Version: 2026.03

pwdata Data Conversion

Overview

This page applies to pwdata >= 0.5.10.

pwdata is MatPL's data-preprocessing toolkit. It can:

Convert among atom.config (PWmat), POSCAR (VASP), lmp.init (LAMMPS), and cp2k.init (CP2K).
Build supercells, scale lattices, and perturb lattices or atomic positions.
Extract MOVEMENT (PWmat), OUTCAR (VASP), LAMMPS dump, and CP2K MD trajectories, or read pwmlff/npy, extxyz, deepmd/npy, deepmd/raw, and the Meta OMAT24 dataset, then write pwmlff/npy or extxyz. For Meta datasets, CPU-parallel queries make it practical to search databases containing over 100 million structures.

Supported Data Types

Software	file	multi-Image	label	format
PWmat	MOVEMENT	True	True	'pwmat/movement'
PWmat	OUT.MLMD	False	True	'pwmat/movement'
PWmat	atom.config	False	False	'pwmat/config'
VASP	OUTCAR	True	True	'vasp/outcar'
VASP	poscar	False	False	'vasp/poscar'
LAMMPS	lmp.init	False	False	'lammps/lmp'
LAMMPS	dump	True	False	'lammps/dump'
CP2K	stdout, xyz, pdb	True	True	'cp2k/md'
CP2K	stdout	False	True	'cp2k/scf'
MatPL	*.npy	True	True	'pwmlff/npy'
DeepMD (read)	.npy, .raw	True	True	'deepmd/npy', 'deepmd/raw'
* (extended xyz)	*.xyz	True	True	'extxyz'
Meta (read)	*aselmdb	True	True	'meta'

Installation

Install with pip:

pip install pwdata

# Install or upgrade pwdata
pip install pwdata --upgrade

# List available versions
pip index versions pwdata
# Example output:
# pwdata (0.5.10)
# Available versions: 0.5.10, 0.5.9, 0.5.8, 0.5.7, 0.5.6, 0.5.5, 0.5.4, 0.5.3, 0.4.7, 0.4.0, 0.3.2, 0.2.16

# Install a specific version
pip install pwdata==n.m.o

Command-Line Usage

pwdata supports both command-line use and Python API integration.

Command list: Use -h for descriptions, options, and examples. pwdata -h lists all commands.

pwdata -h
pwdata convert_config or cvt_config -h
pwdata convert_configs or cvt_configs -h
pwdata scale_cell or scale -h
pwdata super_cell or super -h
pwdata perturb -h
pwdata count -h

pwdata convert_config: Convert a Structure

Converts between structure formats. Use convert_config or cvt_config. Options:

pwdata convert_config [-h] -i INPUT -f INPUT_FORMAT -o OUTPUT_FORMAT [-s SAVENAME] [-c] [-t ATOM_TYPES [ATOM_TYPES ...]]

`-h`

Displays all options and their descriptions.

`-i`

Required input path, absolute or relative.

`-f`

Optional input format. If omitted, it is inferred. Supported values: pwmat/config, vasp/poscar, lammps/lmp, and cp2k/scf.

`-o`

Required output format: pwmat/config, vasp/poscar, or lammps/lmp.

`-s`

Output filename used with -o. Defaults are atom.config, POSCAR, and lammps.lmp for the respective formats.

`-c`

Use -c to write Cartesian coordinates; otherwise fractional coordinates are used. pwmat/config always uses fractional coordinates.

`-t`

For lammps/lmp or lammps/dump, specifies atom types as symbols or atomic numbers in input-file order. Ignored for other formats.

See the repository's examples:

# convert_config example: atom.config to POSCAR
# Creates examples/test_workdir/cvtcnf_atom.POSCAR
pwdata cvt_config -i examples/pwmat_data/LiGePS_atom.config -s examples/test_workdir/cvtcnf_atom.POSCAR -o vasp/poscar

pwdata convert_configs: Convert Training Data

Extracts supported trajectories and datasets—including MOVEMENT, OUTCAR, LAMMPS dump, CP2K MD, pwmlff/npy, extxyz, deepmd/npy, deepmd/raw, and Meta OMAT24—to pwmlff/npy or extxyz. Use convert_configs or cvt_configs.

