2021), coordinated by Pauline Ezanno, Sébastien Picault and Timothée
Vergne (INRAE).
Vergne (INRAE). It is published as supplementary information for the
following article:
> "The African swine fever modelling challenge: objectives, model description and synthetic data generation" by Sébastien Picault, Timothée Vergne, Matthieu Mancini, Servane Bareille and Pauline Ezanno (submitted to _Epidemics_)
<!-- markdown-toc start - Don't edit this section. Run M-x markdown-toc-refresh-toc -->
**Table of Contents**
-[ASF Challenge 2020](#asf-challenge-2020)
-[Content](#content)
-[Running model M0](#running-model-m0)
-[Data provided to ASF Challenge participants ("players")](#data-provided-to-asf-challenge-participants-players)
<!-- markdown-toc end -->
Content of the repository
-------------------------
-`README.md`: this file
-`model`: directory with model files
-`ppa.yaml`: model file (in EMULSION modelling language)
-`ppa.py`: additional Python file to implement specific features not provided by EMULSION (kernel diffusion, special logging, specific control measures...)
-`outputs/challenge`: simulation outputs corresponding to the trajectory used for the challenge
-`data`: directory containing input data used to feed the model
-`boars.csv`: spatial distribution of the centres of home ranges of wild boar
-`herds.csv`: spatial distribution and features of pig farms
-`moves.csv`: pre-defined trade movements between pig farms
-`create_tiles.sh`: script to discretize the whole island into 15x15 km tiles, create the diffusion kernels (as sparse matrices), and store them into `tiles_15000`
-`list_of_tiles.txt`: list of non-empty 15x15 km$`^2`$ tiles
-`extract_area.py`: Python script to (re)build 1 specific tile
-`fences.csv`, `buffer.csv`: list of 15x15 km$`^2`$ tiles which represent the fenced area and the surrounding buffer, respectively
-`kick-of_meeting`: presentation slides used for the kick-off meeting (August 2020)
-`Data_all_players`: initial data and situation reports provided for each challenge period
Content of the README
-------------------------
-``repository.md``: this file
-``model``: directory with model files
-``ppa.yaml``: model file (in EMULSION modelling language)
-``ppa.py``: additional Python file to implement specific features not provided by EMULSION (kernel diffusion, special logging, specific control measures...)
-``outputs/challenge``: simulation outputs corresponding to the trajectory used for the challenge
-``data``: directory containing input data used to feed the model
-``boars.csv``: spatial distribution of the centres of home ranges of wild boar
-``herds.csv``: spatial distribution and features of pig farms
-``moves.csv``: pre-defined trade movements between pig farms
-``create_tiles.sh``: script to discretize the whole island into 15x15 km tiles, create the diffusion kernels (as sparse matrices), and store them into ``tiles_15000``
-``list_of_tiles.txt``: list of non-empty 15x15 km$^2$ tiles
-``extract_area.py``: Python script to (re)build 1 specific tile
-``fences.csv``, ``buffer.csv``: list of 15x15 km$^2$ tiles which represent the fenced area and the surrounding buffer, respectively
-``kick-of_meeting``: presentation slides used for the kick-off meeting (August 2020)
-``Data_all_players``: initial data and situation reports provided for each challenge period
Data provided to ASF Challenge participants ("players")
The installation procedure for EMULSION is fully described on [the software website](https://sourcesup.renater.fr/www/emulsion-public/Install.html). All simulations were carried out under Linux. Install version 1.1rc5 which was used to produce the synthetic data used for the ASF Challenge and discussed in our papers. Ensure that your `PYTHONPATH` variable is set properly (i.e. containing current repository `.`).
2. Clone the git repository and move into the `model/` directory:
Input data for the ASF model is stored by default in the `data/` directory. The git repository contains all necessary files, except for the exponential kernels and neighbourhood matrices used for calculating the forces of infection in the different transmission pathways. These files must be rebuilt before running the model. The trunctation of the kernels (set to 0 below a threshold) make it possible to discretize the whole area into 15x15 km$`^2`$ tiles. Beware, this step maybe time-consuming, and it produces 8.2 Gb data.
