dptools.cli package

class dptools.cli.BaseCLI(parser)

Bases: object

Base class for CLI commands. More or less just a template for reference.

Parameters:: parser (argparse.ArgumentParser) – argparse parser for parsing CLI command arguments

add_args(): Command specific arguments

help_info = ''

main(args): Command specific main method

class dptools.cli.MyFormatter(prog, indent_increment=2, max_help_position=24, width=None)

Bases: ArgumentDefaultsHelpFormatter

HelpFormatter that replaces reST refs in docstrings with readthedocs URLs to print to console. Also dedents and retains docstring format.

dptools.cli.main()

Submodules

dptools.cli.convert module

class dptools.cli.convert.CLI(parser)

Bases: BaseCLI

Convert between structure ASE/VASP/LAMMPS file types.

Complete documentation here

Examples:

$ dptools convert thing.cif thing.traj
$ dptools convert md_000/vasprun.xml md_001/vasprun.xml full_md.traj
$ dptools convert md_???/vasprun.xml full_md.traj # equivalent to above
$ dptools convert -i ::10 vasprun.xml condensed_traj.traj

add_args(): Command specific arguments

get_kwargs(inputs)

help_info = 'Convert between structure file types (e.g., .xml to .db)'

main(args): Command specific main method

dptools.cli.get module

class dptools.cli.get.CLI(parser)

Bases: BaseCLI

Get params.yaml for specific simulation type. Can also be used to get in.json file (deepmd-kit training parameters).

Complete documentation here

Examples:

$ dptools get cellopt
$ dptools get list # see available simulation options
$ dptools get nvt-md
$ dptools get nvt-md.900K # custom simulation params
$ dptools get in.json # get training param file

add_args(): Command specific arguments

help_info = 'Get params.yaml for specific simulation type'

main(args): Command specific main method

dptools.cli.info module

class dptools.cli.info.CLI(parser)

Bases: BaseCLI

Display information on set DP models and Slurm parameters.

Complete documentation here

Examples:

$ dptools info # display all set environments
$ dptools info -m water # display specific labelled environment

add_args(): Command specific arguments

static get_env_name(env_file)

Get corresponding name for .env.* file, e.g. water from .env.water.

Parameters:: env_file (str) – path to .env file to get corresponding label (name).
Returns:: label (str) – name of environment that is appended to .env file.

help_info = 'Show loaded DP models and sbatch parameters'

main(args): Command specific main method

set_info(env_file)

Get and set all env key-value info.

Parameters:: env_file (str) – path to .env file to get and set info for.

summarize(): Print out semi-formatted env info.

dptools.cli.input module

class dptools.cli.input.CLI(parser)

Bases: BaseCLI

Create deepmd-kit training set folder from ASE .db, .traj, or vasprun.xml files.

Complete documentaion here

Examples:

$ dptools input 00_system1.db 00_system2.db
$ dptools input 0*_sys*.db # equivalent to above
$ dptools input 00_system1/vasprun.xml 00_system2/vasprun.xml
$ dptools input -p /path/to/dataset/folder 0*.db

add_args(): Command specific arguments

help_info = 'Set up deepmd-kit training input from ASE/VASP output'

main(args): Command specific main method

set_name(input_file)

dptools.cli.parity module

class dptools.cli.parity.CLI(parser)

Bases: BaseCLI

Generate parity plots comparing DP model accuracy for energy and force (and stress if available) predictions with ab-initio values.

Complete documentation here

Examples:

$ dptools parity
$ dptools parity /path/to/dataset/system*/test/set*
$ dptools parity -m ../old_graph.pb test_set.traj
$ dptools parity -l mae

add_args(): Command specific arguments

help_info = 'Generate energy and force (and stress if available) prediction parity plots for DP model'

main(args): Command specific main method

static read_systems()

Load test sets from system paths in in.json training parameter file.

Returns:: list of str – Paths to system test sets found in training json file.

dptools.cli.reset module

class dptools.cli.reset.CLI(parser)

Bases: BaseCLI

Reset models or sbatch parameters for environments made using dptools set ....

Complete documentation here

Examples:

$ dptools reset all
$ dptools reset -m water all
$ dptools reset sbatch

add_args(): Command specific arguments

help_info = 'Reset model or sbatch params for default or labeled dptools env'

main(args): Command specific main method

dptools.cli.run module

class dptools.cli.run.CLI(parser)

Bases: BaseCLI

Run LAMMPS simulations using trained DP models.