Options:

pwdata [-h] -i INPUT [INPUT ...] [-f INPUT_FORMAT] [-s SAVEPATH] [-o OUTPUT_FORMAT] [-r]
              [-m MERGE] [-g GAP] [-q QUERY] [-n CPU_NUMS] [-t ATOM_TYPES [ATOM_TYPES ...]]

`-h`

Displays all options and descriptions.

`-i`

Required list of input files or directories, absolute or relative. For pwmlff/npy, extxyz, deepmd/npy, deepmd/raw, and Meta OMAT24, pwdata searches directories recursively, so only the data root is required.

For example, to read all dataset subdirectories under examples/pwmlff_data/LiSiC, use -i examples/pwmlff_data.

For many files or directories, such as a Meta database, list the paths in JSON and pass that file with -i.

{
    "datapath" : [
        "/share/public/PWMLFF_test_data/eqv2-models/datasets/decompress/Omat24/train/rattled-1000-subsampled",
        "/share/public/PWMLFF_test_data/eqv2-models/datasets/decompress/Omat24/train/rattled-300",
        "/share/public/PWMLFF_test_data/eqv2-models/datasets/decompress/Omat24/train/rattled-300-subsampled"
    ]
}

`-f`

Optional input format; inferred by default. Supported values are pwmat/movement, vasp/outcar, lammps/dump, cp2k/md, pwmlff/npy, deepmd/npy, deepmd/raw, extxyz, and meta.

`-o`

Optional output format: pwmlff/npy (default) or extxyz.

`-s`

Optional output directory; defaults to the current directory.

`-m`

Optional integer for extxyz output. -m 1 (default) writes all structures to one XYZ file; -m 0 separates files by element combination.

`-g`

Optional integer trajectory interval. The default is 1.

`-q`

Optional query string for meta input. Quote expressions such as H>4 so the shell does not interpret > as redirection. See the Meta query examples.

`-n`

Optional CPU count for parallel Meta queries. The default is one core.

`-t`

For LAMMPS input, specifies atom types as symbols or atomic numbers in input order.

For Meta input, selects structures containing only these element types.

pwdata convert_configs Examples

See the repository's examples:

Example 1. Extract every fifth frame from two PWmat trajectories to examples/test_workdir/3_1_configs_extxyz.

pwdata convert_configs -i examples/pwmat_data/50_LiGePS_movement examples/pwmat_data/lisi_50_movement -s examples/test_workdir/3_1_configs_extxyz -o extxyz -g 5

Example 2. Convert all deepmd/npy data under examples/deepmd_data/alloy to pwmlff/npy.

pwdata convert_configs -i examples/deepmd_data/alloy -s ./test_workdir/7_0_configs_PWdata

Example 3. Convert all XYZ files under examples/xyz_data to pwmlff/npy in examples/test_workdir/5_0_configs_PWdata.

pwdata convert_configs -i examples/xyz_data -s examples/test_workdir/5_0_configs_PWdata -g 1

Example 4. Query all .aselmdb databases under examples/meta_data/alex_val for structures containing only Pt and Ge, then save them under ./test_workdir/10_1_configs_extxyz.

pwdata convert_configs -i examples/meta_data/alex_val -s ./test_workdir/10_1_configs_extxyz -t Pt Ge

Example 5. Query the .aselmdb databases listed in meta_data['data_path'] for Pt–Ge structures and save them under ./test_workdir/10_1_configs_extxyz.

pwdata convert_configs -i examples/meta_data.json -s ./test_workdir/10_1_configs_extxyz -o extxyz -t Pt Ge

Meta Query Examples for convert_configs

Example 1. Query structures containing only Pt and Ge and write XYZ.

pwdata convert_configs -i examples/meta_data/alex_val/alex_go_aao_001.aselmdb examples/meta_data/alex_val/alex_go_aao_002.aselmdb -s examples/test_workdir/10_1_configs_extxyz -o extxyz -t Pt Ge

Example 2. Use -q to find all structures containing Cu.

