-
Notifications
You must be signed in to change notification settings - Fork 3
/
training_data_forecasts.py
552 lines (468 loc) · 22.1 KB
/
training_data_forecasts.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
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
68
69
70
71
72
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
129
130
131
132
133
134
135
136
137
138
139
140
141
142
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
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
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
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
#!/usr/bin/env python3
from __future__ import annotations
import warnings
import datetime
import numpy as np
import xarray as xr
import climetlab as cml
from climetlab import Dataset
from climetlab.normalize import normalize
from climetlab.indexing import PerUrlIndex
from ..config import baseurl
from ..utils import convert_to_datetime
__version__ = "0.1.2"
_terms_of_use = """By downloading data from this dataset, you agree to the terms and conditions defined at
https://github.com/Climdyn/climetlab-eumetnet-postprocessing-benchmark/blob/main/LICENSE
and
https://github.com/Climdyn/climetlab-eumetnet-postprocessing-benchmark/blob/main/DATA_LICENSE
If you do not agree with such terms, do not download the data. """
class TrainingDataForecast(Dataset):
terms_of_use = _terms_of_use
_BASEURL = baseurl
_PATTERN = ""
_ANALYSIS_PATTERN = (
"{url}data/ana/{leveltype}/"
"EU_analysis_{leveltype}_params_{isodate}.grb"
)
_ensemble_alias = ["ensemble", "ens", "proba", "probabilistic"]
def __init__(self, *args, **kwargs):
"""Do almost nothing. To be overridden by the inherithing class."""
self.parameter = list()
self.date = ""
self.leveltype = ""
self.kind = ""
self.obs_source = None
self.isodate = ""
self.level = None
self.year = ""
self.month = ""
self.step = []
def get_observations_as_xarray(self, fcs_kwargs=None, **obs_kwargs):
if fcs_kwargs is None:
fcs_kwargs = dict()
fcs = self.source.to_xarray(**fcs_kwargs)
fcs_valid_time = fcs.valid_time.to_pandas()
fcs_time_list = list(map(convert_to_datetime, fcs_valid_time.iloc[0, :]))
days = dict()
for t in fcs_time_list:
year_month = str(t.year).rjust(4, '0') + str(t.month).rjust(2, '0')
if year_month not in days:
days[year_month] = list()
day = year_month + str(t.day).rjust(2, '0')
if day not in days[year_month]:
days[year_month].append(day)
sources_list = list()
for year_month in days:
request = {"param": self.parameter,
"date": days[year_month],
# Parameters passed to the filename mangling
"url": self._BASEURL,
"leveltype": self.leveltype,
"isodate": "-".join([year_month[:4], year_month[4:]])
}
if self.level is not None:
request.update({'levelist': self.level})
source = cml.load_source("indexed-urls", PerUrlIndex(self._ANALYSIS_PATTERN), request)
sources_list.append(source)
self.obs_source = cml.load_source("multi", *sources_list)
obs = self.obs_source.to_xarray(**obs_kwargs)
obs_valid_time = obs.valid_time.to_pandas()
obs_time_list = list(map(convert_to_datetime, obs_valid_time.iloc[:, 0]))
idx = list()
for i, t in enumerate(obs_time_list):
if t in fcs_time_list:
idx.append(i)
obs_fcs = obs.isel(time=idx)
# reshape obs to fit fcs TODO: still messy, should be reworked
obs_dict = obs_fcs.to_dict()
_, obs_fcs = xr.align(fcs, obs_fcs, join='left', exclude=['number'])
new_obs_dict = obs_fcs.to_dict()
new_obs_dict['coords']['valid_time']['data'] = [fcs_time_list]
for var in new_obs_dict['data_vars']:
new_obs_dict['data_vars'][var]['data'] = list(np.array(obs_dict['data_vars'][var]["data"]).swapaxes(1, 2))
obs_fcs = obs_fcs.from_dict(new_obs_dict)
return obs_fcs
class TrainingDataForecastEfi(TrainingDataForecast):
name = None # TODO
home_page = "-" # TODO
licence = "-" # TODO
documentation = "-" # TODO
citation = "-" # TODO
dataset = None
_PATTERN = (
"{url}data/fcs/efi/"
"EU_forecast_efi_params_{year}-{month}_0.grb"
)
_efi_parameters = ["capesi", "10fgi", "capei", "sfi", "10wsi", "2ti", "mx2ti", "mn2ti", "tpi", "all"]
@normalize("parameter", _efi_parameters)
@normalize("date", "date(%Y%m%d)")
def __init__(self, date, parameter):
TrainingDataForecast.__init__(self)
if isinstance(date, (list, tuple)):
warnings.warn('Please note that you can only download one forecast date per `climmetlab.load_dataset` call.\n' +
'Providing a list of dates might lead to a failure.')
