Those lines try to locate the file
gmic_film_cluts.gmz in various local folders, and if not found, try to download it from the G'MIC server.
$> is the indice of the loop and has here values
0 (first iteration) or
1 (second iteration). The first iteration tries to load the clut file locally, while the second iteration tries to download the file from the Internet (if the first iteration failed, or if the requested clut is not present in the local file).
Anyway, allowing a custom path for storing the clut file has not been planed. And I wouldn't recommend you have a copy of this file in your own custom path, because it happens I add some new CLUTs from time to time, and at the end, the update file always comes in the G'MIC resource folder (which is
~/.config/gmic on Unix).
Why not simply creating a folder
~/.config.gmic if it doesn't exist and uses this to access G'MIC specific files ?
All important data files needed by filters are stored there, so I don't see the point of having a copy of these files elsewhere