pwdata convert_configs -i examples/meta_data/alex_val/alex_go_aao_001.aselmdb examples/meta_data/alex_val/alex_go_aao_002.aselmdb -s examples/test_workdir/10_1_configs_extxyz -o extxyz -q 'Cu'

Example 3. Find structures with fewer than three H atoms.

pwdata convert_configs -i examples/meta_data/alex_val/alex_go_aao_001.aselmdb examples/meta_data/alex_val/alex_go_aao_002.aselmdb -s examples/test_workdir/10_1_configs_extxyz -o extxyz -q 'H<3'

Example 4. Find structures containing Cu and fewer than three H atoms.

pwdata convert_configs -i examples/meta_data/alex_val/alex_go_aao_001.aselmdb examples/meta_data/alex_val/alex_go_aao_002.aselmdb -s examples/test_workdir/10_1_configs_extxyz -o extxyz -q 'Cu,H<3'

Example 5. Find structures with at least two H atoms and at least one O atom.

pwdata convert_configs -i examples/meta_data/alex_val/alex_go_aao_001.aselmdb examples/meta_data/alex_val/alex_go_aao_002.aselmdb -s examples/test_workdir/10_1_configs_extxyz -o extxyz -q 'H2O'

Additional query expressions:

| ---query string---| -------------------------Function Explanation---------------------------- |
| v3                | has ‘v3’ key                                                              |
| abc=bla-bla       | has key ‘abc’ with value ‘bla-bla’                                        |
| v3,abc=bla-bla    | both of the above                                                         |
| calculator=nwchem | calculations done with NWChem                                             |
| 2.2<bandgap<3.0   | ‘bandgap’ key has value between 2.2 and 3.0                               |
| natoms>=10        | 10 or more atoms                                                          |
| id=2345           | specific id                                                               |
| age<1h            | not older than 1 hour                                                     |
| age>1y            | older than 1 year                                                         |
| pbc=TTT           | Periodic boundary conditions along all three axes                         |
| pbc=TTF           | Periodic boundary conditions along the first two axes (F=False, T=True)   |
https://databases.fysik.dtu.dk/ase/ase/db/db.html#id7

tip

pwdata infers formats and searches dataset directories automatically. For pwmlff/npy, extxyz, deepmd/npy, deepmd/raw, and Meta OMAT24, specify only the data root.

If no format is specified, every supported data source under the input directory is included.

pwdata scale_cell: Scale a Lattice

Use scale_cell or scale. Options:

pwdata scale_cell [-h] -r SCALE_FACTOR [SCALE_FACTOR ...] -i INPUT -f INPUT_FORMAT [-s SAVENAME] [-o OUTPUT_FORMAT] [-c] [-t ATOM_TYPES [ATOM_TYPES ...]]

`-h`

Displays all options and descriptions.

`-r`

Required space-separated scale factors in (0,1), with $Lattice_{new}=factor\times Lattice_{old}$ . For example, -r 0.97 0.98 0.99 creates three scaled structures.

`-i`

Required input path, absolute or relative.

`-f`

Optional input format, inferred by default. Supported values: pwmat/config, vasp/poscar, lammps/lmp, and cp2k/scf.

`-o`

Optional output format: pwmat/config, vasp/poscar, or lammps/lmp. Defaults to the input format, except cp2k/scf defaults to pwmat/config.

`-s`

Optional base output filename used with -o; the scale factor is prefixed. Defaults are atom.config, POSCAR, and lammps.lmp. Example: -o pwmat/config -s atom.config -r 0.99 writes 0.99_atom.config.

`-c`

Use -c for Cartesian coordinates; otherwise fractional coordinates are used. PWmat always uses fractional coordinates.

`-t`

For LAMMPS input, specifies atom types as symbols or atomic numbers in input order.

