# read the orginal fldas data and save to .RDS file.

#install.packages('ncdf4')
require(ncdf4)
source('GetReady.R')
source('Rfunction/Fun.readnc.R')

fl=readOGR(file.path(dir.predata, 'FLDAS_GCS.shp'))@data
head(fl)
# years=2017:2018
dirs = file.path(dir.fldas, years)

ndir = length(dirs)
fn=list.files(dir.fldas, pattern=glob2rx('*.nc'), recursive = T, full.names = T)[1]

fid=nc_open(fn)
xloc = round(fid$dim$X$vals,2)
yloc = round(fid$dim$Y$vals, 2)
nx=length(xloc)
ny = length(yloc)

#===================================================
xc = fl[,'xcenter']
yc = fl[,'ycenter']

xid = match(round(xc, 2), xloc)
yid = match(round(yc,2), yloc)
xyid=cbind(xid,yid)

sn = paste0('X',xc*100, 'Y', yc*100)
ns = length(sn)

vns = names(fid$var)
vns = vns[!(vns %in% 'time_bnds')] # don't need the time_bnds

mat=ncvar_get(fid, vns[2])
mat=mat*0+1
png.control(fn=paste0('Rawdata','_FLDAS_location.png'), path = file.path(dir.png), ratio=1)
image(xloc, yloc, mat, xlab='Lon', ylab='Lat', main='Coverage of FLDAS') ; grid()
points(xloc[xid], yloc[yid], col=3)
plot(wbd.gcs, add=T)
dev.off()

nv=length(vns)
for(idd  in 1:ndir){ # for each year dir
# library(foreach)
# library(doMC)
# library(doParallel)
# registerDoMC(4)
# foreach (idd = 1:ndir) %dopar%{
  cdir <- dirs[idd]
  fns = list.files(cdir, pattern=glob2rx('*.nc'), recursive = T, full.names = T)
  nf = length(fns)
  x.arr = array(0, dim=c(ns, nv, nf) )
  x.t= character(nf)
  for(j  in 1:nf){  # files in each year
    fn=fns[j]
    t=substr(basename(fn), 22, 29)
    message(j, '/', nf, '\t', t)
    x.mat = readnc(fn, xyid=xyid, vns=vns)
    x.t[j] = t
    x.arr[,,j ] = x.mat 
  }
  dimnames(x.arr) = list(sn, vns,  x.t)
  fn.rds = file.path(dir.predata, paste0(prjname,'-', basename(cdir), '.RDS'))
  saveRDS(x.arr, file=fn.rds)
}