improve badelf and warren_lab testing

[jacksund]

@SWeav02 a couple bugs were showing up when I was doing last minute checks before the v0.15.0 release. The bugs were all in the badelf and warren_lab apps, so we'll need to add unit tests for them. Also, I wasn't able to run a quick test on your full badelf workflow -- it was too much for my tiny laptop haha. Could we add a lighter workflow preset that runs in <5min on 4 cores for Ca2N?

[SWeav02]

Yes there should be a way. We'd want an incredibly small grid, like 5x5x5 or something. The badelf workflow has a cores parameter which we can set to 4. Also, We'll need to set the check_for_covalency parameter to false because a grid that size will almost certainly throw an error there. It may also fail to find an electride site due to the rough grid, but hopefully it succeeds because the algorithm behaves differently with and without a dummy atom. The test would be much more inclusive if it contains the dummy atom.

I was going to reset my github fork today so that I could fix any issues that might arise from the name changes in #580. I also just ran the algorithm on a bunch of electrides and there's an index error that only shows up occasionally that I need to figure out. I'll work on these today and also see if I can make a quick test for the alg.