Simple percolation model

A simple percolation model intended as a 'computer experiment' for Physics 1B undergraduate labs at the University of Edinburgh, used during the Spring semester of 2021 when courses had to be delivered remotely.

WARNING for UoE students taking P1B labs:

If you are on Notable and see this, it means you have cloned the wrong branch!

Delete the whole folder and go through the steps to clone the repository again, but this time replace master with p1b in the 'branch' box.

Installation

Local installation

Clone the repo using one of the numerous options available! For example,

git clone https://github.com/marshrossney/percolation.git
cd percolation

This project has rather minimal dependencies, and should run fine with reasonably up-to-date versions of NumPy, SciPy and Matplotlib.

To use the Jupyter notebooks, you also need...Jupyter. Alternatively, you can run everything from the command line, which will require the ConfigArgParse tool.

The easiest option is to simply create a Conda environment using the environment file provided,

conda create -n perc -f environment.yml
conda activate perc

and then install the package there.

python -m pip install -e .

Installation via Noteable

While it is easier to just use the 'p1b' branch, it is possible to install this as a package on Notable.

Open Notable and launch a Standard Notebook (Python 3)
Clone the repository using the '+GitRepo' tool
Open a terminal instance and run

cd percolation
python -m pip install -e .

You should now be able to run the Jupyter notebooks in Notable

Usage

In your own scripts

Basic usage will look like the following.

from percolation.lattice import SquareLattice
from percolation.model import PercolationModel
from percolation.scripts.parameter_scan import parameter_scan

# Create a 100x100 lattice
lattice = SquareLattice(n_rows=100, n_cols=100)

# Instantiate a model
model = PercolationModel(lattice, inert_prob=0.4)

Have a look at how it evolves with these parameters
model.animate(n_steps=200)

# Plot an estimate for the probability that the model 'percolates' over a
# range of values of 'frozen_prob'
parameter_scan(model, start=0.1, stop=0.7, num=25, repeats=25)

Look at the options available by running e.g.

help(parameter_scan)

Command line

Installing the package will install a few scripts that can be run from the command line. At the moment these are:

perc-anim which saves an animation as a gif
perc-scan which runs a 'parameter scan' (ideally over the percolation transition) and produces a nice plot.
perc-time which just runs timeit on a couple of things and is mostly just useful to me.

Run e.g.

perc-anim --help

to see what arguments are required, or possible, to run the script.

You can supply arguments via the command line, but it's probably easier to store them in a configuration file, labelled input.yml below.

perc-scan -f input.yml

See the examples at percolation/examples/ for some basic input files.

Running the tests

Install pytest (included in conda environment).

In the root of the repository, run

pytest

There you go...

To do

Things I'm planning to do if I get some time...

More types of lattices and networks to use in place of the square lattice.
More tools to analyse percolation transition.
More options that make it more intuitive to relate the model to a pandemic.
Jupyter notebooks for teaching and demonstrating.

Queries and feedback

Open a Github issue, pull request or just email me < Joe.Marsh-Rossney 'at' ed.ac.uk >.