Outbreak-Sim

A large-scale Epidemiological Agent-Based model.

Repository Layout

Synthetic Environment Generator and Reporting Utilities (.\python)

• Synthetic Environment Generator - A series of Python notebooks dedicated to the task of generating Synthetic Environments, the required input for Outbreak-Sim, which defines the starting state of the simulation, and contains components like agent, household, and workplace positions, agent attributes such as age and workplace allocations, etc.

• Report Analysis - A series of Python notebooks for parsing and analysing the report files created by simulation runs of Outbreak-Sim.

Note: the rest of this README is dedicated to Outbreak-Sim, the Rust simulation runner, see the above README's for detailed information on the other processes and components

Synthetic Environment Schema (.\schema\model.fbs)

A Flatbuffers schema describing the format of a Synthetic Environment data file. The schema is used for generating code with the ./codegen.sh script.

Outbreak-Sim

The runner for the epidemiological simulations. Consists of a rust library, outbreak_sim, an example binary main.rs and a set of benchmarks in .\benches.

Requirements

• Rust
• FlatBuffer Compiler that's accessible on the path as flatc with a version of at least 2.0.0

Setup

Note: Development has been tested on Windows using Windows Subsystem for Linux (WSL), other OS's and combinations may work but are not supported.

• run .\codegen.sh to generate code using Flatbuffers from the schema found in .\schema\model.fbs
• If not using one of the pre-supplied synthetic environment files found within .\python\synthetic_environments\examples then follow the instructions within the respective README to generate a new one.

Usage

• Modify or use .\src\main.rs as an example in how to use the library. The example binary includes a rudimentary form of visualising the simulation as it runs. This will in the future be moved to a better implementation and made optional through a feature flag.
• The following are key parameters to be modified (inputting parameters is a work-in-progress and intended to be less involved and more explicit in the future):
  • The parameters of outbreak_sim::Sim::new require:
    • a path to the directory containing the synthetic environment file
    • the name of the file (without file-extension)
  • outbreak_sim::reporting::intialise_reporting_files requires, as a parameter, a path to a directory to write reports to
  • outbreak_sim::Sim::new currently has transmission_chance hard-coded in the function body. This should be modified as needed, however in future-development the MixingStrategy implementation will be continued, and this will move to a sensible place.
  • outbreak_sim::shared has the following global variables which are hard-coded and will be updated similarly as described above:
    • SEED_INFECTION_CHANCE: The chance of an agent being infected at the start of the simulation
    • TIME_STEPS_PER_DAY: The number of simulation time-steps in each day of in-simulation time. When containers in public transport routing is implemented this will need to be 1440 (a time-step being equivalent to a minute) whenever the feature is turned on. Currently this can be safely changed to a smaller number to speed up simulation speed.