if parameter == "all":
self.parameter = self._efi_parameters
else:
self.parameter = parameter
self.date = date
self.year = date[:4]
self.month = date[4:6]
self.step = ["0-24", "24-48", "48-72", "72-96", "96-120", "120-144", "144-168"] # TODO : deal with the 240-360 steprange for 10wsi, 2ti, tpi
request = {"param": self.parameter,
"date": self.date,
"step": self.step,
# Parameters passed to the filename mangling
"url": self._BASEURL,
"month": self.month,
"year": self.year}
self.source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
def get_observations_as_xarray(self, fcs_kwargs=None, **obs_kwargs):
warnings.warn("Observations are not available for the EFI forecasts.")
return None
class TrainingDataForecastSurface(TrainingDataForecast):
name = None # TODO
home_page = "-" # TODO
licence = "-" # TODO
documentation = "-" # TODO
citation = "-" # TODO
dataset = None
_PATTERN = (
"{url}data/fcs/{leveltype}/"
"EU_forecast_{kind}_{leveltype}_params_{isodate}_0.grb"
)
_surf_parameters = ["2t", "10u", "10v", "tcc", "100u", "100v", "cape", "stl1", "sd",
"tcw", "tcwv", "swvl1", "vis", "all"]
_surf_parameters += ["cin"]
@normalize("parameter", _surf_parameters)
@normalize("date", "date(%Y%m%d)")
def __init__(self, date, parameter, kind):
TrainingDataForecast.__init__(self)
if isinstance(date, (list, tuple)):
warnings.warn('Please note that you can only download one forecast date per `climmetlab.load_dataset` call.\n' +
'Providing a list of dates might lead to a failure.')
if parameter == "all":
self.parameter = self._surf_parameters
else:
self.parameter = parameter
self.date = date
self.leveltype = "surf"
self.kind = kind
self.obs_source = None
self.isodate = "-".join([date[:4], date[4:6], date[6:]])
if kind in self._ensemble_alias:
request = {"param": self.parameter,
# "date": self.date, ## Not needed
# Parameters passed to the filename mangling
"url": self._BASEURL,
"kind": "ens",
"leveltype": self.leveltype,
"isodate": self.isodate
}
ens_source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
request.update({"date": self.date,
# Parameters passed to the filename mangling
"kind": "ctr",
"isodate": self.isodate[:7]
})
ctr_source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
self.source = cml.load_source("multi", ens_source, ctr_source)
else: # default to highres forecasts
request = {"param": self.parameter,
"date": self.date,
# Parameters passed to the filename mangling
"url": self._BASEURL,
"kind": "hr",
"leveltype": self.leveltype,
"isodate": self.isodate[:7]
}
self.source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
class TrainingDataForecastPressure(TrainingDataForecast):
name = None # TODO
home_page = "-" # TODO
licence = "-" # TODO
documentation = "-" # TODO
citation = "-" # TODO
dataset = None
_PATTERN = (
"{url}data/fcs/{leveltype}/"
"EU_forecast_{kind}_{leveltype}_params_{isodate}_0.grb"
)
_pressure_parameters = ['z', 'u', 'v', 'q', 't', 'r', 'all']
_pressure_parameters_by_level = {500: ['z'],
700: ['u', 'v', 'q'],
850: ['t', 'r']}
@normalize("parameter", _pressure_parameters)
@normalize("date", "date(%Y%m%d)")
def __init__(self, date, parameter, level, kind):
TrainingDataForecast.__init__(self)
if isinstance(date, (list, tuple)):
warnings.warn('Please note that you can only download one forecast date per `climmetlab.load_dataset` call.\n' +
'Providing a list of dates might lead to a failure.')
