Add LTE calculation for GCMC

1. LTE for GCMC
   Every time it will select 2 sites, one Na/Va and one Si/P, and evaluate their energies after flip both of them. So it's a 2 flip instead of single-flip LTE and it should be called LTE2 more appropriately. I tried to do LTE on NASICON however I found Phi_LTE is almost the same as the potential_energy since Phi = E-TS-ux and S -> 0 when T->0; Phi_lowT -> E-ux = potential_energy. Therefore it's easier just use potential energy as Phi_lowT. When summing the energies for all the possible flips, they are taken half to get rid of the double counting.

2. The input keywords for ensemble:
   Now we can use both GrandCanonical/grand_canonical and ChargeNeutralGrandCanonical/charge_neutral_grand_canonical

Charge Neutral Monte Carlo

Charge Neutral Monte Carlo is finally working... (only works for our NASICON system)

Use 'grand_canonical' for ensemble

Initial coupled MC Jerry

Initial coupled MC Jerry