Overview

pyiron is an integrated development environment for computational materials science combining atomic structure objects, simulation codes, feedback loops, data management, visualization, and job management in a Jupyter-based platform. Developed at Max Planck Institut für Eisenforschung and ICAMS.

Installation

Conda (Recommended)

conda install -c conda-forge pyiron

pip

pip install pyiron

Quick Start

Initialize Project

from pyiron import Project

pr = Project("my_project")
pr.create_job_class()

LAMMPS Molecular Dynamics

from pyiron import Project

pr = Project("md_simulation")
job = pr.create_job("Lammps", "my_md_job")
job.structure = pr.create_structure("Al", "fcc", 4.05)
job.calc_md(temperature=300, n_steps=1000)
job.run()

VASP DFT Calculation

from pyiron import Project

pr = Project("dft_calc")
job = pr.create_job("Vasp", "my_vasp_job")
job.structure = pr.create_structure("Cu", "fcc", 3.6)
job.calc_static()
job.run()

When to Use This Skill

• Building complex multi-step simulation workflows
• Managing large-scale high-throughput calculations
• Running DFT, MD, or atomistic simulations interactively
• Developing ab initio thermodynamics protocols
• Performing equation of state calculations
• Managing job databases with SQL/HDF5 storage
• Creating reusable simulation templates
• Running LAMMPS, VASP, GPAW, or Quantum ESPRESSO jobs
• Interactive materials science in Jupyter notebooks
• Prototyping and debugging simulation protocols

Key Features

Structure Handling

structure = pr.create_structure("Al", "fcc", 4.0)
structure.plot3d()  # NGLview 3D visualization

Job Management

• Object-oriented job creation
• SQL database for job tracking
• HDF5 hierarchical data storage
• Job restart and re-run capabilities
• Automatic job status tracking

Supported Codes

Code	Type	Job Class
LAMMPS	Molecular dynamics	Lammps
VASP	DFT	Vasp
GPAW	DFT	GPAW
Quantum ESPRESSO	DFT	Espresso
SPHInX	DFT	Sphinx
Murnaghan	Equation of state	Murnaghan
Phonopy	Phonons	Phonopy

Workflow Patterns

# Master-worker pattern for high-throughput
for i in range(10):
    job = pr.create_job("Lammps", f"job_{i}")
    job.run()

Workflow Examples

Ab Initio Thermodynamics

pr = Project("thermo")
job = pr.create_job("Vasp", "bulk")
job.run()

surface = pr.create_job("Vasp", "surface")
surface.structure = bulk.structure.create_surface()
surface.run()

Equation of State

from pyiron import Project

pr = Project("eos")
job = pr.create_job("Murnaghan", "eos_job")
job.structure = pr.create_structure("Al", "fcc", 4.05)
job.run()
job.plot()  # Display E(V) curve

High-Throughput Screening

from pyiron import Project

pr = Project("screening")
structures = ["Al", "Cu", "Ni", "Fe"]
for elem in structures:
    job = pr.create_job("Vasp", elem)
    job.structure = pr.create_structure(elem, "fcc")
    job.run()

Phonon Calculation

from pyiron import Project

pr = Project("phonons")
job = pr.create_job("Phonopy", "phonon_job")
job.structure = pr.create_structure("Al", "fcc", 4.05)
job.run()
job.plot_phonons()

Molecular Dynamics with Analysis

from pyiron import Project

pr = Project("md_analysis")
job = pr.create_job("Lammps", "md_job")
job.structure = pr.create_structure("Cu", "fcc", 3.6)
job.calc_md(temperature=300, n_steps=10000, save_rate=100)
job.run()

# Analyze results
job.output.mean_squared_displacement()
job.output.energy_pot.plot()

Job Restart and Reuse

from pyiron import Project

pr = Project("restart_demo")
job = pr.create_job("Lammps", "reusable_job")
job.structure = pr.create_structure("Al", "fcc", 4.05)
job.run()

# Later: reload job
reloaded = pr.load("reusable_job")
print(reloaded.output.energy_pot)

# Create similar job from existing
new_job = pr.create_job("Lammps", "new_job")
new_job.input = reloaded.input
new_job.run()

Database Queries

from pyiron import Project

pr = Project("database_demo")
# List all jobs
pr.list_jobs()

# Filter jobs by status
pr.job_table(status="finished")

# Query specific properties
pr.job_table(filter={"hamilton": "Vasp"})

Best Practices

• Use Project to organize calculations hierarchically
• Name jobs descriptively for easy retrieval
• Use job.run() for execution, job.status for monitoring
• Store results in HDF5 format for persistence
• Use pr.pack() and pr.unpack() for archiving projects
• Enable job.server.cores for parallel execution
• Use pr.create_group() for logical job grouping
• Check job.status before accessing outputs
• Use job.interactive_open() for large output datasets

# Project organization
pr = Project("study")
pr_group = pr.create_group("high_temp")
pr_group.create_job("Lammps", "md_500K")
pr_group.create_job("Lammps", "md_600K")

# Archive project
pr.pack("study_archive.h5")

Troubleshooting

Common Issues

Issue	Cause	Solution
VASP executable not found	PATH not configured	Set VASP_COMMAND or configure .pyiron
Job hangs	Queue system issue	Check job.status and queue manager
HDF5 file errors	Corrupted storage	Use pr.job_table() to identify issues
Memory issues	Large structures	Use chunked processing
Import errors	Missing optional dependency	Install with extras: pip install pyiron[lammps]

Configuration

Create ~/.pyiron configuration:

# .pyiron file
[DEFAULT]
PROJECT_PATHS = ~/pyiron_projects

[VASP]
VASP_COMMAND = mpirun -np 4 vasp_std

[LAMMPS]
LAMMPS_COMMAND = lmp_mpi

Debugging Jobs

# Check job status
print(job.status)  # initialized, running, finished, aborted

# View job log
job.view_log()

# Check for errors
if job.status == "aborted":
    print(job.error_message)

# Access raw output
job["output/energy_tot"]

Visualization

structure.plot3d()              # 3D structure view
job.output.plot_energy()        # Energy plots
job.plot()                      # Job-specific plots
job.animate_structures()        # Animation for MD

Metadata

Property	Value
License	BSD-3-Clause
Language	Python
Developers	Max Planck Institute, ICAMS
Dependencies	ASE, numpy, pandas, h5py
Storage	SQL + HDF5

Resources

• GitHub: https://github.com/pyiron/pyiron
• Homepage: https://pyiron.org
• Documentation: https://pyiron.readthedocs.io
• Paper: Janssen et al., Comput. Mater. Sci. 163, 24-36 (2019)