; … ; 3 3 3 3; 3 3 3 3], w = [0.038147161848044925 0.32800134934748143 0.5678137074351091 0.06603778136936453; 0.03588516073953725 0.319724971993188 0.5793635031754644 0.06502636409181044; … ; 0.30958104671667847 0.3210276065472765 0.1880481964656952 0.18134315027034986; 0.3070680754881118 0.31913638642184666 0.19049969226156438 0.1832958458284772], j = [40 40 41 41; 40 40 41 41; … ; 135 135 136 136; 135 135 136 136], lon = [-179.875 -179.875 … -179.875 -179.875; -179.625 -179.625 … -179.625 -179.625; … ; 179.625 179.625 … 179.625 179.625; 179.875 179.875 … 179.875 179.875], i = [808 809 809 808; 808 809 809 808; … ; 136 137 137 136; 136 137 137 136])

This could be expanded to other model grids (for example ID=:CS32), but currently supplying an unrecognized gcmgrid falls back on calculating interpolation weights on the default 2° grid.

γ=GridSpec(ID=:CS32)
 gcmgrid 
  class        = CubeSphere
  path         = /Users/jml1/.julia/artifacts/e425ca76b87758101e7f7543eee62d34f7b4553f/GRID_CS32-1.1/
  fSize        = [(32, 32), (32, 32), (32, 32), (32, 32), (32, 32), (32, 32)]
  ioSize       = [32 192]
  ioPrec       = Float32

λ=interpolation_setup(γ)
(lat = [-89.0 -87.0 … 87.0 89.0; -89.0 -87.0 … 87.0 89.0; … ; -89.0 -87.0 … 87.0 89.0; -89.0 -87.0 … 87.0 89.0], f = [6 6 6 6; 6 6 6 6; … ; 3 3 3 3; 3 3 3 3], w = [0.4266968067160594 0.41659067208909717 0.07741722267931467 0.07929529851552865; 0.4365801480888083 0.40628908780280665 0.07574189695964063 0.08138886714874438; … ; 0.08138952033431214 0.43657942774664493 0.40628852672078924 0.07574252519825363; 0.07929594418742816 0.4266961386495333 0.4165900571683199 0.07741785999471859], j = [16 16 17 17; 16 16 17 17; … ; 16 16 17 17; 16 16 17 17], lon = [-179.0 -179.0 … -179.0 -179.0; -177.0 -177.0 … -177.0 -177.0; … ; 177.0 177.0 … 177.0 177.0; 179.0 179.0 … 179.0 179.0], i = [16 17 17 16; 16 17 17 16; … ; 16 17 17 16; 16 17 17 16])

All other behavior is the same, including interpolation_setup with no arguments, and interpolation_setup with a path to saved JLD2 weight file.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.80%. Comparing base (68950c3) to head (ce5feb2).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #220      +/-   ##
==========================================
+ Coverage   87.75%   87.80%   +0.04%     
==========================================
  Files          37       37              
  Lines        3398     3410      +12     
==========================================
+ Hits         2982     2994      +12     
  Misses        416      416              

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.