library(dplyr)
library(ggplot2)
library(viridis)
library(plotly)


# Define helper functions to get and parse sea ice data from NSIDC --------
read_nsidc_csv <- function(file) {
  # reads and loads a csv file of sea ice extent from NSIDC's ftp server
  hem_code <- substr(file, 1, 2)
  if (hem_code == "NH") {
    hemisphere <- "north"
  } else if (hem_code == "SH") {
    hemisphere <- "south"
  } else {
    stop("Filename is not correctly formatted")
  }
  prefix <- paste0("ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/",
                    hemisphere,
                   "/daily/data/")
  file_loc <- paste0(prefix, file)
  d <- read.csv(file_loc, stringsAsFactors = FALSE)
  d <- d[-1, ] # this row contains the units

  maybe_as_numeric <- function(x) {
    numeric_x <- suppressWarnings(as.numeric(x))
    if (!all(is.na(numeric_x))) x <- numeric_x
    x
  }

  # convert numeric columns to numbers and return data.frame
  lapply(d, maybe_as_numeric) %>%
    as.data.frame(stringsAsFactors = FALSE)
}


get_seaice_data <- function() {
  # gets seaice data from northern and southern hemispheres and joins to one df
  northern_hem <- full_join(read_nsidc_csv("NH_seaice_extent_final_v2.csv"),
                              read_nsidc_csv("NH_seaice_extent_nrt_v2.csv")) %>%
    mutate(hem = "Northern hemisphere")
  southern_hem <- full_join(read_nsidc_csv("SH_seaice_extent_final_v2.csv"),
                            read_nsidc_csv("SH_seaice_extent_nrt_v2.csv")) %>%
    mutate(hem = "Southern hemisphere")
  all_data <- full_join(northern_hem, southern_hem)
  all_data %>%
    mutate(date = ISOdate(Year, Month, Day),
           day_of_year = as.numeric(format(date, "%j")))
}


# load data from NSIDC's ftp server
seaice <- get_seaice_data()



# Visualize ---------------------------------------------------------------
seaice %>%
  ggplot(aes(x = day_of_year, y = Extent, color = Year, group = Year)) +
  geom_line() +
  xlab("Day of year") +
  ylab("Sea ice extent (million sq km)") +
  scale_color_viridis(direction = -1) +
  geom_text(aes(x = day_of_year, y = Extent, label = Year),
            data = subset(seaice, date == max(date)),
            nudge_x = c(15, -17)) +
  ggtitle(paste("Global sea ice extent:",
                 min(format(seaice$date, "%Y-%m-%d")),
                 "to",
                 max(format(seaice$date, "%Y-%m-%d")))) +
  facet_wrap(~ hem)
  ggplotly()