R/data_preparation.R

In gouthaman87/ameshouseR: What the Package Does (One Line, Title Case)

#' Data to Read
#'
#' @param path The path of the data
#'
#' @return Data Frame
#' @export
read_from_path <-
  function (path) {
    tryCatch({

      cli::cli_alert_info(
        cli::col_blue("Reading Raw Data From Path: {path}")
      )
      envir <- environment()
      data_name <- load(path, envir = envir)
      df <- get(data_name)

      cli::cli_alert_success(
        cli::col_green("Successfully Read Data")
      )
      return(df)
    }, error = function(e) {
      cli::cli_alert_warning(cli::col_br_red("{e}"))
    })
}


#' Train/Test Split
#'
#' @param DF The Data Frame
#' @param prop The proportion of Training Data
#' @param resp The name of response variable
#'
#' @return rsplit
#' @export
data_split <-
  function(DF, prop, resp) {

    tryCatch({
      cli::cli_alert_info(
        cli::col_blue("Split Data on Proportion : {prop}")
      )

      splt <- withr::with_seed(
        seed = 1987,
        rsample::initial_split(
          data = DF, prop = prop, strata = dplyr::all_of(resp)
        )
      )

      cli::cli_alert_success(
        cli::col_green("Successfully Splitted Data")
      )

      return(splt)

    }, error = function(e) {
      cli::cli_alert_danger(cli::col_br_red("{e}"))
    })
}


#' The Data Recipe with Feature Engineering
#'
#' @param DF The Data to create Recipe
#' @param features The predictor variables to fit
#'
#' @importFrom dplyr matches
#' @importFrom dplyr starts_with
#'
#' @return The recipe
#' @export
data_recipe <-
  function(DF, features) {

    tryCatch({
      cli::cli_alert_info(
        cli::col_blue("Create Recipe")
      )

      rec <- recipes::recipe(Sale_Price ~ ., data = DF) |>
        recipes::step_select(dplyr::matches(features))
      # recipes::step_log(dplyr::matches("Gr_Liv_Area"), base = 10) |>
      # recipes::step_other(dplyr::matches("Neighborhood"), threshold = tune::tune()) |>
      #
      # # Add Feature Hashing
      # # embed::step_feature_hash(recipes::all_nominal_predictors()) |>
      #
      # recipes::step_dummy(recipes::all_nominal_predictors()) |>
      #
      # # Add Interaction Term
      # recipes::step_interact(
      #  terms = ~ matches("Gr_Liv_Area"):starts_with("Bldg_Type_")
      # ) |>
      # recipes::step_zv(recipes::all_predictors()) |>
      #
      # # Add Spline Features
      # recipes::step_ns(dplyr::matches("Latitude"), deg_free = tune::tune("lat df")) |>
      # recipes::step_ns(dplyr::matches("Longitude"), deg_free = tune::tune("long df")) |>
      #
      # # Add PCA
      # recipes::step_pca(dplyr::matches("(SF$)|(GR_LIV)"))

      cli::cli_alert_success(
        cli::col_green("Successfully Created Recipe")
      )

      return(rec)

    }, error = function(e) {
      cli::cli_alert_danger(cli::col_br_red("{e}"))
    })
}