R_tasks/prepare_SDQ20.R

In gorkang/jsPsychHelpeR: Standardize and automatize data preparation and analysis of jsPsych experiments created with jsPsychMaker

##' Prepare SDQ20
##'
##' Template for the functions to prepare specific tasks. Most of this file should not be changed
##' Things to change: 
##'   - Name of function: prepare_SDQ20 -> prepare_[value of short_name_scale_str] 
##'   - dimensions parameter in standardized_names()
##'   - 2 [ADAPT] chunks
##'
##' @title prepare_SDQ20
##'
##' @param short_name_scale_str 
##' @param DF_clean
##'
##' @return
##' @author gorkang
##' @export
prepare_SDQ20 <- function(DF_clean, short_name_scale_str, output_formats) {

  # DEBUG
  # targets::tar_load_globals()
  # jsPsychHelpeR::debug_function(prepare_SDQ20)
  

  
  
  # [ADAPT 1/3]: Items to ignore and reverse, dimensions -----------------------
  # ****************************************************************************
  
  description_task = "" # Brief description here
  
  items_to_ignore = c("000") # Ignore these items: If nothing to ignore, keep as is
  items_to_reverse = c("000") # Reverse these items: If nothing to reverse, keep as is
  
  ## NameDimension1, NameDimension2 should be the names of the dimensions
  ## Inside each c() create a vector of the item numbers for the dimension
  ## Add lines as needed. If there are no dimensions, keep as is
  items_dimensions = list(
    SintomasDisociativosTractoGenitoUrinario = c("001", "010", "028"),
    AlteracionesExteroSensoriales = c("004", "007", "016", "025", "031", "034", "037", "040"),
    AlteracionesInteroceptivas = c("013", "022", "043", "058"),
    AlteracionesMotoras = c("019", "046", "049", "052", "055")
  )
  
  # TEXT_items = c("SDQ20_003", "SDQ20_006", "SDQ20_009", "SDQ20_012", "SDQ20_015", "SDQ20_018", "SDQ20_021", "SDQ20_024", "SDQ20_027", "SDQ20_030", "SDQ20_033", "SDQ20_036", "SDQ20_039", "SDQ20_042", "SDQ20_045", "SDQ20_048", "SDQ20_051", "SDQ20_054", "SDQ20_057", "SDQ20_060")
  
  
  
  # [END ADAPT 1/3]: ***********************************************************
  # ****************************************************************************
  
  
  # Standardized names ------------------------------------------------------
  names_list = standardized_names(short_name_scale = short_name_scale_str, 
                                  dimensions = names(items_dimensions),
                                  help_names = FALSE) # [KEEP as FALSE]
  
  # Create long -------------------------------------------------------------
  DF_long_RAW = create_raw_long(DF_clean, 
                                short_name_scale = short_name_scale_str, 
                                numeric_responses = FALSE, # [TRUE or FALSE]
                                is_experiment = FALSE, 
                                keep_time = FALSE, # Keep time stamp for each response
                                help_prepare = FALSE) # Show n of items, responses,... [CHANGE to TRUE to debug] 
  
  
  # Create long DIR ------------------------------------------------------------
  DF_long_DIR = 
    DF_long_RAW |> 
    # If using keep_time = TRUE above, use this and add timestamp to the select() call
    # dplyr::mutate(timestamp = as.POSIXlt(datetime, format = "%Y-%m-%dT%H%M%S")) |> 
    dplyr::select(id, trialid, RAW) |>
    
    
    
  # [ADAPT 2/3]: RAW to DIR for individual items -------------------------------
  # ****************************************************************************
  
    # Transformations
    dplyr::mutate(
      DIR =
       dplyr::case_when(
         RAW == "Si" ~ 1,
         RAW == "No" ~ 0,
         
         RAW == "Nunca" ~ 1,
         RAW == "Pocas veces" ~ 2,
         RAW == "De vez en cuando" ~ 3,
         RAW == "Frecuentemente" ~ 4,
         RAW == "Casi todo el tiempo" ~ 5,
         RAW == "Casi todo o todo el tiempo" ~ 5,
          is.na(RAW) ~ NA_real_, # OR NA_character_,
          trialid %in% paste0(short_name_scale_str, "_", items_to_ignore) ~ NA_real_, # OR NA_character_,
          TRUE ~ 9999 # OR "9999"
        )
    ) |> 
    
