ML_Trajectories module

MLTrajTask

class esteem.tasks.ml_trajectories.MLTrajTask(**kwargs)

Defines a task to use a Machine Learning calculator to generate Molecular Dynamics trajectories

The main routine is esteem.tasks.ml_trajectories.run()

run()

Main routine for the ML_Trajectories task

make_parser()

Makes a parser to setup input variables

Task Arguments

usage: __main__.py [-h] [--seed SEED] [--traj_suffix TRAJ_SUFFIX]
                   [--calc_seed CALC_SEED] [--calc_suffix CALC_SUFFIX]
                   [--calc_dir_suffix CALC_DIR_SUFFIX]
                   [--calc_prefix CALC_PREFIX] [--target TARGET]
                   [--md_timestep MD_TIMESTEP] [--md_friction MD_FRICTION]
                   [--md_steps MD_STEPS]
                   [--md_init_traj_link MD_INIT_TRAJ_LINK]
                   [--geom_prefix GEOM_PREFIX] [--continuation CONTINUATION]
                   [--debugger DEBUGGER] [--store_full_traj STORE_FULL_TRAJ]
                   [--freq FREQ] [--temp TEMP] [--ntraj NTRAJ] [--nsnap NSNAP]
                   [--nequil NEQUIL] [--which_trajs WHICH_TRAJS]
                   [--carve_trajectory_radius CARVE_TRAJECTORY_RADIUS]
                   [--recalculate_carved_traj RECALCULATE_CARVED_TRAJ]
                   [--constraints CONSTRAINTS] [--dynamics DYNAMICS]

Named Arguments

--seed, -s	Base name stem for the calculation (often the name of the molecule)
--traj_suffix, -S
	Suffix for the trajectory files to be generated Default: “mldyn”
--calc_seed, -Z
	Seed for the calculator
--calc_suffix, -C
	Suffix for the calculator (often specifies ML hyperparameters) Default: “”
--calc_dir_suffix, -D
	Prefix for the calculator (often specifies directory)
--calc_prefix, -P
	Prefix for the calculator (often specifies directory) Default: “”
--target, -t	Excited state index, zero for ground state Default: 0
--md_timestep, -q
	Timestep in ASE units Default: 0.023759962463473984
--md_friction, -r
	Langevin friction coefficient Default: 0.002
--md_steps, -Q	Number of MLMD steps between each snapshot Default: 100
--md_init_traj_link, -I
	Path to initial MD trajectory file, relative to base directory. May contain {{solu}} and {{solv}} for substitution by solute and solvent names respectively.
--geom_prefix, -G
	String to append to filenames for initial geometries. May contain {{solu}} and {{solv}} for substitution by solute and solvent names respectively. Default: “”
--continuation, -o
	Whether to continue an existing run Default: False
--debugger, -g	Debugger class to run on trajectory after each snapshot, as a sanity check
--store_full_traj, -f
	Store full step-by-step trajectory rather than just snapshots every md_steps Default: False
--freq, -F	Post-process trajectory into IR spectrum Default: False
--temp, -T	Temperature for thermostat Default: 300.0
--ntraj, -n	Number of separate trajectories in full ensemble Default: 1
--nsnap, -N	Number of snapshots to record in trajectory Default: 200
--nequil, -e	Number of discarded equilibration snapshots before data is recorded Default: 10
--which_trajs, -w
	Which of the separate trajectories are to be run in this task
--carve_trajectory_radius, -R
	Radius around solute to carve trajectory at
--recalculate_carved_traj, -J
	Use snap_wrapper (if present, wrapper if not) to recalculate energies and forces after carving Default: False
--constraints, -c
	Constraints (ASE constraints class)
--dynamics, -d	Dynamics (ASE Dynamics class)

Standalone module routines

Defines a task to use a Machine Learning calculator to generate Molecular Dynamics trajectories, and also to process the results to generate infrared spectra from trajectory data

esteem.tasks.ml_trajectories.load_trajectory_dipole(seed_state_str, traj_suffix, ntraj, nsnaps, mdsteps, extension='.traj')

Loads a set of saved trajectory files and extracts the dipole moment as a function of time

esteem.tasks.ml_trajectories.calculate_ir_spectrum(mu_t, dt, freq_scale_fac, sigma)

Processes the dipole moment as a function of time for a collection of trajectories, to calculate IR absorption spectrum