################################################################################
# - 레이다유량계 : RFM
# - 초음파수위계 : UWL
# - 초음파관유량계(S) : PFS
# - 기상관측계 : CWS
# - 지하수위계 : GWL
# - 다층 토양수분계 : MSM
iwmp.convert.exported.data.to.station.data <- function(EnvList){

  library(data.table)
  library(dplyr)
  library(tidyr)

  prjdir <- EnvList$prjdir
  sensor.type <- EnvList$sensor.type
  station.ids <- EnvList$station.ids
  start_year <- EnvList$start_year
  end_year <- EnvList$end_year

  raw.dir <- file.path(prjdir, "01_download")
  stn.dir <- file.path(prjdir, "02_extracted.station")

  if(!dir.exists(stn.dir)) dir.create(stn.dir, showWarnings = F, recursive = T)

  flist <- list.files(raw.dir, pattern = glob2rx("*.csv"), full.names = F)
  if(any(sensor.type == "all")){
    stypes <- unique(matrix(unlist(strsplit(flist, "-")), nrow=length(flist), byrow=T)[,1])
  } else {
    stypes <- sensor.type
  }

  for(i in 1:length(stypes)){
    stype <- stypes[i]
    if(any(sensor.type == "all")){

      flist <- list.files(raw.dir, pattern = glob2rx(sprintf("%s*-all*.csv", stype)), full.names = T)
      data.list <- list()
      for(j in 1:length(flist)){
        fname <- flist[j]
        data.list[[j]] <- read.csv(fname, skip = 1, header=F)
      }
      data.dt <- data.table::rbindlist(data.list)

      ## add column names of raw data --> change to function
      if(stype == "cws"){
        colnames(data.dt) <- c("id", "date", "hr", "prcp",  "temp", "dew", "rhum", "wspd", "wdir", "rsds", "sshine", "lux", "pet", "presure", "bat")
      }
      if(stype == "msm"){
        colnames(data.dt) <- c("id", "date", "hr", "sm10", "st10", "sm20", "st20", "sm30", "st30", "sm40", "st40", "sm50", "st50", "bat")
      }
      if(stype == "pfs"){
        colnames(data.dt) <- c("id", "date", "flow.day", "flow.acu", "flow.sen", "para", "bat")
      }
      if(stype == "rfm"){
        colnames(data.dt) <- c("id", "date", "hr", "vel00", "vel05", "vel10", "vel15", "vel20", "vel25", "vel30", "vel35", "vel40", "vel45", "vel50", "vel55",
                               "dep00",  "dep05", "dep10", "dep15", "dep20", "dep25", "dep30", "dep35", "dep40", "dep45", "dep50", "dep55", "h.ref", "bat")
      }
      if(stype == "uwl"){
        colnames(data.dt) <- c("id", "date", "hr", "dep00",  "dep05", "dep10", "dep15", "dep20", "dep25", "dep30", "dep35", "dep40", "dep45", "dep50", "dep55", "dep.avg", "h.ref", "bat")
      }

      stnids <- unique(data.dt$id)

      for(j in 1:length(stnids)){
        stnid <- stnids[j]
        stn.dt <- as.data.table(data.dt[which(data.dt$id == stnid), ])

        if(stype == "cws"){
          stn.dt <- setorder(stn.dt, date)
          stn.dt <- setDT(stn.dt)[, c("bat") := NULL]

          stn.dt$time <- as.POSIXct(sprintf("%s %02d:00:00", stn.dt$date, as.numeric(stn.dt$hr)))
          stn.dt$time <- format(stn.dt$time, format = "%Y-%m-%d %H:%M:%S")
          stn.dt$time <- as.character(stn.dt$time)

          first.date <- as.Date(sprintf("%d-01-01", start_year))
          min.date <- min(stn.dt$date)
          if(min.date > first.date ) {
            sdate <- as.Date(min.date)
          } else {
            sdate <- as.Date(first.date)
          }

          last.date <- as.Date(sprintf("%d-12-31", end_year))
          max.date <- max(stn.dt$date)
          if(max.date < last.date ) {
            edate <- as.Date(max.date)
          } else {
            edate <- as.Date(last.date)
          }

          stime <- as.POSIXct(sprintf("%s 01:00:00", sdate))
          etime <- as.POSIXct(sprintf("%s 01:00:00", edate))
          time.ref <- seq(from = stime, to = etime, by = "hour")
          time.ref <- as.data.table(format(time.ref, format = "%Y-%m-%d %H:%M:%S"))
          setnames(time.ref, old = names(time.ref), new = "time")

          stn.out <- dplyr::left_join(time.ref, stn.dt)
          stn.out$id <- stnid
          stn.out <- setDT(stn.out)
          stn.out <- stn.out[, c("date", "hr", "id") := NULL]

