'''
Mask Class to hold the mask information and operation on mask
@author: Mr. Srikanth Nagella
'''
# pylint: disable=E1103
# pylint: disable=no-name-in-module
import numpy as np
from netCDF4 import Dataset
import logging
from scipy import ndimage
import matplotlib.pyplot as plt
from pynemo.utils import gcoms_break_depth
from PyQt4.QtGui import QMessageBox
from PyQt4 import QtCore
class Mask(object):
"""This is a Mask holder. which reads from a netCDF bathymetry file and
stores it in 'data' member variable"""
min_depth = 200.0
shelfbreak_dist = 200.0
mask_type = 0
def __init__(self, bathymetry_file=None, mask_file=None, min_depth = 200.0, shelfbreak_dist = 200.0):
"""Initialises the Mask data"""
self.data = None
self.bathy_data = None
self.mask_file = None
self.logger = logging.getLogger(__name__)
self.set_mask_file(mask_file)
self.set_bathymetry_file(bathymetry_file)
self.min_depth = min_depth
self.shelfbreak_dist = shelfbreak_dist
def set_mask_file(self, mask_file):
"""
Reads the mask data from the mask file
Assumes the mask file is 2D
"""
self.mask_file = mask_file
#if mask file is not set then reset the data
if self.mask_file == None:
self.data = None
return
try:
mask_nc = Dataset(str(self.mask_file), mode="r")
data = mask_nc.variables['mask']
self.data = data[:,:]
except KeyError:
self.logger.error('Mask file missing have mask variable')
raise
except (IOError, RuntimeError):
self.logger.error('Cannot open mask file '+self.mask_file)
self.data = None
raise
def set_bathymetry_file(self, bathy_file):
""" This reads the bathymetry file and sets the land to 0 and ocean to 1 """
if bathy_file == None:
return
try:
self.bathymetry_file = str(bathy_file)
#open the bathymetry file
self.bathy_nc = Dataset(self.bathymetry_file)
self.lon = np.asarray(self.bathy_nc.variables['nav_lon'])
self.lat = np.asarray(self.bathy_nc.variables['nav_lat'])
self.bathy_data = self.bathy_nc.variables['Bathymetry'][:,:]
try: #check if units exists otherwise unknown. TODO
self.data_units = self.bathy_nc.variables['Bathymetry'].units
except AttributeError:
self.data_units = "unknown"
if self.data is None:
self.data = self.bathy_nc.variables['Bathymetry']
self.data = np.asarray(self.data[:, :])
self.data = np.around((self.data + .5).clip(0, 1))
#apply default 1px border
self.apply_border_mask(1)
except KeyError:
self.logger.error('Bathymetry file does not have Bathymetry variable')
raise
except (IOError, RuntimeError):
self.logger.error('Cannot open bathymetry file '+self.bathymetry_file)
raise
def save_mask(self, mask_file):
"""Reads the mask data from the mask file"""
if mask_file == None:
mask_file = self.mask_file
try:
self.logger.info('Writing mask data to %s' % mask_file)
msgbox = QMessageBox()
msgbox.setWindowTitle("Saving....")
msgbox.setText("Writing mask data to file, please wait...")
msgbox.setWindowModality(QtCore.Qt.NonModal)
msgbox.show()
mask_nc = Dataset(str(mask_file), mode="w")
mask_nc.createDimension('y', size=len(self.bathy_nc.dimensions['y']))
mask_nc.createDimension('x', size=len(self.bathy_nc.dimensions['x']))
nav_lat = mask_nc.createVariable('nav_lat', 'f4', ('y','x',))
nav_lon = mask_nc.createVariable('nav_lon', 'f4', ('y','x',))
mask_var = mask_nc.createVariable('mask', 'f4', ('y','x',))
mask_var[...] = self.data
nav_lat[...] = self.lat
nav_lon[...] = self.lon
msgbox.close()
except (IOError, RuntimeError):
QMessageBox.information(None,"pyNEMO","Failed to write the mask file, please check the permissions")
self.logger.info('Cannot open mask file for writing '+mask_file)
raise
def apply_border_mask(self, pixels):
""" pixels is number of pixels in the border that need applying mask"""
if self.data is not None and pixels < self.data.shape[0] and pixels < self.data.shape[1]:
if pixels != 0:
tmp = np.ones(self.data.shape, dtype=bool)
tmp[pixels:-pixels, pixels:-pixels] = False
else:
tmp = np.zeros(self.data.shape, dtype=bool)
self.reset_mask()
self.data = self.data * -1
self.data[tmp] = -1
self.select_the_largest_region()
def add_mask(self, index, roi):
""" Adds the masks for the given index values depending on the type of mask selected"""
out_index = None
if self.mask_type == None or self.mask_type == 0:
out_index = index