See the repository's examples:

# scale_cell example
pwdata scale_cell -r 0.98 0.99 0.97 0.95 -i examples/pwmat_data/lisi_atom.config -f pwmat/config -s examples/test_workdir/scale_atom.config -o pwmat/config

# or the abbreviated command
pwdata scale -r 0.98 0.99 0.97 0.95 -i examples/pwmat_data/lisi_atom.config -f pwmat/config -s examples/test_workdir/scale_atom.config -o pwmat/config

# Creates four scaled files under examples/test_workdir

pwdata super_cell: Build a Supercell

Use super_cell or super. Options:

pwdata super_cell [-h] -m SUPERCELL_MATRIX [SUPERCELL_MATRIX ...] -i INPUT -f INPUT_FORMAT [-s SAVENAME] [-o OUTPUT_FORMAT] [-c] [-p PERIODICITY [PERIODICITY ...]] [-l TOLERANCE] [-t ATOM_TYPES [ATOM_TYPES ...]]

`-h`

Displays all options and descriptions.

`-m`

Required 3×3 supercell matrix, supplied as three or nine values. -m 2 0 0 0 2 0 0 0 2 and -m 2 2 2 both define a 2×2×2 supercell.

`-i`

Required input path.

`-f`

Optional input format, inferred by default: pwmat/config, vasp/poscar, lammps/lmp, or cp2k/scf.

`-o`

Optional output format: pwmat/config, vasp/poscar, or lammps/lmp. Defaults to input format, except CP2K defaults to PWmat.

`-s`

Optional output filename used with -o. Defaults are atom.config, POSCAR, and lammps.lmp. Example: -o pwmat/config -s super_atom.config.

`-c`

Use -c for Cartesian coordinates; otherwise fractional coordinates are used. PWmat always uses fractional coordinates.

`-t`

For LAMMPS input, specifies atom types in input order.

`-p`

Optional periodic boundaries. [1,1,1] (default) is periodic in x, y, and z.

`-l`

Optional fractional-coordinate tolerance. The default is 1e-5 and prevents tiny negative coordinates from being mapped incorrectly.

See the repository's examples:

# super_cell example
pwdata super_cell -m 2 3 4 -i examples/pwmat_data/lisi_atom.config -s examples/test_workdir/super_atom.config -o pwmat/config 

# or the abbreviated command
pwdata super -m 2 0 0 0 3 0 0 0 4 -i examples/pwmat_data/lisi_atom.config -s examples/test_workdir/super_atom.config -o pwmat/config 

# Creates examples/test_workdir/super_atom.config with a 2×3×4 supercell

pwdata perturb: Perturb a Lattice and Atomic Positions

Perturbs a structure's lattice or atomic positions. Options:

pwdata perturb [-h] [-d ATOM_PERT_DISTANCE] [-e CELL_PERT_FRACTION] -n PERT_NUM -i INPUT -f  INPUT_FORMAT [-s SAVENAME] [-o OUTPUT_FORMAT] [-c] [-t ATOM_TYPES [ATOM_TYPES ...]]

`-h`

Displays all options and descriptions.

`-d`

Optional atomic perturbation distance in ångströms. Each coordinate is shifted by a value sampled uniformly from [-atom_pert_distance, atom_pert_distance]. The default 0 disables atomic perturbation.

`-e`

Optional cell-deformation fraction. Each of the nine lattice values is perturbed by a uniformly sampled value in [-cell_pert_fraction, cell_pert_fraction]. For example, 0.03 represents approximately 3% deformation. The default 0 disables lattice perturbation.

`-n`

Required number of perturbed structures to generate.

`-i`

Required input path.

`-f`

Optional input format, inferred by default: pwmat/config, vasp/poscar, lammps/lmp, or cp2k/scf.

`-o`

Optional output format: pwmat/config, vasp/poscar, or lammps/lmp. Defaults to input format, except CP2K defaults to PWmat.

`-s`

Optional output filename used with -o. Defaults are atom.config, POSCAR, and lammps.lmp.

`-c`

Use -c for Cartesian coordinates; otherwise fractional coordinates are used. PWmat always uses fractional coordinates.

`-t`

For LAMMPS input, specifies atom types in input order.

See the repository's examples:

# perturb example
pwdata perturb -e 0.01 -d 0.04 -n 20 -i examples/pwmat_data/lisi_atom.config -f pwmat/config -s examples/test_workdir/perturb_atom -o pwmat/config
# Generates 20 perturbed structures under examples/test_workdir/pertub_atom

pwdata count: Count Structures

Counts structures in supported trajectories and datasets, including MOVEMENT, OUTCAR, LAMMPS dump, CP2K MD, pwmlff/npy, extxyz, deepmd/npy, deepmd/raw, and Meta OMAT24.

