siasiesiv.py

# coding=utf-8
"""Compute the sea ice extent , area and volume  in both hemispheres or a specified region"""
import os
import six

import iris
import iris.analysis
import iris.coords
import iris.util
from bscearth.utils.log import Log

from earthdiagnostics.constants import Basins
from earthdiagnostics.diagnostic import Diagnostic, DiagnosticBasinListOption, DiagnosticBoolOption
from earthdiagnostics.modelingrealm import ModelingRealms
from earthdiagnostics.utils import Utils, TempFile


# noinspection PyUnresolvedReferences


class Siasiesiv(Diagnostic):
    """
    Compute the sea ice extent , area and volume  in both hemispheres or a specified region.

    Parameters
    ----------
    data_manager: DataManager
    startdate: str
    member: int
    chunk: init
    domain: ModellingRealm
    variable: str
    basin: list of Basin
    mask: numpy.array
    omit_vol: bool
    """

    alias = 'siasiesiv'
    "Diagnostic alias for the configuration file"

    e1t = None
    e2t = None
    gphit = None

    def __init__(self, data_manager, startdate, member, chunk, masks, var_manager, data_convention, omit_vol):
        Diagnostic.__init__(self, data_manager)
        self.startdate = startdate
        self.member = member
        self.chunk = chunk
        self.masks = masks
        self.generated = {}
        self.var_manager = var_manager
        self.omit_volume = omit_vol
        self.sic_varname = self.var_manager.get_variable('sic').short_name
        self.sit_varname = self.var_manager.get_variable('sit').short_name
        self.data_convention = data_convention

        self.results = {}
        for var in ('siarean', 'siareas', 'sivoln', 'sivols', 'siextentn', 'siextents'):
            self.results[var] = {}

    def __str__(self):
        return 'Siasiesiv Startdate: {0.startdate} Member: {0.member} Chunk: {0.chunk} ' \
               'Basins: {1} Omit volume: {0.omit_volume}'.format(self,
                                                                 ','.join(str(basin) for basin in self.masks.keys()))

    @classmethod
    def generate_jobs(cls, diags, options):
        """
        Create a job for each chunk to compute the diagnostic

        :param diags: Diagnostics manager class
        :type diags: Diags
        :param options: basin
        :type options: list[str]
        :return:
        """
        options_available = (DiagnosticBasinListOption('basins', [Basins().Global]),
                             DiagnosticBoolOption('omit_volume', False))
        options = cls.process_options(options, options_available)

        basins = options['basins']
        if not basins:
            Log.error('Basins not recognized')
            return ()

        masks = {}
        basins.sort()
        for basin in basins:
            masks[basin] = Utils.get_mask(basin)

        job_list = list()
        for startdate, member, chunk in diags.config.experiment.get_chunk_list():
            job_list.append(Siasiesiv(diags.data_manager, startdate, member, chunk, masks,
                                      diags.config.var_manager, diags.config.data_convention,
                                      options['omit_volume']))

        e1t = iris.load_cube('mesh_hgr.nc', 'e1t')
        e2t = iris.load_cube('mesh_hgr.nc', 'e2t')
        Siasiesiv.area = e1t * e2t

        return job_list

    def request_data(self):
        """Request data required by the diagnostic"""
        if not self.omit_volume:
            self.sit = self.request_chunk(ModelingRealms.seaIce, self.sit_varname,
                                          self.startdate, self.member, self.chunk)
        self.sic = self.request_chunk(ModelingRealms.seaIce, self.sic_varname,
                                      self.startdate, self.member, self.chunk)

    def declare_data_generated(self):
        """Declare data to be generated by the diagnostic"""
        if not self.omit_volume:
            self._declare_var('sivols')
            self._declare_var('sivoln')

        self._declare_var('siareas')
        self._declare_var('siextents')

        self._declare_var('siarean')
        self._declare_var('siextentn')

