Newer
Older
# coding=utf-8
from bscearth.utils.date import parse_date, add_months
from bscearth.utils.log import Log
from earthdiagnostics.diagnostic import Diagnostic, DiagnosticVariableOption, DiagnosticDomainOption, \
DiagnosticIntOption, DiagnosticFloatOption
from earthdiagnostics.utils import Utils, TempFile
from earthdiagnostics.variable_type import VariableType
import numpy as np
import iris
from iris.cube import Cube
import iris.coord_categorisation
from iris.time import PartialDateTime
import iris.exceptions
import iris.coords
import math
import psutil
import six
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
class Discretize(Diagnostic):
"""
Discretizes a variable
:param data_manager: data management object
:type data_manager: DataManager
:param variable: variable to average
:type variable: str
"""
alias = 'discretize'
"Diagnostic alias for the configuration file"
Percentiles = np.array([0.1, 0.25, 0.33, 0.5, 0.66, 0.75, 0.9])
def __init__(self, data_manager, startdate, domain, variable, num_bins, min_value, max_value):
Diagnostic.__init__(self, data_manager)
self.startdate = startdate
self.variable = variable
self.domain = domain
self.realizations = None
self.num_bins = num_bins
self._bins = None
self.cmor_var = data_manager.variable_list.get_variable(variable, silent=True)
if not math.isnan(min_value):
self.min_value = min_value
self.check_min_value = False
elif self.cmor_var and self.cmor_var.valid_min:
self.min_value = float(self.cmor_var.valid_min)
self.check_min_value = False
else:
self.min_value = None
self.check_min_value = True
if not math.isnan(max_value):
self.max_value = max_value
self.check_max_value = False
elif self.cmor_var and self.cmor_var.valid_min:
self.max_value = float(self.cmor_var.valid_max)
self.check_max_value = False
else:
self.max_value = None
self.check_max_value = True
self.process = psutil.Process()
def print_memory_used(self):
Log.debug('Memory: {0:.2f} GB'.format(self.process.memory_info().rss / 1024.0**3))
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
@property
def bins(self):
if self._bins is None:
return self.num_bins
return self._bins
@bins.setter
def bins(self, value):
self._bins = value
def __eq__(self, other):
return self.domain == other.domain and self.variable == other.variable and self.num_bins == other.num_bins and \
self.min_value == other.min_value and self.max_value == other.max_value and \
self.startdate == other.startdate
def __str__(self):
return 'Discretizing variable: {0.domain}:{0.variable} Startdate: {0.startdate} ' \
'Bins: {0.num_bins} Range: [{0.min_value}, {0.max_value}]'.format(self)
@classmethod
def generate_jobs(cls, diags, options):
"""
Creates a job for each chunk to compute the diagnostic
:param diags: Diagnostics manager class
:type diags: Diags
:param options: domain, variable, percentil number, maximum depth (level)
:type options: list[str]
:return:
"""
options_available = (DiagnosticDomainOption(),
DiagnosticVariableOption(),
DiagnosticIntOption('bins', 2000),
DiagnosticFloatOption('min_value', float('nan')),
DiagnosticFloatOption('max_value', float('nan')),
)
options = cls.process_options(options, options_available)
job_list = list()
for startdate in diags.config.experiment.startdates:
job_list.append(Discretize(diags.data_manager, startdate, options['domain'], options['variable'],
options['bins'], options['min_value'], options['max_value']))
return job_list
def request_data(self):
self.original_data = self.request_chunk(self.domain, self.variable, self.startdate, None, None)
def declare_data_generated(self):
var_name = '{0.variable}_dis'.format(self)
self.discretized_data = self.declare_chunk(self.domain, var_name, self.startdate, None, None,
vartype=VariableType.STATISTIC)
def compute(self):
"""
Runs the diagnostic
"""
self.print_memory_used()
iris.FUTURE.netcdf_promote = True
self._load_cube()
self.print_memory_used()
self.print_memory_used()
Log.info('Range: [{0}, {1}]', self.min_value, self.max_value)
self._get_distribution()
self.print_memory_used()
self.print_memory_used()
del self.distribution
del self.data_cube
self.print_memory_used()
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
def _load_cube(self):
handler = Utils.openCdf(self.original_data.local_file)
if 'realization' in handler.variables:
handler.variables[self.variable].coordinates = 'realization'
handler.close()
data_cube = iris.load_cube(self.original_data.local_file)
date = parse_date(self.startdate)
lead_date = add_months(date, 1, self.data_manager.config.experiment.calendar)
leadtimes = {1: PartialDateTime(lead_date.year, lead_date.month, lead_date.day)}
def assign_leadtime(coord, x):
leadtime_month = 1
partial_date = leadtimes[leadtime_month]
while coord.units.num2date(x) >= partial_date:
leadtime_month += 1
try:
partial_date = leadtimes[leadtime_month]
except KeyError:
new_date = add_months(date, leadtime_month, self.data_manager.config.experiment.calendar)
partial_date = PartialDateTime(new_date.year, new_date.month, new_date.day)
leadtimes[leadtime_month] = partial_date
return leadtime_month
iris.coord_categorisation.add_categorised_coord(data_cube, 'leadtime', 'time', assign_leadtime)
self.data_cube = data_cube
def _save_results(self):
Log.debug('Saving results...')
