jplots: Some Little Plots

#' #' Interpolation in a dataframe
#' #' 
#' #' Plotting envelopes or computing min/max/percentile statistics over a time series requires that the same year intervals be used in each series. This function is intended to fill in intermediate values through interpolation, in the context of a completed data frame, where NA valeus represent values to be filled.
#' #' 
#' #' This function is an awkward fit for this package, but I didn't know where to put it!
#' #' 
#' #' The additional variable specifications, result_var and interp_var, can be useful for debugging because they provide the constituent detailed data.
#' #' 
#' #'@param data within which to interpolate
#' #'@param x_var
#' #'@param y_var string or symbol of the value variable being interpolated
#' #'@param by_vars character vector of variables that define grouping
#' #'
#' #'@import dplyr
#' #'@import tidyr
#' #'
#' #'@export
#' interpolate_data <- function(.data, x_var, y_var, result_var, interp_var, by_vars) {
#'   
#'   # currently all required.
#'   # TODO: add back result_var and interp_var?
#'   x_var <- rlang::ensym(x_var)
#'   y_var <- rlang::ensym(y_var)
#'   by_vars   <- rlang::syms(by_vars)
#'   
#'   
#'   #Complete the implicit missing years/values
#'   data_complete_w_na <- .data %>%
#'     tidyr::complete(nesting(!!!by_vars), !!x_var) %>%
#'     arrange(!!!by_vars, !!x_var) # does approx need ordered?
#'   
#'   ### INTERPOLATION SECTION
#'   
#'   #This section performs linear interpolation.
#'   #It takes as input data frames which are already "completed" with the NA values that will be filled in through interpolation.
#'   #Apply the interpolation function to the entire dataset (only within data NAs are interpolated, outside core data.
#'   #NAs are returned as NAs.
#'   
#'   data_interpolated <- data_complete_w_na %>%
#'     group_by(!!!by_vars) %>%
#'     nest() %>%
#'     mutate(int_pts = map(data,
#'                          .f = ~ safe_approx(x=.[[x_var]], y=.[[y_var]], xout = .[[x_var]]) )) %>%
#'     mutate(int_pts = map(int_pts,
#'                          ~select(.x, !!y_var := y,
#'                                  ))) %>%
#'     mutate(data = map(data, ~select(.x, -!! y_var))) %>%
#'     
#'     unnest(cols = c(data, int_pts)) %>%
#'     
#'     select(!!!by_vars, !!x_var, !!y_var)
#'   
#'   data_interpolated
#' }
#' 
#' #wraps the approx call, checking to see if there are multiple points to interpolate or not
#' 
#' safe_approx <- function(x, y, xout) {
#'   
#'   if(sum(!is.na(x) & !is.na(y)) > 1) {
#'     approx(x, y, xout) %>% as_tibble()
#'   } else {
#'     tibble(x=x, y=y)
#'   }
#'   
#' }