farmrisk_dev

Development repository for the FARMR!SK algorithm. This repo contains the modular components and developer tools used to build and run the farmrisk algorithm locally. The app-ready repository can be found at: https://github.com/VetEpi-UAB/farmrisk.

This algorithm uses the mcmodule R package for modular Monte Carlo risk analysis. Installing the latest version from the GitHub repository is recommended. Before reinstalling packages from GitHub, first remove them using remove.packages() and then restart R.

Repository layout

• R/: custom functions, most of them are documented using roxygen2 standards.
• unified_script/: .R scripts generated from .qmd modules, used to generate the unified script for each model (e.g. cattle.R)
• input_files/:
  • admin/: files provided by the model developers
    • key_dictionary.csv: conversion between machine-readable keys and human-readable literals. Also indicates the key type and, if it's a factor level, the parent factor key. The default literal lit is English. Translations to Spanish (lit_esp), Catalan (lit_cat), and other languages will be provided.
    • administrative_levels.csv: list of regional administrative levels for Andorra, Austria, Belgium, Germany, Spain, Estonia, Finland, France, Greece, Italy, Luxembourg, Poland, and Portugal. Includes GADM hierarchy, names, and GID codes for municipalities (level 3), provinces (level 2), regions (level 1), and countries (level 0), or their equivalent levels. Note that some municipality (level 3) names are missing.
    • Model-specific files (e.g., cattle):
      • data/: tables containing bibliography and expert based-model parameters.
      • [model]_core_exp.R: general expressions to calculate risk step probabilities.
      • [model]_link_exp.R: expressions to connect inputs (survey and admin parameters) with core expressions.
      • [model]_data_keys.R: list that defines the columns and keys (fields that uniquely identify each row in a table) for each of the tables in data/.
      • [model]_mctable.R: table defining all the model inputs and their probability distributions. See model parameters documentation for more information.
  • user/: only needed to run the model locally. Should contain a directory named after the farm_id, which includes the farm's bsg and mov files.
• output_files/: default location for storing model outputs when run locally. A file with the farm_id will be created.
• forms/: JSON files to create SurveyJS forms, their version change logs, and migrations
• cattle.R: unified script that runs the cattle farm biosecurity risk analysis model
• run_farmrisk_local.R: unified script that runs farmrisk from local user inputs in the input_files/user/ directory
• run_farmrisk.R: unified script that runs farmrisk from user inputs retrieved from an SQL database (see connect to database)

Workflow

farmrisk development

[On progress, add process flow diagram, including unify_scripts()]

$$...$$

FARMR!SK app

Connect to SQL database

Save SQL database credentials in a text file called db_keys.txt in the same directory as run_farmrisk.R. Use the following script to create this file:

# Specify database credentials
hostname  <- ""
port      <- ""
db_name   <- ""
user      <- ""
password  <- ""

# Write keys file
writeLines(
  c(paste0("hostname=", hostname),
    paste0("port=", port),
    paste0("db_name=", db_name),
    paste0("user=", user),
    paste0("password=", password)),
  con = "db_keys.txt")

Run algorithm from command line

To launch the simulation from the command line, use the Rscript command to execute run_farmrisk.R and provide a farm_id:

Rscript run_farmrisk.R [farm_id]

Outputs

First, the simulation returns a detailed log of the algorithm processes in the console, useful for debugging.

Then, after the following message:

####################################

 FARM [farm_id] COMPLETED!

####################################

A JSON summarizing the simulation results is returned.

Contact

For questions or bug reports, please open an issue or contact Natalia Ciria (Natalia.Ciria@uab.cat)

Acknowledgements

farmrisk was developed with support from:

• BIOSECURE, a European Union HORIZON Europe FARM2FORK project
• INNOTUB, funded by the INTERREG POCTEFA 2021–2027 European territorial cooperation program
• Ministerio de Agricultura, Pesca y Alimentación de España
• Universitat Autònoma de Barcelona

Views and opinions expressed are those of the author(s) only and do not necessarily reflect those of the European Union or REA. Neither the European Union nor the granting authority can be held responsible for them.