# coding=utf-8

import os

import Ngl
import numpy as np
import numpy.ma as ma

from netCDF4 import Dataset

from livvkit.util import elements as el


describe = """
Average of the annual average surface mass balance (kg m^-2 a^-1) over Greenland for every 
summer (June-July-August; JJA) from 1980--1999 for CESM (left) and RACMO 2.3 (middle), 
and the difference between them (right; CESM - RACMO 2.3). The black solid lines denote 
the 0, 1000, 2000, and 3000 meter Greenland ice sheet elevation contours as seen by 
the models (CESM's contours shown in the difference plot).
"""

title = "Summer surface mass balance"


def make_plot(config, out_path='.'):
    # ---------------- Data source in TITAN ------------------------
    f_cism = os.path.join(config['cism_data'], 'Greenland_5km_v1.1_SacksRev_c110629.nc')
    f_model_lnd_climo_jja = os.path.join(config['model_lnd_climos'],
                                        'b.e10.BG20TRCN.f09_g16.002_JJA_climo.nc')
    f_racmo_smb_jja = os.path.join(config['racmo_data'], 'racmo23_GRN_monthly.smb.1980-1999.JJA.nc')
    f_racmo_smb_jja_remapped = os.path.join(config['racmo_data'],
                                            'racmo23_GRN_monthly.smb.1980-1999.remap2cesm.JJA.nc')
    f_racmo_mask = os.path.join(config['racmo_data'], 'RACMO23_masks_ZGRN11.nc')
    # --------------------------------------------------------------

    img_list = []

    # f_cism get the following, matrixsize[301,561]
    # usrf(0,:,:), lat(0,:,:), lon(0,:,:)
    # Goal: plot 3 figures, radiation overlaying on elevation usrf

    # read f_cism, elevation of 5km CESM
    # adding the [:] transform the variable to type numpy.ndarray
    ncid0 = Dataset(f_cism)
    usrf = ncid0.variables['usrf'][0, :, :]
    lat = ncid0.variables['lat'][0, :, :]
    lon = ncid0.variables['lon'][0, :, :]

    # read f_model_lnd_climo_jja, CESM variable
    ncid1 = Dataset(f_model_lnd_climo_jja)
    smb_model = ncid1.variables['QICE'][0, :, :]
    smb_model = smb_model * 3.156e7  # convert from mm/s to kg/m^2/yr
    lat1 = ncid1.variables['lat'][:]
    lon1 = ncid1.variables['lon'][:]

    # use gris as a mask to mask smb array
    gris_mask = ncid1.variables['gris_mask'][0, :, :]
    gris_mask = ma.masked_equal(gris_mask, 0)

    smb_model_mask = ma.masked_array(smb_model, mask=gris_mask.mask)
    smb_model = smb_model_mask

    # read f_racmo_smb_jja_remapped, the remapped RACMO file to calculate difference
    ncid3 = Dataset(f_racmo_smb_jja_remapped)
    smb_remap = ncid3.variables['smb'][0, :, :]
    smb_remap = smb_remap

    smb_remap_mask = ma.masked_array(smb_remap, mask=gris_mask.mask)
    smb_remap = smb_remap_mask

    # read f_racmo_smb_jja, RACMO variable
    ncid2 = Dataset(f_racmo_smb_jja)
    smb = ncid2.variables['smb'][0, :, :]
    # lat2 = ncid2.variables['lat'][:]
    # lon2 = ncid2.variables['lon'][:]
    smb = smb

    # read f_racmo_mask, the lat/lon for RACMO data
    ncid4 = Dataset(f_racmo_mask)
    lat4 = ncid4.variables['lat'][:]
    lon4 = ncid4.variables['lon'][:]
    racmo_elev = ncid4.variables['Topography'][0, 0, :, :]

    # Use gris as a mask to mask smb array.
    gris_mask = ncid4.variables['GrIS_mask'][:]
    gris_mask = ma.masked_equal(gris_mask, 0)

    smb_mask = ma.masked_array(smb, mask=gris_mask.mask)
    smb = smb_mask

    diff = smb_model - smb_remap

    # diffmin = (np.min(diff))
    # diffmax = (np.max(diff))
    # print("Min ANN SMB anomaly: {}".format(diffmin))
    # print("Max ANN SMB anomaly: {}".format(diffmax))

