lammps-ast 0.1.91

A LAMMPS script parser and sanitizer using Lark

LAMMPS-AST

LAMMPS-AST is a Python package for sanitizing and parsing LAMMPS input scripts into abstract syntax trees (ASTs). It is built on Lark and is intended for structural analysis, validation, comparison, and downstream workflows around LAMMPS input files.

What It Provides

• script sanitization before parsing
• parsing of LAMMPS input scripts into ASTs
• AST transformation and comparison utilities
• repository examples showing how the parser can be used in notebook and pipeline workflows

Install

Install from PyPI:

pip install lammps_ast

If you need the optional visualization tooling used in some example workflows, you may also want a local Graphviz install.

Minimal Usage

from lammps_ast.sanitizer import sanitize
from lammps_ast.parser import parse_to_AST

script = """
units metal
atom_style atomic
boundary p p p
"""

sanitized = sanitize(script)
tree, errors = parse_to_AST(sanitized, lint=True)

parse_to_AST(..., lint=True) returns a parse tree plus collected parse errors. With lint=False, it behaves like a direct parser call and returns either a tree or an exception object.

Repository Layout

• lammps_ast/: package source, including parser, sanitizer, grammar, and AST utilities
• examples/: small examples of using the parser directly
• publication/: notebook-based workflow used for the publication-oriented evaluation example
• ez-pipeline/: script-oriented evaluation pipeline built on top of lammps_ast

The PyPI distribution is focused on the lammps_ast package itself. The notebook and pipeline folders are repository examples and supporting workflows.

Development Install

To work from a local clone:

pip install -e .

Citation

If you use LAMMPS-AST or the evaluation workflow in academic work, please cite the associated publication.

@article{lammps_ast_paper,
  title        = {Evaluating LLM-generated code for domain-specific languages: molecular dynamics with LAMMPS},
  author       = {Holbrook, Ethan W. and Verduzco, Juan C. and Strachan, Alejandro},
  year         = {2026},
  eprint       = {2603.20630},
  archivePrefix = {arXiv},
  primaryClass = {cs.SE},
  url          = {https://arxiv.org/abs/2603.20630}
}