R/ss_runsetter.R

In poales/sharkeySpec: Handle and analyze SharkeySpec data

#' A narrow function to read through an appended dataset and attempt to automatically split it up.
#'
#' Use on appended data files where multiple measurements are made in one file. Will create an "ID" column for each individual measurement, as determined by threshhold.
#' When the difference in the chosen column is greater than the provided threshhold it will decide that it is a new measurement.
#' @param df The data table (singular) to analyze
#' @param threshhold The difference in the chosen column that is to be considered a new measurement.
#' @param aslist When FALSE, returns a dataframe with an ID column that distinguishes separate measurements. When TRUE, returns the same but divided into a list, with one element per measurement.
#' @param column The chosen column to study for differences in threshhold. Defaults to "Time," used by me to differentiate different measurements taken in the same file but separated by time.
#' @name ss_runsetter
#' @export



ss_runsetter<- function(df,threshhold=2,aslist = FALSE,column = "Time"){
  #scan through the dataframe and look for gaps
  #set gaps to be the break in runs with an ID column
  TDIFF <- function(df){
    diff <- df[-1,column]-df[-nrow(df),column]
    changes <- which(diff>threshhold | diff<(-1*threshhold))
    if(length(changes)>0){
      for(i in 1:length(changes)){
        df$ID[(changes[i]+1):length(df$ID)] <- df$ID[(changes[i]+1):length(df$ID)]+1
      }
    }
    return(df)
  }

  if((column %in% colnames(df))){
    if(!("ID" %in% colnames(df))){

      df <- tibble::add_column(df,ID=1)
      df <- TDIFF(df)
      if(aslist){
        mylist <- list()
        for(i in unique(df$ID)){
          mylist[[i]] <- dplyr::filter(df,df$ID==i)
        }
        df <- mylist
        #df <- lapply(df,function(x) select(x,-ID))
      }

    } else {
      cat("Error setting runs: ID column already extant in df")
    }

  } else{
    cat("Error setting runs: Chosen column name not in df")
  }

  return(df)

}