elif self.mask_type == 1: #maximum depth
out_index = self._get_bathy_depth_index(index,self.min_depth)
out_index = self.remove_small_regions(out_index)
out_index = self.fill_small_regions(out_index)
elif self.mask_type == 2: # shelf break
#dummy, shelf_break = gcoms_break_depth.gcoms_break_depth(self.bathy_data[index])
#out_index = self._get_bathy_depth_index(index, shelf_break)
out_index = gcoms_break_depth.polcoms_select_domain(self.bathy_data, self.lat,
self.lon, roi, self.shelfbreak_dist)
out_index = np.logical_and(index, out_index)
out_index = self.remove_small_regions(out_index)
#out_index = self.fill_small_regions(out_index)
#if index is not empty
if out_index is not None:
tmp = self.data[out_index]
tmp[tmp == -1] = 1
self.data[out_index] = tmp
self.select_the_largest_region()
def _get_bathy_depth_index(self, index, depth):
""" returns the indices from the input field `index` which have bathymetry depth greater
than the input field `depth`
"""
output_index = self.bathy_data < depth
output_index = np.logical_and(index,output_index)
return output_index
def remove_mask(self,index,roi):
""" Removes the mask for the given index values depending on the type of mask selected """
out_index = None
if self.mask_type == None or self.mask_type == 0:
out_index = index
elif self.mask_type == 1: #minimum depth
out_index = self._get_bathy_depth_index(index,self.min_depth)
out_index = self.remove_small_regions(out_index)
out_index = self.fill_small_regions(out_index)
elif self.mask_type == 2: # shelf break
# dummy, shelf_break = gcoms_break_depth.gcoms_break_depth(self.bathy_data[index])
# out_index = self._get_bathy_depth_index(index, shelf_break)
out_index = gcoms_break_depth.polcoms_select_domain(self.bathy_data, self.lat,
self.lon, roi, self.shelfbreak_dist)
out_index = np.logical_and(index, out_index)
out_index = self.remove_small_regions(out_index)
#out_index = self.fill_small_regions(out_index)
tmp = self.data[out_index]
tmp[tmp == 1] = -1
self.data[out_index] = tmp
self.select_the_largest_region()
def set_minimum_depth_mask(self, depth):
self.min_depth = depth
def set_mask_type(self, mask_type):
""" Sets the mask type """
self.mask_type = mask_type
def remove_small_regions(self, index):
""" Removes the small regions in the selection area and takes only the largest area
for mask"""
#prepare the regions
mask_data = np.zeros(self.data.shape)
mask_data[index] = self.data[index]
#connected components
label_mask, num_labels = ndimage.label(mask_data)
mean_values = ndimage.sum(np.ones(self.data.shape),label_mask,range(1, num_labels+1))
max_area_mask = None
if mean_values.size != 0:
max_area_index = np.argmax(mean_values)+1
max_area_mask = label_mask == max_area_index
return max_area_mask
def fill_small_regions(self, index):
""" This method fills the small regions of the selection area and fills them up """
#prepare the region with mask and land as 0, ocean as 1
mask_data = np.zeros(self.data.shape)
mask_data[index] = 1
#remove the small unmask holes
mask_withoutholes = ndimage.binary_fill_holes(mask_data)
return np.where(mask_withoutholes==1)
def reset_mask(self):
""" This method resets the data back to no mask with land fill """
self.data = np.around((self.bathy_data + .5).clip(0, 1))*-1
def select_the_largest_region(self):
""" This method tidies up the mask by selecting the largest masked region. this is to avoid two disconnected masked regions """
mask_data = np.zeros(self.data.shape)
mask_data[:,:] = self.data[:,:]
mask_data[mask_data == -1] = 0
#connected components
label_mask, num_labels = ndimage.label(mask_data)
if num_labels == 0: #if mask is empty/clear
return
mean_values = ndimage.sum(np.ones(self.data.shape),label_mask,range(1, num_labels+1))
max_area_mask = None
if mean_values.size != 0:
max_area_index = np.argmax(mean_values)+1
max_area_mask = label_mask == max_area_index
self.data = np.around((self.bathy_data + .5).clip(0, 1))
self.data[self.data == 1] = -1
self.data[max_area_mask] = self.data[max_area_mask] * -1
def apply_mediterrian_mask(self):
""" This is mediterrian mask specific for the test bathymetry file """
tmp = self.data[0:59, 280:350]
tmp[tmp == 1] = -1
self.data[0:59, 280:350] = tmp