Create a hyperscan resource

Created by: scarlehoff

As said in PR #1311 even if the "beyond chi2 hyperscan" is still a fluid concept, it is true that we want them to be reproducible, and at the same time I don't think they should be treated as regular fit, so I've created a separate hyperscans folder in the server.

from validphys.loader import Loader
l = Loader()
hyperscan = l.check_hyperscan("test_hyperopt_fit_300621")

Right now a hyperscan is little more than a special case of a fit with tries.json inside (subtlety: having done postfit makes the object into a fit even if tries.json can be found inside, which I think it is fair). Anyway, now the infrastructure is there so future hyperopt-related improvements can be built using the same "technology" as for fits*.

@scarrazza @Zaharid I've added a index-hyperscan.py file to the server and seems to work. Didn't want to mess with the crontab (don't even know if I have the right permissions to do so) so please add it to all the right places (now one would need to run it manually)

I'll rebase #1311 on top of this, so it uses the hyperscan.

wrt vp-hyperoptplot, I know there's a long way to go there, the only thing I will do in this PR with respect to that script is that one could use a hyper parameter scan from the server but then it will fall back to the old scripty behaviour. Later on it can be made to use the hyperscan spec as a spec. But one thing at a time.

A lot of cargo-cult programming in this PR, please let me know whether I missed something important.