graph2lammps

This program enables the generation of LAMMPS input data based on networkx graphs.

Install Instructions

Installation with pip:

git clone git@github.com:sjiang87/graph2lammps.git
cd graph2lammps
pip install -e .

Quickstart

import numpy as np
import networkx as nx
from graph2lammps.convert import gen_pos, write_pdb, write_lammps

Create a polymer graph, e.g., a star polymer:

def gen_star(num_arm, len_branch):
    G = nx.Graph()
    G.add_node(0)

    for i in range(1, num_arm + 1):
        arm = nx.path_graph(len_branch)
        arm = nx.relabel_nodes(arm, {j: i*len_branch + j for j in range(len_branch)})
        G = nx.union(G, arm)
        G.add_edge(0, i*len_branch)
    return G

G = gen_star(num_arm=5, len_branch=5)
mapping = {old_label: new_label for new_label, old_label in enumerate(sorted(G.nodes()))}
G = nx.relabel_nodes(G, mapping)

Assign properties to each node or edge. This can be done after graph creation or directly in the custom 'gen_star' function.

for node in G.nodes():
    G.nodes[node]['type'] = random.randint(1, 2) # bead type 1 or 2
    G.nodes[node]['name'] = string.ascii_uppercase[G.nodes[node]['type'] ] # some random node names
    
for edge in G.edges():
    G.edges[edge]['type'] = random.randint(1, 2) # bond type 1 or 2

G.graph['mass'] = {1: 1, 2:2} # graph-level metadata, e.g., bead mass, charge

Create node position matrix:

pos = gen_pos(G, rmin=0.5, bdist=1.0, bmin=0.9, bmax=1.1, niter=1000)

rmin : float
    Minimum separation distance between polymer beads.
bdist : float
    Ideal bond distance between connected polymer beads.
bmin : float
    Minimum bond length.
bmax : float
    Maximum bond length.
niter : int
    Number of iterations for position generation.

Write LAMMPS data file.

write_lammps(G, pos, fname='your/dir/sys.data')

The above-mentioned procedure can be completed by:

from graph2lammps import graph2lammps

G = gen_star(num_arm=5, len_branch=5)
graph2lammps(G, rmin=0.5, bdist=1.0, bmin=0.9, bmax=1.1, niter=1000, fname='your/dir/sys.data)

TODO

Currently supports bead/atom and bond types. Support for angles, dihedrals, and other types will be added.
An executable command will be added to convert a pickled file containing multiple graphs to multiple LAMMPS data files.

Contribution

If you need to add additional functionality, please contact sj0161@princeton.edu or submit a pull request.