Options:

pwdata convert_configs [-h] [-h] -i INPUT [INPUT ...] [-f INPUT_FORMAT] [-q QUERY] [-n CPU_NUMS] [-t ATOM_TYPES [ATOM_TYPES ...]]

`-h`

Displays all options and descriptions.

`-i`

Required list of input files or directories. For pwmlff/npy, extxyz, deepmd/npy, deepmd/raw, and Meta OMAT24, specify only the data root; pwdata searches recursively.

For example, to count all datasets under examples/pwmlff_data/LiSiC, use -i examples/pwmlff_data.

For many inputs, list their paths in JSON and pass the file with -i.

{
    "datapath" : [
        "/share/public/PWMLFF_test_data/eqv2-models/datasets/decompress/Omat24/train/rattled-1000-subsampled",
        "/share/public/PWMLFF_test_data/eqv2-models/datasets/decompress/Omat24/train/rattled-300",
        "/share/public/PWMLFF_test_data/eqv2-models/datasets/decompress/Omat24/train/rattled-300-subsampled"
    ]
}

`-f`

Optional input format, inferred by default. Supported values are pwmat/movement, vasp/outcar, lammps/dump, cp2k/md, pwmlff/npy, deepmd/npy, deepmd/raw, extxyz, and meta.

`-q`

Optional query string for Meta input. See the Meta query examples.

`-n`

Optional CPU count for parallel Meta queries. The default is one core.

`-t`

For LAMMPS input, specifies atom types in input order.

For Meta input, selects structures containing only the specified element types.

Using pwdata as a Python Library

Use the pwdata API to generate datasets or convert data programmatically.

pwdata Config: Read Data

Config (self, format: str, data_path: str, pbc = None, atom_names = None, index = ':', **kwargs)

Parameters:

format: String. Input format: pwmat/config, vasp/poscar, lammps/dump, lammps/lmp, pwmat/movement, vasp/outcar, cp2k/md, cp2k/scf, pwmlff/npy, deepmd/npy, deepmd/raw, extxyz, or meta.
pwmat/config: PWmat structure, e.g. atom.config.
pwmat/movement: PWmat MD trajectory, e.g. MOVEMENT.
lammps/dump: LAMMPS dump, e.g. dump.lammptraj.
lammps/lmp: LAMMPS structure, e.g. in.lmp.
vasp/poscar: VASP structure, e.g. POSCAR.
vasp/outcar: VASP MD trajectory, e.g. OUTCAR.
cp2k/md: CP2K output plus positions and forces, e.g. cp2k.out, *pos-1.xyz, *pos-1.pdb, or *frac-1.xyz.
cp2k/scf: CP2K output, e.g. cp2k.out.
pwmlff/npy: MatPL dataset, e.g. energies.npy.
deepmd/npy: DeepMD dataset, e.g. force.npy.
deepmd/raw: DeepMD dataset, e.g. force.raw.
extxyz: Extended XYZ, e.g. *.xyz.
meta: Open Meta data with .aselmdb extension.

caution

Set PRINT_LEVEL MEDIUM in CP2K input so the standard output contains lattice information.

data_path: String, required. Input path.
pbc: Optional list of periodic boundaries. Default None; [1,1,1] is periodic in x, y, and z.
atom_names: Optional list of atom names for lammps/dump, e.g. ['C','H','O']. Default None.
index: Optional integer, slice, or string selecting structures from a multi-structure file. Default : reads all.
- index=0: first structure
- index=-2: second-to-last structure
- index=':' or slice(None): all structures
- index='-3:' or slice(-3,None): last three structures
- index='::2' or slice(0,None,2): even indices
- index='1::2' or slice(1,None,2): odd indices
kwargs: Optional keyword arguments for the reader.
unit: Optional LAMMPS unit style. Default metal.
style: Optional LAMMPS atom style. Default atomic. See LAMMPS atom_style.
sort_by_id: Optional Boolean controlling atom sorting by ID for LAMMPS. Default True.