    # ------- PLOT --------
    #  Open a workstation for drawing the plots
    wkres = Ngl.Resources()
    # wkres.wkColorMap = "matlab_jet"
    wkres.wkColorMap = "BlueWhiteOrangeRed"
    # wkres.wkOrientation = "portrait"  # "portrait" or "landscape"
    wks_type = "png"
    wks_img = str(os.path.join(out_path, "CESM_RACMO23_smb_ANN"))
    wks = Ngl.open_wks(wks_type, wks_img, wkres)

    # --- for the map -------
    # Define plotting area, Greenland
    mres = Ngl.Resources()
    mres.nglDraw = False  # Don't draw individual plots
    mres.nglFrame = False  # Don't advance frame.

    mres.pmTickMarkDisplayMode = "Never"  # Turn off map tickmarks.
    mres.mpGridAndLimbOn = False  # Turn off grid and limb lines.
    mres.mpProjection = "Aitoff"
    mres.mpLimitMode = "LatLon"  # limit map via lat/lon, to zoom in
    mres.mpCenterLatF = 70.  # map area
    mres.mpCenterLonF = -44.
    mres.mpMinLatF = 57.
    mres.mpMaxLatF = 85.
    mres.mpMinLonF = -55.
    mres.mpMaxLonF = -30.
    mres.mpOutlineOn = False
    mres.mpFillOn = False
    mres.mpPerimOn = True  # add box around map

    # --- for the smbedo contour of CESM -------
    res1 = Ngl.Resources()
    res1.nglDraw = False  # Don't draw individual plots
    res1.nglFrame = False  # Don't advance frame.
    res1.cnLineLabelsOn = False
    res1.cnFillOn = True
    res1.cnLinesOn = False
    res1.cnLineLabelsOn = False
    res1.cnFillMode = "RasterFill"
    res1.cnLevelSelectionMode = "ExplicitLevels"
    res1.cnLevels = [-4000, -2000, -1000, -700, -500, -300, -200, -100, -50, -20, 20, 50, 100, 200, 300, 500, 700, 1000,
                     2000, 5000]
    res1.lbLabelBarOn = True  # Turn on labelbar.
    res1.lbLabelFontHeightF = 0.04
    # res1.pmLabelBarOrthogonalPosF = -0.05

    res1.sfXArray = lon1
    res1.sfYArray = lat1

    # --- for the smbedo contour of RACMO -------
    res2 = Ngl.Resources()
    res2.nglDraw = False  # Don't draw individual plots
    res2.nglFrame = False  # Don't advance frame.
    res2.cnLineLabelsOn = False
    res2.cnFillOn = True
    res2.cnLinesOn = False
    res2.cnLineLabelsOn = False
    res2.cnFillMode = "RasterFill"
    res2.cnLevelSelectionMode = "ExplicitLevels"
    res2.cnLevels = [-4000, -2000, -1000, -700, -500, -300, -200, -100, -50, -20, 20, 50, 100, 200, 300, 500, 700, 1000,
                     2000, 5000]
    # res2.cnLevels = [-12800,-6400,-3200,-1600,-800,-400,-200,-50,-25,0,25,50,100,200,400,800,1600,3200,6400]
    res2.lbLabelBarOn = True  # Turn on labelbar.
    res2.lbOrientation = "Vertical"  # Verticle labelbar
    # res2.pmLabelBarHeightF    = 0.5         # Change height of labelbar
    # res2.pmLabelBarWidthF     = 0.2           # Change width of labelbar
    # res2.pmLabelBarOrthogonalPosF = -0.02      # Move labelbar closer to plot
    res2.lbLabelFontHeightF = 0.04  # Make fonts smaller.

