instanton rate with fix_atoms active

[Dominik Primožič]

Hello,

I've been playing around with instanton rate calculations and I started using fix_atoms to maintain the system I am intersted in. This introduces reduced degrees of freedom that the engine itself easily computes. 

The trouble comes from analysis as I' ve noticed there is no filter option for TS or instanton case in the ipi.utils.tools.instanton_postproc.py script. I did implement a filter that cleans out removed dof from all arrays but now I come to a standstill.

1) Should I reduce the system completely, natoms, beads.q, m, beads.m3 (which in clean_hessian function says is supposed to be vector with degrees of freedom) because then I dont know if sprint_hessian, B_N and spring_pot are computed correctly. 

2) I read the references and the code and I can't seem to place all the terms so I am not too confident about rewriting too much 

for instanton rate Q_vib its 

logQvib= (-np.sum( np.log(betaP * hbar * np.sqrt(np.absolute(np.delete(d, [1])))))) + nzeros * np.log(nbeads)+ np.log(nbeads)
and I don't know where the secong logN comes from

Thank you very much if anyone is more familiar with this