Issue with GAG parameters

[gitoliver]

The old gag builder (glycam.org/gag pre October 2021, currently legacy.glycam.org/gag) uses prep files that contain LIdopA in the different ring shapes (1C4, 4C1, 2SO). I reckon Arunima made them. The website code loads them in and modifies them depending on what's decorating the sugar. All fine, except that the charges used aren't the same as what we use for LIdopA in the regular carbohydrate builder. If you compare the prep files, they have different charges. I imagine what happened is that when they were developing the charges in this paper:
Extension and validation of the GLYCAM force field parameters for modeling glycosaminoglycans
They also calculated new charges for LIdopA (though it's not explicitly mentioned) and that's what Arunima used in the GAG builder. But no-one went back and changed the charges we use for LIdopA whenever the ring form isn't specified. So if you build a sugar on the GAG builder and specify the ring form for IdoA, you get different charges than if you didn't specify it, because different prep files get used. We won't be fixing this on legacy, but we do need to generate the prep files so the new website gives you up-to-date parameters. Oliver has copied the "undecorated" versions of the LIdopA prep files with different shapes into
gmml/dat/prep/GLYCAM_06j-1_GAGS.prep
so we have them somewhere that isn't legacy. But we also need to generate all the decorated versions.