Returns: A list of Image objects, one per structure.

Example:

from pwdata import Config

data_file = "./cp2k.out"
format = "cp2k/scf"
config = Config(format, data_file)

pwdata Config.to(): Write Data

Writes loaded data in the requested format.

Config.to (self, data_path, format = None, **kwargs)

Parameters:

data_path: String, required. Output path.
format: String, required. Output format: pwmat/config, pwmat/movement, vasp/poscar, lammps/lmp, extxyz, or pwmlff/npy.
Kwargs: Additional output keyword arguments:
For pwmat/config, vasp/poscar, lammps/lmp, and extxyz:
- data_name: String, required. Output filename.
- sort: Optional Boolean controlling sorting by atomic number. Default None.
- wrap: Optional Boolean controlling wrapping into the simulation cell. Default False.
- direct: Boolean, required. Selects fractional versus Cartesian coordinates.
For labeled pwmlff/npy output:
- data_name: String, required. Dataset-directory name.
- random: Optional Boolean controlling shuffling.
- seed: Optional random seed. Default 2024.

caution

For CP2K data, set sort=False; CP2K is already sorted and re-sorting corrupts ordering.
pwmlff/npy stores labeled datasets for model training.
PWmat structures always use fractional coordinates, so direct is ignored.
LAMMPS structures always use Cartesian coordinates, so direct is ignored.

Example 1:

from pwdata import Config

data_file = "./POSCAR"
format = "vasp/poscar"
config = Config(format, data_file)
config.to(data_path = "./", format = "lammps/lmp", data_name = "lmp.init", direct = False, sort = True)

Multiple configurations with the same structure can be combined with .append() before calling .to().

Example:

from pwdata import Config

raw_data = ["./OUTCAR0", "./OUTCAR1", "./OUTCAR2"]    # the same atoms...
format = "vasp/outcar"
multi_data = Config(format, raw_data[0])
for data in raw_data[1:]:
   image_data = Config(format, data)
   multi_data.append(image_data)
multi_data.to(data_path = "./PWdata", format='pwmlff/npy')

Example 2:

Convert pwmat/movement, vasp/outcar, cp2k/md, or lammps/dump trajectories to individual pwmat/config, vasp/poscar, or lammps/lmp structures.

from pwdata import Config
from pwdata.utils.constant import FORMAT

def trajs2config():
    atom_types = ["Hf", "O"] # for lammps
    input_file = "/data/home/wuxingxing/codespace/pwdata/examples/lmps_data/HfO2/30.lammpstrj"
    input_format="lammps/dump"
    save_format = "pwmat/config"
    image = Config(data_path=input_file, format=input_format, atom_names=atom_types)
    tmp_image_data = image.images
    save_dir = "./tmp_test"
    for id, config in enumerate(tmp_image_data):
        savename = "{}_{}".format(id, FORMAT.get_filename_by_format(save_format))
        image.iamges = [config]
        image.to(data_path = save_dir,
            data_name = savename,
            format = save_format,
            sort = True)

if __name__=="__main__":
    trajs2config()

pwdata build.supercells: Supercell Construction

build.supercells.make_supercell (image_data, supercell_matrix: list, pbc: list = None, wrap=True, tol=1e-5)

Construct a supercell from the input structure and supercell matrix.

Parameters:

image_data: Required. An Image object containing information about the original structure.
supercell_matrix: List, required. The 3x3 supercell matrix. For example, [[2, 0, 0], [0, 2, 0], [0, 0, 2]] creates a 2x2x2 supercell.
pbc: List, optional. Periodic boundary conditions. The default is None. For example, [1, 1, 1] makes the system periodic along the x, y, and z directions.
wrap: Boolean, optional. Whether to wrap atoms into the simulation cell when periodic boundary conditions are used. The default is True.
tol: Float, optional. Tolerance for fractional coordinates. The default is 1e-5. It prevents slightly negative coordinates from being wrapped into the simulation cell.

Returns: A new Image object (or list of Image objects) containing information about the supercell.

