R/data_cleaning.R

In AGRBDCenter/aghelpr2: Helper functions for osfr and Ag Department RBDC

# This script houses data cleaning functions for different types of data.

#' Clean NIR data and put into tidy format. This requires that JA's input identifiers as unique and standardized. 
#' Format: First three letters of the project name, followed by an underscore, followed by the first three letters of the crop, 
#' followed by an underscore, followed by the treatment number as "T1", followed by the block as "B1". If my project were "Mosaic Potato",
#' and I were running treatment 1, block 3, the format would be: "Mos_Pot_T1B3".
#'
#' @param path The path to the file in the directory.
#' @param proj_name The name of the project exactly as found in the "Admin Sheet" on OSF.
#' @param label The first three letters of the project and the first three letters of the crop. ie "Mos_Pot". Whatever was put into NIR machine.
#' @export clean_nir

clean_nir <- function(path, proj_name, label) {
  # Read in the data and filter by the label
  enc <- readr::guess_encoding(path)[[1]][1]
  data <- read.csv(path, fileEncoding = enc, header = FALSE)
  data <- filter(data, grepl(label, data$V4))
  
  # Move the column headers from row values to the next column's name
  for (i in 6:length(data)) {
    if (i %% 2 == 0) {
      data[[i]] <- as.character(data[[i]])
      name <- unique(data[[i]])
      names(data)[[i + 1]] <- name
    }
  }
  
  # Remove columns which only contain column names
  for (i in 6:length(data)) {
    colname <- paste0("V", i)
    for (name in names(data)) {
      if (name == colname) {
        data <- data[, names(data) != name]
      }
    }
  }
  
  # Rub-a-dub-dub squeeky clean data
  data <- data %>% 
    gather(variables, value, 6:length(data)) %>% 
    filter(!is.na(value)) %>% 
    mutate(project_name = proj_name) %>% 
    rename(date_collected = V1) %>% 
    mutate(treatments = gsub("T", "", regmatches(V4, regexpr("T([0-9]+)", V4)), ignore.case = TRUE),
           replications = gsub("B", "", regmatches(V4, regexpr("B([0-9]+)", V4)), ignore.case = TRUE),
           date_collected = lubridate::mdy(date_collected)) %>% 
    select(date_collected, project_name, variables, treatments, replications, value)
  
  return(data)
}


#' Clean harvester (Haldrup) data and put into tidy format. This requires that we know the direction of harvest, the direction of the plot map, and a standardized name for the harvest.
#'
#' @param path The path to the file in the directory.
#' @export clean_harvester

clean_harvester <- function(path) {
  return(data)
}


#' Clean Dr. Hopkin's lab results data. This requires that the project name be included in the file treatments column which requires that we send the project name with the samples exactly as we want it entered.
#'
#' @param path The path to the file in the directory.
#' @export clean_labresults

clean_labresults <- function(path) {
  return(data)
}


#' Clean potato weights (harvest) data. This requires that the data be in a standardized format which will be determined.
#'
#' @param path The path to the file in the directory.
#' @export clean_potato

clean_potato <- function(path) {
  return(data)
}


#' Clean specific gravity. This requires that the data be in a standardized format which will be determined.
#'
#' @param path The path to the file in the directory.
#' @export clean_specgrav

clean_specgrav <- function(path) {
  return(data)
}


#' Takes a list of data and puts it together in a single dataframe.
#'
#' @param data A list of dataframes.
#' @export compose_data

compose_data <- function(data) {
  full_data <- data.frame()
  for (i in seq_along(data)) {
    full_data <- plyr::rbind.fill(full_data, data[[i]])
  }
  return(full_data)
}