Complete documentation here

Examples:

$ dptools run opt start.traj # simple atomic position optimization
$ dptools run cellopt start.traj # simple unit cell optimization
$ dptools run /path/to/params.yaml start.traj # custom param file simulation
$ dptools run -s eos 0*/start.traj # submit slurm job eos simulations on multiple structures
$ dptools run -s -m water nvt-md start.traj # submit slurm nvt-md run using set water model

add_args(): Command specific arguments

help_info = "Run simulation using trained DP model (USE COMMAND 'dptools set path/to/graph.pb' first)"

main(args): Command specific main method

run(): Sequentially setup and run simulations on all structure inputs.

set_model(model_label)

Load specific model (if model_label is not None) and set corresponding graph and type_map.

Parameters:: model_label (str or None) – Label of specific model environment to load.

set_params(calc_arg)

Set simulation parameters either from simulation keyword or params.yaml file.

Note

If using parameter file, you must at least retain the .yaml extension.

Parameters:: calc_arg (str) – simulation keyword (e.g. opt) or path to params.yaml file.

set_structures(structures)

Read and set structure inputs as ase.Atoms objects and set corresponding dirs. Only reads last index unless calc_type == spe, in which case single points are ran on all images of structure file.

Parameters:: structures (list of str) – Paths to structure inputs to run simulations on (.traj, .xyz, .cif, etc.).

submit_jobs(sub=True)

Setup and (optionally) submit slurm jobs on all structure inputs.

Parameters:: sub (bool, optional) – Submits slurm jobs if True, else only writes input files without submitting.

dptools.cli.sample module

class dptools.cli.sample.CLI(parser)

Bases: BaseCLI

Intelligently select new training configurations using an ensemble of models and the approach described in DP-GEN (DOI: 10.1016/j.cpc.2020.107206).

Complete documentation here

Examples:

$ dptools sample -n 200 nvt-md.traj
$ dptools sample -n 100 --lo 0.05 --hi 0.25 nvt-md.traj
$ dptools sample -m water_ensemble -p npt-md.traj

add_args(): Command specific arguments

help_info = 'Select new training configs from MD traj using force prediction deviations from ensemble of DPs'

load_ensemble(ensemble)

main(args): Command specific main method

plot(steps=False)

sample(configs, args)

set_configs(configs)

dptools.cli.set module

class dptools.cli.set.CLI(parser)

Bases: BaseCLI

Set DP models, Slurm settings, calculation parameters, or deepmd-kit training parameters

Complete documentation here

Examples:

$ dptools set /home/user/projects/dp_models/h2o.pb
$ dptools set 00/graph.pb 01/graph.pb 02/graph.pb 03/graph.pb # ensemble
$ dptools set -m water /home/user/projects/dp_models/h2o.pb
$ dptools set -m hpc_para parallel_settings.sh
$ dptools set -m hpc_serial serial_settings.sh
$ dptools set params.yaml
$ dptools set /path/to/in.json

add_args(): Command specific arguments

help_info = 'Set DP model defaults, calculation parameters, or sbatch settings'

main(args): Command specific main method

set(thing, **kwargs)

Set file parameters to .env or simulation parameters to yaml file.

Note

Multiple .pb inputs allowed for setting model ensembles.

Parameters:: thing (str) – Path to file you want to set (.pb, .sh, .yaml, or .json file).

dptools.cli.train module

class dptools.cli.train.CLI(parser)

Bases: BaseCLI

Setup and submit jobs to train single or ensemble of deepmd-kit models.

Complete documentation here

Examples:

$ dptools train /path/to/dataset # simple single model
$ dptools train -e /path/to/dataset # ensemble (4) of models
$ dptools train -e -s /path/to/dataset # submit 4 slurm jobs to train ensemble
$ dptools train -p /path/to/training/dir /path/to/dataset # specify dir to train in
$ dptools train -i /path/to/in.json /path/to/dataset # specify in.json parameter file

add_args(): Command specific arguments

static get_hpc_info()

help_info = 'Setup and submit jobs to train deepmd-kit models'

link_dirs(in_json)

Find and add all system training and validation set folders found in dataset directory to in_json.

Parameters:: in_json (dict) – jsonnable dictionary with deepmd-kit training parameters.

main(args): Command specific main method

set_types(in_json)

Load type_map.json from dataset directory and add types to in_json.

Parameters:: in_json (dict) – jsonnable dictionary with deepmd-kit training parameters.

setup(): Loads training parameter json file, updates with relevant info (random seeds, training/validation dirs, type map), and writes to training dir.

submit_jobs(): Write and submit Slurm job(s) to train model or ensemble of models.

static write_json(src, dest)

Writes training parameter json file to specified training directory.

Parameters:

src (dict) – jsonnable dictionary with deepmd-kit training parameters.
dest (str) – Path to training directory to write .json file to