Building a module

The process of making a module is essentially 1. Write an R function 2. Run BuildModule with the function and metadata

and optionally 3. Upload to the zoon modules repository

An example

Here is a simple function that will become our module. It is a model module that uses general additive models.

NewModule <- function(df){

  zoon:::GetPackage("gam")

  covs <- as.data.frame(df[, 6:ncol(df)])
  names(covs) <- names(df)[6:ncol(df)]
  m <- gam::gam(formula = df$value ~ .,
         data = covs,
         family = binomial)

}

We then run BuildModule adding fairly extensive meta data and directing BuildModule to save the file in the working directory '.'. As this module has no parameters other than df which is not user specified, set paras to list().

BuildModule(NewModule, type = 'model', dir = '.', title = 'GAM sdm model', 
  description = 'This is my mega cool new model. It does GAMs using the gam package.',
  paras = NULL, author = 'Z. Oon', email = 'zoon@zoon.com')

This is now a runeable module.

rm(NewModule)
LoadModule('NewModule.R')
work1 <- workflow(occurrence = UKAnophelesPlumbeus,
                  covariate = UKAir,
                  process  = OneHundredBackground,
                  model = NewModule,
                  output   = PrintMap)

Once we're happy with the module, we will hopefully upload it to the zoon repository.

Module IO definitions for module developers

The input arguments and return values of modules are strict. However, any module type can have additional input arguments but these must be named. A lot of the data frames include '+ covariates'. This indicates that there the number of covariate columns is flexible.

Occurrence

Out: dataframe with column names: longitude, latitude, value, type, fold

Covariate

out: raster layer or raster stack

Process

in: list( df with values, type, fold, longitude, latitude + covariates,
covariate rasterstack/layer )

out: list(df with values, type, fold, longitude, latitude + covariates, covariate rasterstack/layer )

Model

in: dataframe from process out: model object (Object with a predict method. If not, please define a method within the module. See BiomodModel for an example.)

Output

in: list: list$model (model object from above) and list$data is a dataframe from process.output + predictions. Rasterlayer out: Anything

Pictoral description of inputs and outputs