bins = np.zeros(self.num_bins)
bins_bounds = np.zeros((self.num_bins, 2))
for x in range(self.num_bins):
bins[x] = (self.bins[x+1] - self.bins[x]) / 2 + self.bins[x]
bins_bounds[x, 0] = self.bins[x]
bins_bounds[x, 1] = self.bins[x+1]
bins_coord = iris.coords.DimCoord(bins, var_name='bin', units=self.data_cube.units, bounds=bins_bounds)
cubes = iris.cube.CubeList()
for leadtime, distribution in six.iteritems(self.distribution):
leadtime_cube = Cube(distribution.astype(np.uint32), var_name=self.data_cube.var_name,
standard_name=self.data_cube.standard_name, units='1')
leadtime_cube.add_dim_coord(bins_coord, 0)
leadtime_cube.add_dim_coord(self.data_cube.coord('latitude'), 1)
leadtime_cube.add_dim_coord(self.data_cube.coord('longitude'), 2)
leadtime_cube.add_aux_coord(iris.coords.AuxCoord(np.array((leadtime,), np.int8), var_name='leadtime',
units='months'))
cubes.append(leadtime_cube)
temp = TempFile.get()
iris.FUTURE.netcdf_no_unlimited = True
iris.save(cubes.merge_cube(), temp, zlib=True)
self.discretized_data.set_local_file(temp, rename_var=self.data_cube.var_name)
def _get_distribution(self):
self.distribution = {}
Log.debug('Discretizing...')
for leadtime in set(self.data_cube.coord('leadtime').points):
Log.debug('Discretizing leadtime {0}', leadtime)
leadtime_cube = self.data_cube.extract(iris.Constraint(leadtime=leadtime))
for realization in self.data_cube.coord('realization').points:
Log.debug('Discretizing realization {0}', realization)
self.print_memory_used()
try:
realization_cube = leadtime_cube.extract(iris.Constraint(realization=realization))
except iris.exceptions.CoordinateNotFoundError:
realization_cube = leadtime_cube
if realization_cube is None and realization == 0:
realization_cube = leadtime_cube
if leadtime not in self.distribution:
self.distribution[leadtime] = self._calculate_distribution(realization_cube)
else:
self.distribution[leadtime] += self._calculate_distribution(realization_cube)
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
def _get_value_interval(self):
if self.check_min_value or self.check_max_value:
Log.debug('Calculating max and min values...')
for time_slice in self.data_cube.slices_over('time'):
if self.check_min_value:
file_min = np.amin(time_slice.data)
if self.min_value is None:
self.min_value = file_min
self.min_value = min(self.min_value, file_min)
if self.check_max_value:
file_max = np.amax(time_slice.data)
self.max_value = max(self.max_value, file_max)
def _calculate_distribution(self, data_cube):
def calculate_histogram(time_series):
histogram, self.bins = np.histogram(time_series, bins=self.bins,
range=(self.min_value, self.max_value))
return histogram
return np.apply_along_axis(calculate_histogram, 0, data_cube.data)