if parameter == "all":
self.parameter = self._pressure_parameters_by_level[int(level)]
else:
self.parameter = parameter
self.date = date
self.leveltype = "pressure"
self.kind = kind
self.level = str(level)
self.obs_source = None
self.isodate = "-".join([date[:4], date[4:6], date[6:]])
if kind in self._ensemble_alias:
request = {"param": self.parameter,
# "date": self.date, ## Not needed
"levelist": self.level,
# Parameters passed to the filename mangling
"url": self._BASEURL,
"kind": "ens",
"leveltype": self.leveltype,
"isodate": self.isodate
}
ens_source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
request.update({"date": self.date,
# Parameters passed to the filename mangling
"kind": "ctr",
"isodate": self.isodate[:7]
})
ctr_source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
self.source = cml.load_source("multi", ens_source, ctr_source)
else: # default to highres forecasts
request = {"param": self.parameter,
"date": self.date,
"levelist": self.level,
# Parameters passed to the filename mangling
"url": self._BASEURL,
"kind": "hr",
"leveltype": self.leveltype,
"isodate": self.isodate[:7]
}
self.source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
class TrainingDataForecastSurfacePostProcessed(TrainingDataForecast):
name = None # TODO
home_page = "-" # TODO
licence = "-" # TODO
documentation = "-" # TODO
citation = "-" # TODO
dataset = None
_PATTERN = (
"{url}data/fcs/{leveltype}/"
"EU_forecast_{kind}_{leveltype}_params_{isodate}_0.grb"
)
_surf_pp_parameters = ["tp", "sshf", "slhf", "ssr", "str", "cp", "ssrd", "strd", "10fg6", "mn2t6", "mx2t6"]
# _surf_pp_parameters += ["cin"]
_not_6 = []
for par in _surf_pp_parameters:
if '6' not in par:
_not_6.append(par)
_to_diff = ["tp", "sshf", "slhf", "ssr", "str", "cp", "ssrd", "strd"] # accumulated parameters to differentiate
_parameters_ufunc = {"10fg6": "max",
"mn2t6": "min",
"mx2t6": "max"}
for par in _not_6:
_parameters_ufunc[par] = 'sum'
_parameters_base = {"10fg6": 0,
"mn2t6": 0,
"mx2t6": 0}
for par in _not_6:
_parameters_base[par] = 0
_parameters_loffset = {"10fg6": 0,
"mn2t6": 0,
"mx2t6": 0}
for par in _not_6:
_parameters_loffset[par] = '5H'
@normalize("parameter", _surf_pp_parameters)
@normalize("date", "date(%Y%m%d)")
def __init__(self, date, parameter, kind):
TrainingDataForecast.__init__(self)
if isinstance(date, (list, tuple)):
warnings.warn('Please note that you can only download one forecast date per `climmetlab.load_dataset` call.\n' +
'Providing a list of dates might lead to a failure.')
if isinstance(parameter, (tuple, list)):
if 'tp' in parameter and len(parameter) > 1:
warnings.warn("For technical reason, the parameter 'tp' can only be downloaded alone. \nRemoving this parameter from the list.\n"
"Please make a new separate request with only 'tp' as parameter.")
npar = parameter
parameter = list()
for par in npar:
if par != 'tp':
parameter.append(par)
else:
parameter = [parameter]
self.parameter = parameter
self.date = date
self.leveltype = "surf"
self.kind = kind
self.obs_source = None
self.isodate = "-".join([date[:4], date[4:6], date[6:]])
if kind in self._ensemble_alias:
request = {"param": self.parameter,
# "date": self.date, ## Not needed
# Parameters passed to the filename mangling
"url": self._BASEURL,
"kind": "ens",
"leveltype": self.leveltype,
"isodate": self.isodate
}
ens_source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
request.update({"date": self.date,
# Parameters passed to the filename mangling
"kind": "ctr",
"isodate": self.isodate[:7]
})
ctr_source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
self.source = cml.load_source("multi", ens_source, ctr_source)
else: # default to highres forecasts
request = {"param": self.parameter,
"date": self.date,
# Parameters passed to the filename mangling
"url": self._BASEURL,
"kind": "hr",
"leveltype": self.leveltype,
"isodate": self.isodate[:7]
}
self.source = cml.load_source("indexed-urls", PerUrlIndex(self._PATTERN), request)
def to_xarray(self, **kwargs):
