---
id: dpdata
version: "1.0.2"
license: LGPL-3.0-or-later
license_treatment: copyleft
maintenance: active
---

# dpdata — Manipulating data formats of DeePMD-kit, VASP, QE, PWmat, and LAMMPS, etc.

License: copyleft · Maintenance: active · Downloads: 128.2K/mo

## What it is and what it does

dpdata is a Python toolkit for reading, writing, and converting atomistic simulation data between the formats used by different computational chemistry and machine learning software. It acts as a bridge between tools like VASP (quantum chemistry), LAMMPS (molecular dynamics), DeePMD-kit (machine learning potentials), and others, letting you prepare training data or share results across different simulation packages without manual file translation.

The package provides both command-line tools for quick format conversion and Python APIs for programmatic data manipulation. It depends on numpy, scipy, h5py for numerical work, and uses monty, wcmatch, lmdb, and msgpack-numpy for file handling and data serialization. It is actively maintained, supports Python 3.10 and later, and has no known security vulnerabilities.

Use it for:

- Convert VASP output (OUTCAR) to DeePMD-kit format for training machine learning interatomic potentials
- Prepare molecular dynamics trajectories from LAMMPS or GROMACS for analysis or machine learning
- Parse quantum chemistry calculations from Gaussian or ABACUS and reformat for downstream tools
- Batch convert atomistic datasets between formats in a computational workflow or data pipeline
- Extract and standardize atomic coordinates and energies from mixed simulation software outputs

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

dpdata converts and manipulates atomistic simulation data between formats used by computational chemistry and machine learning packages like VASP, LAMMPS, DeePMD-kit, and others.

Yes. dpdata is actively maintained, has low install friction, no known vulnerabilities, and fills a clear niche in computational chemistry workflows. The copyleft license is standard for research software. Install if you work with multiple atomistic simulation packages and need to convert or prepare data between them.

## Install

pip install dpdata
uv add dpdata
poetry add dpdata

## Installing dpdata

Before you install: Low friction: pure Python wheel, active maintenance (last commit 2026-08-14), and only seven runtime dependencies including standard scientific libraries (numpy, scipy, h5py). Requires Python 3.10 or later.

License in practice: LGPL-3.0-or-later (copyleft): you may use and modify dpdata freely, but any derivative work must remain under LGPL and source must be made available. Acceptable for most research and open-source projects; check your organization's policy if bundling into proprietary software.

Quickstart:

pip install dpdata

import dpdata

# Convert VASP OUTCAR to DeePMD-kit format
system = dpdata.System('OUTCAR', fmt='vasp/outcar')
system.to('deepmd/npy', 'deepmd_data')

Requires Python 3.10 or later. Some runtime dependencies (h5py, lmdb, msgpack-numpy) may require compilation on systems without pre-built wheels; conda installation is recommended if pip fails.

Verify before relying:

- Whether msgpack-numpy is a pure-Python dependency or requires compilation on your platform
- Performance characteristics when handling large atomistic datasets (gigabytes or larger)
- Whether all supported formats (VASP, LAMMPS, Gaussian, ABACUS, etc.) are equally mature or if some are experimental

## Package facts

- License: LGPL-3.0-or-later (copyleft)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 128.2K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

atomistic data format conversion, vasp lammps data conversion, deepmd-kit data preparation, molecular dynamics data manipulation, quantum chemistry output parsing, computational chemistry data tools, atomic structure file format conversion, computational-chemistry, data-conversion, machine-learning-data

[View on SkillFed](https://skillfed.io/packages/dpdata) · [View on PyPI](https://pypi.org/project/dpdata/)