        }
        if(stype == "msm"){
          stn.dt <- setorder(stn.dt, date, hr)
          stn.dt <- stn.dt[, bat := NULL]

          cols_to_check <- 3:ncol(stn.dt)
          cols_to_remove <- names(stn.dt)[cols_to_check][sapply(stn.dt[, ..cols_to_check], function(x) sum(x) == 0)]
          stn.dt <- as.data.table(stn.dt[, !cols_to_remove, with = FALSE])

          stn.dt$hr <- stn.dt$hr + 1
          stn.dt$time <- as.POSIXct(sprintf("%s %02d:00:00", stn.dt$date, stn.dt$hr))
          stn.dt$time <- format(stn.dt$time, format = "%Y-%m-%d %H:%M:%S")
          stn.dt$time <- as.character(stn.dt$time)

          first.date <- as.Date(sprintf("%d-01-01", start_year))
          min.date <- min(stn.dt$date)
          if(min.date > first.date ) {
            sdate <- as.Date(min.date)
          } else {
            sdate <- as.Date(first.date)
          }

          last.date <- as.Date(sprintf("%d-12-31", end_year))
          max.date <- max(stn.dt$date)
          if(max.date < last.date ) {
            edate <- as.Date(max.date)
          } else {
            edate <- as.Date(last.date)
          }

          stime <- as.POSIXct(sprintf("%s 01:00:00", sdate))
          etime <- as.POSIXct(sprintf("%s 01:00:00", edate))
          time.ref <- seq(from = stime, to = etime, by = "hour")
          time.ref <- as.data.table(format(time.ref, format = "%Y-%m-%d %H:%M:%S"))
          setnames(time.ref, old = names(time.ref), new = "time")

          stn.out <- dplyr::left_join(time.ref, stn.dt)

          stn.out$id <- stnid
          stn.out <- setDT(stn.out)
          stn.out <- stn.out[, c("date", "hr", "id") := NULL]
        }
        if(stype == "pfs"){
          stn.dt <- setorder(stn.dt, date)
          stn.dt <- stn.dt[, bat := NULL]

          first.date <- as.Date(sprintf("%d-01-01", start_year))
          min.date <- min(stn.dt$date)
          if(min.date > first.date ) {
            sdate <- as.Date(min.date)
          } else {
            sdate <- as.Date(first.date)
          }

          last.date <- as.Date(sprintf("%d-12-31", end_year))
          max.date <- max(stn.dt$date)
          if(max.date < last.date ) {
            edate <- as.Date(max.date)
          } else {
            edate <- as.Date(last.date)
          }

          time.ref <- seq(from = sdate, to = edate, by = "day")
          time.ref <- as.data.table(format(time.ref, format = "%Y-%m-%d"))
          setnames(time.ref, old = names(time.ref), new = "date")

          stn.dt$date <- as.character(stn.dt$date)
          stn.out <- dplyr::left_join(time.ref, stn.dt)
          stn.out$id <- stnid
          stn.out <- setDT(stn.out)
          stn.out <- stn.out[, c("id", "flow.sen") := NULL]
        }
        if(stype == "rfm"){
          stn.dt <- setorder(stn.dt, date)
          stn.dt <- setDT(stn.dt)[, c("bat") := NULL]

          first.date <- as.Date(sprintf("%d-01-01", start_year))
          min.date <- min(stn.dt$date)
          if(min.date > first.date ) {
            sdate <- as.Date(min.date)
          } else {
            sdate <- as.Date(first.date)
          }

          last.date <- as.Date(sprintf("%d-12-31", end_year))
          max.date <- max(stn.dt$date)
          if(max.date < last.date ) {
            edate <- as.Date(max.date)
          } else {
            edate <- as.Date(last.date)
          }

          stime <- as.POSIXct(sprintf("%s 01:00:00", sdate))
          etime <- as.POSIXct(sprintf("%s 23:55:00", edate))
          time.ref <- seq(from = stime, to = etime, by = "5 min")
          time.ref <- as.data.table(format(time.ref, format = "%Y-%m-%d %H:%M:%S"))
          colnames(time.ref) <-"time"

