perses.rjmc.geometry.BootstrapParticleFilter

class perses.rjmc.geometry.BootstrapParticleFilter(growth_context, atom_torsions, initial_positions, beta, n_particles=18, resample_frequency=10)[source]

Implements a Bootstrap Particle Filter (BPF) to sample from the appropriate degrees of freedom. Designed for use with the dimension-matching scheme of Perses.

__init__(growth_context, atom_torsions, initial_positions, beta, n_particles=18, resample_frequency=10)[source]
Parameters:
growth_context : simtk.openmm.Context object

Context containing appropriate “growth system”

atom_torsions : dict

parmed.Atom : parmed.Dihedral dict that specifies what torsion to use to propose each atom

initial_positions : np.ndarray [n,3]

The positions of existing atoms.

beta : simtk.unit.Quantity

The inverse temperature, with units

n_particles : int, optional

The number of particles in the BPF (note that this is NOT the number of atoms). Default 18.

resample_frequency : int, optional

How often to resample particles. default 10

Methods

__init__(growth_context, atom_torsions, …)
Parameters: