kausaltech / reina-model

Agent-based simulation model for COVID-19 spread in society and patient outcomes

Agent-based COVID-19 simulation

This software implements an agent-based (or individual-based) model for simulating the spread of the coronavirus (SARS-CoV-2) in a healthcare district.

It models healthcare capacity (hospital beds and ICU units) and different public health simulations.

More documentation is also available in the repository.

Installation

If you're using Ubuntu/Debian, you might need to install a couple of packages:

apt install python3.8-dev libpython3.8-dev python3.8-venv \
    libfreetype6-dev libpng-dev libqhull-dev pkg-config \
    gcc gfortran libopenblas-dev liblapack-dev cython

Install a Python venv with Python 3.7 or 3.8. Install the requirements:

pip install -r requirements.txt

Compile translations:

pybabel compile -d locale

Usage

Run the simulation with:

python -m calc.simulation

Or visualize using Dash:

python -m corona

Installing and running with Docker

You can run Reina with Docker. Check reina-ui project to ../reina-ui and set following environment variables to .env file:

ENV=dev
FLASK_ENV=development
NEXT_PUBLIC_GRAPHQL_ENDPOINT=http://localhost:8080/graphql

Then run

docker-compose up

The first time you run it, it will build the container for Reina, which takes some time. Once build done, Reina UI is available at localhost:8080 by default. If you want to use some other port, set it in REINA_PORT environment variable.

While the container is running, you can run the simulation like this:

docker exec -ti reina-model python -m calc.simulation

Configuring the hospital district

Currently, Reina supports Helsinki and Uusimaa district (HUS) and Varsinais-Suomi hospital district out of the box. By default, it's configured to run for HUS, but you can set the district as Varsinais-Suomi by setting the environment variable AREA_NAME as follows:

AREA_NAME=Varsinais-Suomi

If you're running your setup via docker-compose, you can set the environment variable in .env file in the project root directory and it will work directly.

Development

Localisation

Extract new translation strings to the template:

pybabel extract -w 120 -F babel.cfg -o locale/messages.pot .

Merge new strings to language-specific translation files:

pybabel update -w 120 -i locale/messages.pot -d locale