          # https://blog.naver.com/regenesis90/222357629320
          vel.long <- stn.dt %>%
            tidyr::pivot_longer(cols = starts_with('vel'), names_to = 'minute', values_to = 'velocity') %>%
            mutate(minute = as.numeric(gsub("vel", "", minute))) %>%
            mutate(time = as.POSIXct(paste(date, sprintf("%02d:%02d:00", hr, minute)), format = "%Y-%m-%d %H:%M:%S")) %>%
            dplyr::select(time, velocity) %>%
            arrange(time)

          dep.long <- stn.dt %>%
            tidyr::pivot_longer(cols = starts_with('dep'), names_to = 'minute', values_to = 'depth') %>%
            mutate(minute = as.numeric(gsub("dep", "", minute))) %>%
            mutate(time = as.POSIXct(paste(date, sprintf("%02d:%02d:00", hr, minute)), format = "%Y-%m-%d %H:%M:%S")) %>%
            dplyr::select(time, depth, h.ref) %>%
            arrange(time)

          vel.long <- as.data.table(vel.long)
          dep.long <- as.data.table(dep.long)
          vel.long$time <- format(vel.long$time, format = "%Y-%m-%d %H:%M:%S")
          dep.long$time <- format(dep.long$time, format = "%Y-%m-%d %H:%M:%S")

          stn.out <- dplyr::left_join(time.ref, vel.long, by="time")
          stn.out <- dplyr::left_join(stn.out, dep.long, by="time")

        }
        if(stype == "uwl"){
          stn.dt <- setorder(stn.dt, date)
          stn.dt <- stn.dt[, c("bat", "dep.avg") := NULL]

          first.date <- as.Date(sprintf("%d-01-01", start_year))
          min.date <- min(stn.dt$date)
          if(min.date > first.date ) {
            sdate <- as.Date(min.date)
          } else {
            sdate <- as.Date(first.date)
          }

          last.date <- as.Date(sprintf("%d-12-31", end_year))
          max.date <- max(stn.dt$date)
          if(max.date < last.date ) {
            edate <- as.Date(max.date)
          } else {
            edate <- as.Date(last.date)
          }

          stime <- as.POSIXct(sprintf("%s 00:05:00", sdate))
          etime <- as.POSIXct(sprintf("%s 23:55:00", edate))
          time.ref <- seq(from = stime, to = etime, by = "5 min")
          time.ref <- as.data.table(format(time.ref, format = "%Y-%m-%d %H:%M:%S"))
          colnames(time.ref) <-"time"

          # https://blog.naver.com/regenesis90/222357629320
          dep.long <- stn.dt %>%
            tidyr::pivot_longer(cols = starts_with('dep'), names_to = 'minute', values_to = 'depth') %>%
            mutate(minute = as.numeric(gsub("dep", "", minute))) %>%
            mutate(time = as.POSIXct(paste(date, sprintf("%02d:%02d:00", hr, minute)), format = "%Y-%m-%d %H:%M:%S")) %>%
            dplyr::select(time, depth, h.ref) %>%
            arrange(time)

          dep.long$time <- format(dep.long$time, format = "%Y-%m-%d %H:%M:%S")

          stn.out <- dplyr::left_join(time.ref, dep.long, by="time")

        }

        OldDFiles <- list.files(stn.dir, pattern = glob2rx(sprintf("%s-%s-%s_*.csv", stype, stnid, gsub("-", "", sdate))), full.names = T)
        file.remove(OldDFiles)

        OutDFile <- file.path(stn.dir, sprintf("%s-%s-%s_%s.csv", stype, stnid, gsub("-", "", sdate), gsub("-", "", edate)))
        fwrite(stn.out, OutDFile, na="")

      } # Station IDs

    } # If sensor type is all

  } # Sensor types


}