Example:

from pwdata import Config, make_supercell

data_file = "./atom.config"
config = Config('pwmat/config', data_file)
supercell_matrix = [[2, 0, 0], [0, 2, 0], [0, 0, 2]]
supercell = make_supercell(config, supercell_matrix, pbc=[1, 1, 1])
supercell.to(data_path = "./", data_name = "atom_2x2x2.config", format = "pwmat/config", sort = True)

pwdata pertub.perturbation: Structure Perturbation

Perturb a structure.

pertub.perturbation.perturb_structure (image_data, pert_num:int, cell_pert_fraction:float, atom_pert_distance:float)

Parameters:

image_data: Required. The Image object to perturb, containing information about the original structure.
pert_num: Integer, required. Number of perturbed structures to generate.
cell_pert_fraction: Float, required. Magnitude of the cell deformation. For example, 0.03 means that the cell is deformed by 3% relative to the original cell.
atom_pert_distance: Float, required. Atomic perturbation distance, in angstroms, that controls how far atoms move from their original positions. For example, 0.01 means that atoms are displaced by 0.01 angstrom.

Returns: A list of new Image objects, each containing information about a perturbed structure.

Example:

from pwdata import Config, perturb_structure

data_file = "./atom.config"
config = Config('pwmat/config', data_file)
pert_num = 50
cell_pert_fraction = 0.03
atom_pert_distance = 0.01
save_format = "pwmat/config"
perturbed_structs = perturb_structure(config, pert_num, cell_pert_fraction, atom_pert_distance)
perturbed_structs.to(data_path = "~/pwdata/test/pertubed/",
        data_name = "pertubed",
        format = save_format,
        direct = True,
        sort = True)

pwdata pertub.scale: Cell Scaling

pertub.scale.scale_cell (image_data, scale_factor:float) 

Parameters:

image_data: Required. The Image object to scale, containing information about the original structure.
scale_factor: Float, required. Scaling factor for the cell.

Returns: A new Image object (or list of Image objects), each containing information about a scaled structure.

Example:

from pwdata import Config, scale_cell

data_file = "./atom.config"
config = Config('pwmat/config', data_file)
scale_factor = 0.95
scaled_structs = scale_cell(config, scale_factor)
scaled_structs.to(data_path = "~/test/scaled/",
        data_name = "scaled",
        format = "pwmat/config",
        direct = True,
        sort = True)

pwdata API Example: Converting MPtraj Files to LMDB Format

The open-source MPtraj dataset is provided in JSON format. After converting it to LMDB, use the cvt_configs command with the -t and -q options to quickly locate selected structures and convert them to extxyz or pwmlff/npy format for training.

import json
from pwdata.fairchem.datasets.ase_datasets import LMDBDatabase
from ase import Atoms
from ase.db.row import AtomsRow
from pwdata.utils.constant import get_atomic_number_from_name
from tqdm import tqdm
import numpy as np

def MPjson2lmdb():
    mp_file = "/data/home/wuxingxing/codespace/pwdata/examples/mp_data/mptest.json"
    save_file = "/data/home/wuxingxing/codespace/pwdata/examples/mp_data/sub.aselmdb"
    Mpjson = json.load(open(mp_file))
    db = LMDBDatabase(filename=save_file, readonly=False)
    for key_1, val_1 in tqdm(Mpjson.items(), total=len(Mpjson.keys())):
        for key_2, val_2 in val_1.items():
            _atomrow, data = cvt_dict_2_atomrow(val_2)
            db._write(_atomrow, key_value_pairs={}, data=data)
    db.close()
    
def cvt_dict_2_atomrow(config:dict):
    cell = read_from_dict('matrix', config['structure']['lattice'], require=True)
    atom_type_list = get_atomic_number_from_name([_['label'] for _ in config['structure']['sites']])
    position = [_['xyz'] for _ in config['structure']['sites']]
    magmom = read_from_dict('magmom', config, require=True)
    atom = Atoms(positions=position,
                numbers=atom_type_list,
                magmoms=magmom,
                cell=cell)

