molpy

A programmable toolkit for molecular simulation workflows

Documentation  ·  Quick start  ·  Examples  ·  Ecosystem

MolPy is a Python toolkit for the full molecular-system workflow — parsing, building, editing, typing, analyzing, and reading/writing simulation formats. Packing a box is the job of the companion package molpack.

Under active development. Public APIs may change between minor releases.

Vision

Molecular modeling is fragmented. Every simulation code has its own file formats and conventions; every task — parsing, building, typing, analysis, visualization — lives in a separate library; and moving a system between them means writing throwaway glue.

molpy aims to be the common foundation beneath that workflow: one explicit, programmable representation of a molecular system that every stage can share. Parse a structure into it, build on it, type and analyze it — then hand the same object onward, with no conversion step in between.

That representation is meant to be built on, not just used. It is the data model the MolCrafts ecosystem extends — visualization, experiment management, agent access — and it reads the same whether a human writes it or an agent calls it.

Capabilities

Each row is one src/molpy/ module — parse or build a structure, edit and type it, analyze or minimize it, then read and write it across formats.

Module	Capability
core	Explicit data model — editable Atomistic topology graph, Frame/Block columnar arrays, ForceField, Box
parser	SMILES / SMARTS (SmilesIR, SmartsPattern)
builder	System assembly — polymer planning and polydispersity, nanostructures, crystals, virtual sites; site-graph assembly (mp.Assembler) is native, on the molpy root
conformer	3D coordinate generation (native ETKDG + MMFF cleanup)
typifier	Force-field typing — OPLS-AA and MMFF94 (native), GAFF / GAFF2 via AmberTools
potential · optimize	Energy & force potentials with L-BFGS minimization
compute	Analysis under molpy.compute — rdf/msd/pmsd/jacf/order/voronoi/… modules, plus dielectric and vibrational-spectrum classes on the package itself (native kernels)
io	Read/write — PDB, GRO, LAMMPS data, XYZ, force fields, trajectories, …
engine	MD input generation & run management — LAMMPS, CP2K, OpenMM
wrapper · adapter	External CLIs (Antechamber, tleap, …) and optional RDKit in-memory bridge

Install

pip install molcrafts-molpy

# Browser (Pyodide) — needs molcrafts-molrs Pyodide wheel on PyPI:
# await micropip.install("molcrafts-molrs")
# await micropip.install("molcrafts-molpy")

Core dependencies: NumPy and molrs (molcrafts-molrs>=0.15.0,<0.16) plus the MolCrafts logging/config packages. Optional: RDKit (adapter example), AmberTools (GAFF charges).

Until molrs 0.15.0 is published. molpy 0.15 needs molrs 0.15, which is not on PyPI yet, so pip install cannot resolve it. Install from source instead: clone molrs next to molpy (the two checkouts side by side in one directory) and run uv sync in molpy (see Install from source (development) below). uv builds molrs from the sibling checkout named in [tool.uv.sources]; pip ignores that path source, so it does not work for this step.

Nightly builds. Bleeding-edge snapshots are published to the separate project molcrafts-molpy-nightly (versioned X.Y.Z.devN) on every push to the nightly branch. Install with pip install --pre molcrafts-molpy-nightly. It imports as molpy, so it cannot be installed alongside the stable molcrafts-molpy (same as tensorflow vs tf-nightly).

Install from source (development)

git clone https://github.com/MolCrafts/molrs.git   # sibling checkout, see below
git clone https://github.com/MolCrafts/molpy.git
cd molpy
uv sync --extra dev
pre-commit install --hook-type pre-commit --hook-type pre-push
# the two gates (same as the hooks / CI):
uv run --no-project --with 'tox>=4.23' --with ruff==0.16.1 --with ty==0.0.65 tox -e lint
uv run --extra dev python -m pytest tests/ -n auto

[tool.uv.sources] points molcrafts-molrs at the sibling checkout ../molrs/molrs-python, so uv sync builds the Rust core with your toolchain (rustup). After editing molrs, rebuild what uv installed:

uv sync --extra dev --reinstall-package molcrafts-molrs

See docs/developer/development-setup for the full workflow.

Quick start

Parse a SMILES string, assign OPLS-AA types, and write LAMMPS input files:

import molpy as mp

from pathlib import Path

mol       = mp.SmilesIR("CCO").to_atomistic()     # ethanol from SMILES
mol3d, _  = mp.Conformer(seed=42).generate(mol)   # hydrogens + 3D coordinates

typifier  = mp.typifier.OPLSAATypifier()          # carries the OPLS-AA library
typed     = typifier.typify(mol3d)
ff        = typifier.forcefield()                 # parameters of the assigned types

system    = typed.to_frame()
system.box = mp.Box.cube(30.0)
system["atoms"]["mol_id"] = mp.Topology.from_frame(system).connected_components() + 1

# the pair cutoff is a run setting: you declare it, molpy never invents one
ff.get_style("pair", "lj/cut")["cutoff"] = 10.0
ff.get_style("pair", "coul/cut")["cutoff"] = 10.0

out = Path("output"); out.mkdir(exist_ok=True)
mp.io.write_lammps_data(out / "system.data", system)
mp.io.write_lammps_forcefield(out / "system.ff", ff, system)

More workflows — packed solvent boxes, virtual-site models, polymer chains and networks (the stress test for MolPy's editing machinery), AmberTools parameterization — are in the Example Gallery and the task-oriented Guides.

Documentation

Full documentation, including executable notebooks: docs.molcrafts.org/molpy

• Getting Started — install and first example
• Example Gallery — short copy-paste workflows
• Guides — task-oriented notebooks
• Concepts — data model deep dives
• API Reference — full API

MolCrafts ecosystem

Project	Role
molpy	Python toolkit — the shared molecular data model & workflow layer — this repo
molrs	Rust core — molecular data structures & compute kernels (native + WASM)
molpack	Packmol-grade molecular packing (Rust + Python)
molvis	WebGL molecular visualization & editing
molexp	Workflow & experiment-management platform
molnex	Molecular machine-learning framework
molq	Unified job queue — local / SLURM / PBS / LSF
molcfg	Layered configuration library
mollog	Structured logging, stdlib-compatible
molhub	Molecular dataset hub
molmcp	MCP server for the ecosystem
molrec	Atomistic record specification

Contributing

Issues and pull requests are welcome — see Contributing.

License

BSD-3-Clause — see LICENSE.

---

_{Crafted with 💚 by MolCrafts}