aiida-dlpoly

An AiiDA plugin for the DL_POLY molecular/particle dynamics software.

Installation

This plugin is available through PyPI and should be installed using pip or an equivalent python package manager.

pip install aiida-dlpoly

Requirements

To use this plugin a configured AiiDA profile and computer instance are required, see the AiiDA documentation for instructions on how to install and configure AiiDA and AiiDA code setup for how to configure AiiDA to link to external software.

Setup

To configure the DL_POLY plugin, an AiiDA code instance needs to be configured for the DL_POLY executable. The following is an example of a basic YAML configuration file:

label: dlpoly
description: DL_POLY 
computer: localhost 
filepath_executable: /absolute/path/to/DLPOLY.Z
default_calc_job_plugin: dlpoly
use_double_quotes: false 
with_mpi: true 
prepend_text: '' 
append_text: '' 

Write this to a file named dlpoly_config.yml ensuring the value for computer matches the label of your configured computer instance and the filepath_executable entry is the absolute path to a compiled DL_POLY executable. The code can then be configured by running:

verdi code create core.code.installed --config dlpoly_config.yml -n 

If successful this will have created a code with the label dlpoly which can then be used to run DL_POLY jobs within the AiiDA workflow.

Examples

An example script for running an simple molecular dynamics simulation with existing CONTROL, FIELD and CONFIG input files is given below.

from aiida.engine import run
from aiida.orm import SinglefileData, load_code
from aiida import load_profile

load_profile("username") # Replace with your own AiiDA username

builder = load_code("dlpoly").get_builder()
builder.control = SinglefileData(file="/absolute/path/to/CONTROL") # Change these to the absolute paths to input files
builder.field = SinglefileData(file="/absolute/path/to/FIELD")
builder.configuration = SinglefileData(file="/absolute/path/to/CONFIG")

results, node = run.get_node(builder)

By default the calculation will produce three output nodes, a SinglefileData node containing the generated OUTPUT file from DL_POLY, an ArrayData node containing the calculated statistics from the simulation and another SinglefileData node containing the DL_POLY formatted REVCON file. Additionally the following optional output nodes are supported by the aiida-dlpoly plugin:

• Trajectory history (TrajectoryData)
• Radial distribution function (SinglefileData)

These will be generated if the relevant input keys are present in the DL_POLY CONTROL file. Further information on AiiDA data types and how to interact with them can be found in the AiiDA documentation.

The aiida-dlpoly plugin additionally allows alternative input methods for the CONTROL and CONFIG inputs, which can both be optionally specified as a Dict or StructureData node respectively. The plugin itself will then handle the relevant file formatting for to be compatible with DL_POLY. An example showing these alternative inputs methods is shown below, it assumes an AiiDA StructureData node already exists in the user's database.

from aiida.engine import run
from aiida.orm import Dict, load_node, SinglefileData, load_code
from aiida import load_profile

load_profile("username") # Replace with your own AiiDA username

builder = load_code("dlpoly").get_builder()
builder.field = SinglefileData(file="/absolute/path/to/FIELD") # Change these to the absolute paths to input file
builder.configuration = load_node(pk=10) # Replace with the pk of the StructureData node or create a new StructureData node
builder.control = Dict({
    "title": "DL_POLY example MD simulation",
    "temperature": (295.0, "K"),
    "coul_method": "OFF",
    "print_frequency": (100, "steps"),
    "stats_frequency": (10, "steps"),
    "padding": (0.2, "ang"),
    "cutoff": (7.5, "ang"),
    "ensemble": "nve",
    "time_run": (2000, "steps"),
    "time_equilibration": (1000, "steps"),
    "time_job": (3000.0, "s"),
    "time_close": (100.0, "s"),
    "timestep": (0.001, "ps"),
    "rescale_frequency": (5, "steps"),
})

results, node = run.get_node(builder)