pele.landscape.LocalConnect

class pele.landscape.LocalConnect(pot, mindist, tsSearchParams=None, verbosity=1, NEBparams=None, nrefine_max=100, reoptimize_climbing=0, pushoff_params=None, create_neb=<class 'pele.transition_states._nebdriver.NEBDriver'>)

a class to do a single local connect run, i.e. NEB + transition state search

Parameters :

pot : potential object

the potential

mindist : callable

the function which returns the optimized minimum distance between two structures

tsSearchParams: dict

parameters passed to the transition state search algorithm

NEBparams : dict

NEB setup parameters. Use NEBquenchParams for parameters related to the optimization of the band.

nrefine_max : int

the maximum number of NEB transition state candidates to refine

reoptimize_climbing : int

the number of iterations to use for re-optimizing the climbing images after the NEB is done.

pushoff_params : int

parameters for detemining how to find the minima on either side of a transition state

verbosity : int

this controls how many status messages are printed. (not really implemented yet)

See also

DoubleEndedConnect

the routine from which local connect is generally called

pele.transition_states.NEB

one of the core routines

pele.transition_states.NEBDriver

the wrapper which sets up NEB

pele.transition_states.findTransitionState

one of the core routine

Notes

this class takes two minima as input, does an NEB run to find transition state candidates, then refines those candidates into transition states. Finally, we fall off either side of the transition states to fine the minima on either side.

This is the core routine of DoubleEndedConnect. It is separated out in order to make parallelization easier. This class intentionally has no knowledge of the global landscape (database, graph, etc.).

Methods

connect(min1, min2)

NEB to find transition state candidates.