dpdata

library1.0.1pypypi✓ verified 86d ago

A Python library for manipulating data formats of DeePMD-kit, VASP, QE, PWmat, LAMMPS, ABACUS, and other computational chemistry/DFT codes. It provides a unified interface to read, write, convert, and process atomic simulation data including coordinates, forces, energies, and stress tensors. Current version: 1.0.1, with a stable API after the 1.0.0 release. Release cadence: approximately 1-2 months per minor/patch version.

ai-ml data

pip install dpdata

github ↗package ↗homepage ↗docs ↗

INSTALL

IMPORT

SIG · DPDATA

dpdata

ai-mlpythonv1.0.1

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

System

✓ from dpdata import System

✗ from dpdata.system import System

dpdata previously allowed deep imports, but requires top-level import since v0.2.20

MultiSystems

✓ from dpdata import MultiSystems

✗ from dpdata.multisystems import MultiSystems

similar to System, top-level import is recommended

Read a VASP OUTCAR file, print number of atoms and first frame energy, convert to ASE Atoms.

from dpdata import System import numpy as np sys = System('path/to/OUTCAR', fmt='vasp/outcar') print('Number of atoms per frame:', sys['atom_numbs']) print('First frame energies:', sys['energies'][0]) # Convert ASE Atoms back atoms = sys.to_ase_atoms() print(atoms)

Debug

Known issues

gotchadpdata uses 0-based atomic type indices, while some formats (e.g., LAMMPS) use 1-based. Always check atom_type_map after loading.

fix

Access sys.atom_type_map to get mapping from index to species name.

affects: all

gotchaWhen converting to deepmd/npy, the 'set_size' during writing can produce multiple subdirectories (set.001, set.002, ...). Ensure consistency in number of frames per set if you intend to use multiple sets for training.

fix

Specify set_size in the DeepmdNativeFormat or use to_deepmd_npy(set_size=...) to control splitting.

affects: >=1.0.0

breakingPython 3.7 support was dropped in v0.2.22. Python 3.8+ is required.

fix

Upgrade Python to 3.8 or later.

affects: >=0.2.22

deprecatedThe 'label' kwarg in to_deepmd_npy is deprecated in favor of 'set_size'.

fix

Use set_size instead of label.

affects: >=1.0.0

gotchaWriting to LAMMPS data file may lose information if the System contains multiple atom types with different masses; you must manually set mass_map.

fix

Before writing, assign sys.data['atom_mass'] = [...] or use sys.set_masses(mass_map).

affects: all

Errors

Common errors & fixes

ModuleNotFoundError: No module named 'dpdata'

dpdata is not installed.

fix

Run: pip install dpdata

AttributeError: module 'dpdata' has no attribute 'System'

Importing dpdata but not using submodule; may have old version or incorrect import.

fix

Use: from dpdata import System

ValueError: Unknown format: 'outcar'

Format string mismatch: dpdata expects 'vasp/outcar' not just 'outcar'.

fix

Use fmt='vasp/outcar' when reading VASP OUTCAR files.

Upgrade

Version history

1.0.1latest on PyPI · released Feb 28, 2026

Audit

Dependencies

numpyrequiredcore array operations

aseoptionalASE Atoms and Trajectory support, converting dpdata System to/from ASE

Agent activity

13 hits · last 30 days

node

OpenAI (training)

Resources

githubhttps://github.com/deepmodeling/dpdata ↗

packagehttps://pypi.org/project/dpdata/ ↗

homepagehttps://dp.tech ↗

docshttps://docs.deepmodeling.com/projects/dpdata ↗