    # Invert items [CAN BE DELETED IF NOT USED or DIR is non-numeric]
    dplyr::mutate(
      DIR = 
       dplyr::case_when(
          DIR == 9999 ~ DIR, # To keep the missing values unchanged
          trialid %in% paste0(short_name_scale_str, "_", items_to_reverse) ~ (6 - DIR), # REVIEW and replace 6 by MAX + 1
          TRUE ~ DIR
        )
    )
    
  # [END ADAPT 2/3]: ***********************************************************
  # ****************************************************************************
    

  # Create DF_wide_RAW_DIR -----------------------------------------------------
  DF_wide_RAW =
    DF_long_DIR |> 
    tidyr::pivot_wider(
      names_from = trialid, 
      values_from = c(RAW, DIR),
      names_glue = "{trialid}_{.value}") |> 
    
    # NAs for RAW and DIR items
    dplyr::mutate(!!names_list$name_RAW_NA := rowSums(is.na(across((-matches(paste0(short_name_scale_str, "_", items_to_ignore, "_RAW")) & matches("_RAW$"))))),
                  !!names_list$name_DIR_NA := rowSums(is.na(across((-matches(paste0(short_name_scale_str, "_", items_to_ignore, "_DIR")) & matches("_DIR$"))))))
  

  
  # [ADAPT 3/3]: Scales and dimensions calculations ----------------------------
  # ****************************************************************************
  
  # Reliability -------------------------------------------------------------
  # REL1 = auto_reliability(DF_wide_RAW, short_name_scale = short_name_scale_str, items = items_dimensions[[1]])
  # items_RELd1 = REL1$item_selection_string
    
  
  # [USE STANDARD NAMES FOR Scales and dimensions: names_list$name_DIRd[1], names_list$name_DIRt,...] 
  # CHECK with: create_formulas(type = "dimensions_DIR", functions = "sum", names(items_dimensions))
  DF_wide_RAW_DIR =
    DF_wide_RAW |> 
    dplyr::mutate(

      # [CHECK] Using correct formula? rowMeans() / rowSums()
      
      # Score Dimensions (see standardized_names(help_names = TRUE) for instructions)
      !!names_list$name_DIRd[1] := rowSums(across(all_of(paste0(short_name_scale_str, "_", items_dimensions[[1]], "_DIR"))), na.rm = TRUE),
      !!names_list$name_DIRd[2] := rowSums(across(all_of(paste0(short_name_scale_str, "_", items_dimensions[[2]], "_DIR"))), na.rm = TRUE),
      !!names_list$name_DIRd[3] := rowSums(across(all_of(paste0(short_name_scale_str, "_", items_dimensions[[3]], "_DIR"))), na.rm = TRUE),
      !!names_list$name_DIRd[4] := rowSums(across(all_of(paste0(short_name_scale_str, "_", items_dimensions[[4]], "_DIR"))), na.rm = TRUE)
      # Reliability Dimensions (see standardized_names(help_names = TRUE) for instructions)
      # !!names_list$name_RELd[1] := rowMeans(across(all_of(paste0(short_name_scale_str, "_", items_RELd1, "_DIR"))), na.rm = TRUE), 

      # Score Scale
      # !!names_list$name_DIRt := rowSums(across(all_of(matches("_DIR$"))), na.rm = TRUE)
      
    )
    
  # [END ADAPT 3/3]: ***********************************************************
  # ****************************************************************************


  # CHECK NAs -------------------------------------------------------------------
  check_NAs(DF_wide_RAW_DIR)
  
  # Save files --------------------------------------------------------------
  save_files(DF_wide_RAW_DIR, short_name_scale = short_name_scale_str, is_scale = TRUE, output_formats = output_formats)
  
  # Output of function ---------------------------------------------------------
  return(DF_wide_RAW_DIR) 
 
}