    def _declare_var(self, var_name):
        self.generated[var_name] = self.declare_chunk(ModelingRealms.seaIce, var_name,
                                                      self.startdate, self.member, self.chunk)

    def compute(self):
        """Run the diagnostic"""
        coordinates = ' '.join(('time', 'leadtime', 'time_centered',
                                self.data_convention.lon_name, self.data_convention.lat_name))
        handler = Utils.open_cdf(self.sic.local_file)
        handler.variables[self.sic_varname].coordinates = coordinates
        handler.close()
        sic = iris.load_cube(self.sic.local_file)
        if sic.units.origin == '%' and sic.data.max() < 2:
            sic.units = '1.0'
        sic.convert_units('1.0')
        sic *= Siasiesiv.area.data
        extent = sic.data > 0.15

        if not self.omit_volume:
            handler = Utils.open_cdf(self.sit.local_file)
            handler.variables[self.sit_varname].coordinates = coordinates
            handler.close()
            sit = iris.load_cube(self.sic.local_file)

        for basin, mask in six.iteritems(self.masks):
            self.results['siarean'][basin] = self.sum(sic, mask, north=True)
            self.results['siareas'][basin] = self.sum(sic, mask, north=False)

            if not self.omit_volume:
                volume = sic * sit
                self.results['sivoln'][basin] = self.sum(volume, mask, north=True)
                self.results['sivols'][basin] = self.sum(volume, mask, north=False)

            extent_mask = mask * extent
            self.results['siextentn'][basin] = self.sum(sic, extent_mask, north=True)
            self.results['siextents'][basin] = self.sum(sic, extent_mask, north=False)

        self.save()

    def sum(self, data, mask, north=True):
        if north:
            condition = data.coord('latitude').points > 0
        else:
            condition = data.coord('latitude').points < 0
        weights = iris.util.broadcast_to_shape(condition, data.shape, data.coord_dims('latitude')) * mask
        return data.collapsed(('latitude', 'longitude'), iris.analysis.SUM, weights=weights)

    def save(self):
        for var, basins in six.iteritems(self.results):
            results = iris.cube.CubeList()
            for basin, result in six.iteritems(basins):
                result.var_name = var
                if var.startswith('sivol'):
                    result.units = 'm^3'
                else:
                    result.units = 'm^2'
                result.add_aux_coord(iris.coords.AuxCoord(basin.name, var_name='region'))
                results.append(result)
            if not results:
                continue
            self._save_file(results.merge_cube(), var)

    def _save_file(self, data, var):
        generated_file = self.generated[var]
        temp = TempFile.get()
        region = data.coord('region').points
        data.remove_coord('region')
        iris.save(data, temp, zlib=True)
        if len(region) > 1:
            Utils.rename_variable(temp, 'dim0', 'region', False)
            handler = Utils.open_cdf(temp)
            var = handler.createVariable('region2', str, ('region',))
            var[...] = region
            handler.close()
            Utils.rename_variable(temp, 'region2', 'region', True)
        else:
            handler = Utils.open_cdf(temp)
            if 'region' not in handler.dimensions:
                new_file = TempFile.get()
                new_handler = Utils.open_cdf(new_file, 'w')

                new_handler.createDimension('region', 1)
                for dimension in handler.dimensions:
                    Utils.copy_dimension(handler, new_handler, dimension)

                for variable in handler.variables.keys():
                    if variable in (var, 'region'):
                        continue
                    Utils.copy_variable(handler, new_handler, variable)
                old_var = handler.variables[var]
                new_var = new_handler.createVariable(var, old_var.dtype, ('region',) + old_var.dimensions,
                                                     zlib=True, fill_value=1.0e20)
                Utils.copy_attributes(new_var, old_var)
                new_var[0, :] = old_var[:]

                new_var = new_handler.createVariable('region', str, ('region',))
                new_var[0] = region[0]

                new_handler.close()
                os.remove(temp)
                temp = new_file
            handler.close()
        generated_file.set_local_file(temp)