    atom_rows = AtomsRow(atom)
    atom_rows.pbc = np.ones(3, bool)
    # read stress -> xx, yy, zz, yz, xz, xy
    stress = read_from_dict('stress', config, require=True)
    atom_rows.stress = [stress[0][0],stress[1][1],stress[2][2],stress[1][2],stress[0][2],stress[0][1]]
    force = read_from_dict('force', config, require=True)
    energy = read_from_dict('corrected_total_energy', config, require=True)
    atom_rows.__setattr__('force',  force)
    atom_rows.__setattr__('energy', energy)
    data = {}
    data['uncorrected_total_energy'] = read_from_dict('uncorrected_total_energy', config, default=None)
    data['corrected_total_energy'] = read_from_dict('uncorrected_total_energy', config, default=None)
    data['energy_per_atom'] = read_from_dict('energy_per_atom', config, default=None)
    data['ef_per_atom'] = read_from_dict('ef_per_atom', config, default=None)
    data['e_per_atom_relaxed'] = read_from_dict('e_per_atom_relaxed', config, default=None)
    data['ef_per_atom_relaxed'] = read_from_dict('ef_per_atom_relaxed', config, default=None)
    data['bandgap'] = read_from_dict('bandgap', config, default=None)
    data['mp_id'] = read_from_dict('mp_id', config, default=None)
    return atom_rows, data

def read_from_dict(key:str, config:dict, default=None, require=False):
    if key in config:
        return config[key]
    else:
        if require:
            raise ValueError("key {} not found in config".format(key))
        else:
            return default
if __name__=="__main__":
    MPjson2lmdb()

pwdata Data Conversion

Overview​

Supported Data Types​

Installation​

Command-Line Usage​

pwdata convert_config: Convert a Structure​

-h​

-i​

-f​

-o​

-s​

-c​

-t​

pwdata convert_configs: Convert Training Data​

-h​

-i​

-f​

-o​

-s​

-m​

-g​

-q​

-n​

-t​

pwdata convert_configs Examples​

Meta Query Examples for convert_configs​

pwdata scale_cell: Scale a Lattice​

-h​

-r​

-i​

-f​

-o​

-s​

-c​

-t​

pwdata super_cell: Build a Supercell​

-h​

-m​

-i​

-f​

-o​

-s​

-c​

-t​

-p​

-l​

pwdata perturb: Perturb a Lattice and Atomic Positions​

-h​

-d​

-e​

-n​

-i​

-f​

-o​

-s​

-c​

-t​

pwdata count: Count Structures​

-h​

-i​

-f​

-q​

-n​

-t​

Using pwdata as a Python Library​

pwdata Config: Read Data​

pwdata Config.to(): Write Data​

pwdata build.supercells: Supercell Construction​

pwdata pertub.perturbation: Structure Perturbation​

pwdata pertub.scale: Cell Scaling​

pwdata API Example: Converting MPtraj Files to LMDB Format​

Overview

Supported Data Types

Installation

Command-Line Usage

pwdata convert_config: Convert a Structure

`-h`

`-i`

`-f`

`-o`

`-s`

`-c`

`-t`

pwdata convert_configs: Convert Training Data

`-h`

`-i`

`-f`

`-o`

`-s`

`-m`

`-g`

`-q`

`-n`

`-t`

pwdata convert_configs Examples

Meta Query Examples for convert_configs

pwdata scale_cell: Scale a Lattice

`-h`

`-r`

`-i`

`-f`

`-o`

`-s`

`-c`

`-t`

pwdata super_cell: Build a Supercell

`-h`

`-m`

`-i`

`-f`

`-o`

`-s`

`-c`

`-t`

`-p`

`-l`

pwdata perturb: Perturb a Lattice and Atomic Positions

`-h`

`-d`

`-e`

`-n`

`-i`

`-f`

`-o`

`-s`

`-c`

`-t`

pwdata count: Count Structures

`-h`

`-i`

`-f`

`-q`

`-n`

`-t`

Using pwdata as a Python Library

pwdata Config: Read Data

pwdata Config.to(): Write Data

pwdata build.supercells: Supercell Construction

pwdata pertub.perturbation: Structure Perturbation

pwdata pertub.scale: Cell Scaling

pwdata API Example: Converting MPtraj Files to LMDB Format