Merge pull request #370 from ajwheeler/no_CH_in_GES

filter out strong CH lines, because they appear to be problematic (see #365)

src/linelist.jl

@@ -641,9 +641,10 @@ function get_GES_linelist(; include_molecules=true)
     path = joinpath(artifact"Heiter_2021_GES_linelist",
                     "Heiter_et_al_2021_2022_06_17",
                     "Heiter_et_al_2021.h5")
-    h5open(path, "r") do f
+    linelist = h5open(path, "r") do f
         each_species = [Species(s) for s in read(f["species"])]
         filter = ones(Bool, length(each_species))
+
         if !include_molecules
             filter .= [!ismolecule(s) for s in each_species]
         end
@@ -656,6 +657,10 @@ function get_GES_linelist(; include_molecules=true)
               tentotheOrMissing.(Float64.(read(f["gamma_stark"])[filter])),
               read(f["vdW"])[filter])
     end
+
+    # see https://github.com/ajwheeler/Korg.jl/issues/356
+    # there seem to be errors in the CH lines.  Many have very large log(gf) values.
+    linelist = linelist[[!(l.species == Korg.species"CH" && l.log_gf > -1.9) for l in linelist]]
 end
 
 """