R/ensureProductionOutputs.R

In SWS-Methodology/faoswsProduction: Package to perform the imputation of area harvested, production and yield for the FAO production domain

##' Check Production Inputs
##'
##' This function is designed to ensure that the dataset to be saved is valid.
##'
##' @param data A data.table containing the data.
##' @param processingParameters A list containing the parameters to be used in
##'     the processing algorithms. See ?defaultProcessingParameters for a
##'     starting point.
##' @param formulaParameters A list holding the names and parmater of formulas.
##'     See \code{productionFormulaParameters}.
##' @param testImputed logical, whether to test imputation result
##' @param testCalculated logical, whether to test calculated result.
##' @param domain The domain to be saved back
##' @param dataset The dataset to be saved back
##' @param returnData logical, whether the data should be returned
##' @param normalised logical, whether the data is normalised
##'
##' @return The original data if all tests are passed
##'
##' @export
##'

ensureProductionOutputs = function(data,
                                   processingParameters,
                                   formulaParameters,
                                   testImputed = FALSE,
                                   testCalculated = FALSE,
                                   domain = "agriculture",
                                   dataset = "aproduction",
                                   returnData = TRUE,
                                   normalised = TRUE){
    dataCopy = copy(data)

    if(normalised){
        dataCopy = denormalise(dataCopy, "measuredElement")
    }

    with(formulaParameters,
         with(processingParameters,
         {
             suppressMessages({
                 ## Check data inputs
                 ensureDataInput(data = dataCopy,
                                 requiredColumn = c(productionValue,
                                                    productionObservationFlag,
                                                    productionMethodFlag,
                                                    yieldValue,
                                                    yieldObservationFlag,
                                                    yieldMethodFlag,
                                                    areaHarvestedValue,
                                                    areaHarvestedObservationFlag,
                                                    areaHarvestedMethodFlag,
                                                    itemVar,
                                                    yearVar,
                                                    areaVar),
                                 returnData = FALSE)


                 ## Ensure there is no production is zero while area harvested
                 ## is non zero, vice versa.
                 ensureNoConflictingZero(data = dataCopy,
                                         valueColumn1 = productionValue,
                                         valueColumn2 = areaHarvestedValue,
                                         returnData = FALSE,
                                         normalised = FALSE)

                 ## Ensure the range of values are correct
                 ##
                 ## NOTE (Michael): Yield can not be equal to zero
                 ensureValueRange(data = dataCopy,
                                  ensureColumn = yieldValue,
                                  min = 0,
                                  max = Inf,
                                  includeEndPoint = TRUE)
                 ensureValueRange(data = dataCopy,
                                  ensureColumn = areaHarvestedValue,
                                  min = 0,
                                  max = Inf,
                                  includeEndPoint = TRUE)
                 ensureValueRange(data = dataCopy,
                                  ensureColumn = productionValue,
                                  min = 0,
                                  max = Inf,
                                  includeEndPoint = TRUE)


                 ## Ensure missing values are correctly specified
                 ensureCorrectMissingValue(data = dataCopy,
                                           valueVar = yieldValue,
                                           flagObservationStatusVar =
                                               yieldObservationFlag,
                                           returnData = FALSE,
                                           getInvalidData = FALSE)
                 ensureCorrectMissingValue(data = dataCopy,
                                           valueVar = areaHarvestedValue,
                                           flagObservationStatusVar =
                                               areaHarvestedObservationFlag,
                                           returnData = FALSE)
                 ensureCorrectMissingValue(data = dataCopy,
                                           valueVar = productionValue,
                                           flagObservationStatusVar =
                                               productionObservationFlag,
                                           returnData = FALSE)

                 ## Ensure flags are valid
                 ensureFlagValidity(data = dataCopy,
                                    normalised = FALSE)

                 ## Ensure production is balanced
                 ##
                 ## NOTE (Michael): This may be optional in input, but mandatory in
                 ##                 output
                 ##ensureProductionBalanced(data = dataCopy,
                 ##                         areaVar = areaHarvestedValue,
                 ##                         yieldVar = yieldValue,
                 ##                         prodVar = productionValue,
                 ##                         conversion = unitConversion,
                 ##                         returnData = FALSE,
                 ##                         normalised = FALSE)
             })

             if(testImputed){
                 ## Ensure time series are imputed
                 suppressMessages({
                     ensureTimeSeriesImputed(data = dataCopy,
                                             key = c(areaVar, itemVar, elementVar),
                                             returnData = TRUE,
                                             normalised = FALSE,
                                             getInvalidData = FALSE)
                 })
             }

             if(testCalculated){
                 ## Ensure the identity is calculated.
                 suppressMessages({
                     ensureIdentityCalculated(data = dataCopy,
                                              areaVar = areaHarvestedValue,
                                              yieldVar = yieldValue,
                                              prodVar = productionValue,
                                              returnData = FALSE,
                                              normalised = FALSE,
                                              getInvalidData = FALSE)
                 })
             }

         }))

    if(normalised){
        dataCopy = normalise(dataCopy)
    }

    if(returnData)
        return(dataCopy)
}