ABM-COVID-DSL

High-resolution agent-based model of COVID-19 with testing and treatment

Code structure

Details about the code and usage examples can be found in the documentation, and the manual.

The directories are as follows:

• include - header files
• src - source files
• scripts - python scripts for general use throughout the repository
• tests - tests developed for this code
• parameters - data and scripts for fitting of some of the parameters
• simulations - simulations, specifically,
  • NewRochelle_population - model validation and optimization of testing prevalence
  • vaccination_study - vaccination strategies without restrictions
  • code_complexity - code for generating data for complexity estimates Specific features of these simulation sets are available in their own README files.

Compiling and running

Complete compilation scripts are written in Python and can be found in subdirectories of simulations. They are executed by running

python3.X compilation.py

in the terminal.

The code was developed and tested on MacOS and Linux, some of its parts may not be compatible with Windows.

Compiler requirement is availability of the C++11 standard. Python version used througout the development was 3.6, other 3.X version should work as well.

Tests

All the core and most of the side functionality of the code is extensively tested, with tests present in th tests directory. Each testing directory has its own python script that compiles and exectues all its tests. The test outcomes are printed to the screen and pass/fail are color-coded and clearly indicated. To run all of the tests execute:

python3.X run_all_tests.py 

in the tests directory. Some of the tests may at times fail due to insufficient statistical sampling (e.g. in distribution tests) or result precision (contribution_class tests on Linux platforms)

Documentation

The documentation is generated with Doxygen. To create your own, run doxygen Doxyfile in the root (this) directory.

Notes

• Current simulation wrappers and drivers are written in MATLAB. Python versions will soon be available.
• Optimization framework in simulations/NewRochelle_population is hardcoded - careful examination recommended before repurposing
• To run the optimization with MATLAB one needs to adjust the python path called to post-process an intermediate file. On older Linux systems one may also need to manually specify the location of libstdc++ library. General instructions are available through this post. Specifically, on a Red Hat Enterprise Linux Workstation release 7.3 (Maipo) the command sequence was:

locate libstdc++
export LD_PRELOAD=/usr/lib64/libstdc++.so.6.0.19

We encourage the users and developers to report any issues, including platform incompatibility and failed tests on our GitHub page or directly, by sending us an email.

Citation

If you find this software useful, consider citing

Truszkowska A, Behring B, Hasanyan J, Zino L, Butail S, Caroppo E, Jiang ZP, Rizzo A, Porfiri M. COVID‐19 Modeling: High‐Resolution Agent‐Based Modeling of COVID‐19 Spreading in a Small Town (Adv. Theory Simul. 3/2021). Advanced Theory and Simulations. 2021 Mar;4(3):2170005.