    res2.sfXArray = lon4
    res2.sfYArray = lat4

    # --- for the smb contour of CESM-RACMO -------
    res3 = Ngl.Resources()
    res3.nglDraw = False  # Don't draw individual plots
    res3.nglFrame = False  # Don't advance frame.
    res3.cnLineLabelsOn = False
    res3.cnFillOn = True
    res3.cnLinesOn = False
    res3.cnLineLabelsOn = False
    res3.cnFillMode = "RasterFill"
    res3.cnLevelSelectionMode = "ExplicitLevels"
    res3.cnLevels = [-4000, -2000, -1000, -700, -500, -300, -200, -100, -50, -20, 20, 50, 100, 200, 300, 500, 700, 1000,
                     2000, 5000]

    # res3.cnLevels = np.arange(-1000, 1000, 200)
    res3.lbLabelBarOn = True  # Turn on labelbar.
    res3.lbOrientation = "Vertical"  # Verticle labelbar
    res3.lbLabelFontHeightF = 0.04  # Make fonts smaller.

    res3.sfXArray = lon1
    res3.sfYArray = lat1

    # ---- for the elevation -------
    sres = Ngl.Resources()
    sres.nglDraw = False  # Don't draw individual plots
    sres.nglFrame = False  # Don't advance frame.
    sres.cnFillOn = False
    sres.cnLinesOn = True
    sres.cnLineLabelsOn = False
    sres.cnLevelSelectionMode = "ExplicitLevels"
    sres.cnLevels = [0, 1000, 2000, 3000]
    sres.sfXArray = lon
    sres.sfYArray = lat

    # ---- for the elevation of RACMO -------
    sres1 = Ngl.Resources()
    sres1.nglDraw = False  # Don't draw individual plots
    sres1.nglFrame = False  # Don't advance frame.
    sres1.cnFillOn = False
    sres1.cnLinesOn = True
    sres1.cnLineLabelsOn = False
    sres1.cnLevelSelectionMode = "ExplicitLevels"
    sres1.cnLevels = [0, 1000, 2000, 3000]
    sres1.sfXArray = lon4
    sres1.sfYArray = lat4

    # ---- Overlay plots, each one has its own ID
    # overlay ice on base map, and then overlay elevation on ice
    # usrf is elevation 5km CESM, racmo_elev is elevation of RACMO
    usrf_plot1 = Ngl.contour(wks, usrf, sres)
    usrf_plot2 = Ngl.contour(wks, racmo_elev, sres1)
    usrf_plot3 = Ngl.contour(wks, usrf, sres)

    smb_model_plot = Ngl.contour(wks, smb_model, res1)
    smb_plot = Ngl.contour(wks, smb, res2)
    diff_plot = Ngl.contour(wks, diff, res3)
    # diff_plot = Ngl.contour(wks,Remap_smb,res1)

    # Creat multiple figures and draw, which now contains the elevation and radiation
    # "[1,3]" indicates 1 row, 3 columns.
    map_title = ["CESM, SMB", "RACMO, SMB", "CESM-RACMO, SMB"]

    nmap = 3
    plot = []
    for i in range(nmap):
        mres.tiMainString = map_title[i]
        plot.append(Ngl.map(wks, mres))

    # Overlay everything on the map plot.
    Ngl.overlay(plot[0], smb_model_plot)
    Ngl.overlay(plot[0], usrf_plot1)

    Ngl.overlay(plot[1], smb_plot)
    Ngl.overlay(plot[1], usrf_plot2)

    Ngl.overlay(plot[2], diff_plot)
    Ngl.overlay(plot[2], usrf_plot3)

    # Using overlay, CANNOT set panel property
    # text_ndc should be before panel
    # textres               = Ngl.Resources()
    # textres.txFontHeightF = 0.02   # Size of title.
    # Ngl.text_ndc(wks,"SMB",0.48,0.85,textres)

    Ngl.panel(wks, plot, [1, nmap])

    img_link = os.path.join(os.path.basename(out_path),
                            os.path.basename(wks_img + '.' + wks_type))
    img_elem = el.image(title,
                        ' '.join(describe.split()),
                        img_link)
    img_elem['Height'] = config['image_height']
    img_list.append(img_elem)

    return img_list