fcs = self.source.to_xarray(**kwargs)
variables = list(fcs.keys())
ds_list = list()
for var in variables:
if var in self._to_diff: # need to do a finite difference to get the accumulation per step
da = fcs[var].diff('step')
# elif var == 'cin':
# da = fcs[var].isel(step=slice(1, None))
else:
da = fcs[var]
if var == 'p10fg6': # remove first step for wind gusts
var = var[1:]
if self._parameters_ufunc[var] == "sum":
ds_resampled = da.resample({'step': '6H'}, label='right', closed='right',
base=self._parameters_base[var],
loffset=self._parameters_loffset[var]).sum()
elif self._parameters_ufunc[var] == "min":
ds_resampled = da.resample({'step': '6H'}, label='right', closed='right',
base=self._parameters_base[var],
loffset=self._parameters_loffset[var]).min()
elif self._parameters_ufunc[var] == "max":
ds_resampled = da.resample({'step': '6H'}, label='right', closed='right',
base=self._parameters_base[var],
loffset=self._parameters_loffset[var]).max()
else: # for debug, do nothing
ds_resampled = da
ds_list.append(ds_resampled.to_dataset())
ds = xr.merge(ds_list).assign_attrs(fcs.attrs)
for var in variables:
if var in self._not_6:
try:
new_ds = ds.rename_vars({var: var + '6'})
ds = new_ds
except:
pass
return ds.assign_coords(valid_time=ds.time + ds.step)
def get_observations_as_xarray(self, fcs_kwargs=None, **obs_kwargs):
if fcs_kwargs is None:
fcs_kwargs = dict()
fcs = self.to_xarray(**fcs_kwargs)
fcs_valid_time = fcs.valid_time.to_pandas()
fcs_time_list = list(map(convert_to_datetime, fcs_valid_time.iloc[0, :]))
initial_time = fcs_time_list[0]
final_time = fcs_time_list[-1]
previous_time = initial_time - datetime.timedelta(days=1)
tlist = [previous_time] + fcs_time_list
days = dict()
extra_date = None
for t in tlist:
year_month = str(t.year).rjust(4, '0') + str(t.month).rjust(2, '0')
if year_month not in days:
days[year_month] = list()
day = year_month + str(t.day).rjust(2, '0')
if day not in days[year_month]:
days[year_month].append(day)
if t.day == 1: # Edge case when crossing months
pt = t - datetime.timedelta(days=1)
iso_pt = str(pt.year).rjust(4, '0') + str(pt.month).rjust(2, '0') + str(pt.day).rjust(2, '0')
extra_date = {year_month: iso_pt}
parameters = list()
for param in self.parameter:
if param not in self._not_6:
parameters.append(param[:-1])
else:
parameters.append(param)
sources_list = list()
for year_month in days:
request = {"param": parameters,
"date": days[year_month],
# Parameters passed to the filename mangling
"url": self._BASEURL,
"leveltype": self.leveltype,
"isodate": "-".join([year_month[:4], year_month[4:]])
}
if self.level is not None:
request.update({'levelist': self.level})
source = cml.load_source("indexed-urls", PerUrlIndex(self._ANALYSIS_PATTERN), request)
sources_list.append(source)
extra_sources_list = list()
if extra_date is not None: # Edge case when crossing months
for year_month in extra_date:
request = {"param": parameters,
"date": extra_date[year_month],
# Parameters passed to the filename mangling
"url": self._BASEURL,
"leveltype": self.leveltype,
"isodate": "-".join([year_month[:4], year_month[4:]])
}
if self.level is not None:
request.update({'levelist': self.level})
extra_source = cml.load_source("indexed-urls", PerUrlIndex(self._ANALYSIS_PATTERN), request)
extra_sources_list.append(extra_source)
extra_source = cml.load_source("multi", *extra_sources_list)
else:
extra_source = None
self.obs_source = cml.load_source("multi", *sources_list)
obs = self.obs_source.to_xarray(**obs_kwargs)
if extra_source is not None:
extra_obs = extra_source.to_xarray(**obs_kwargs)
new_obs = obs.merge(extra_obs)
else:
new_obs = obs
new_obs = new_obs.stack(datetime=("time", "step")).drop_vars("datetime").swap_dims({"datetime": "time"}).rename({"valid_time": "time"})
obs_time = new_obs.time.to_pandas()
obs_valid_time = (final_time >= obs_time) & (obs_time >= initial_time - datetime.timedelta(hours=5))
obs_fcs = new_obs.isel(time=obs_valid_time)
# filter obs to fit fcs
variables = list(obs_fcs.keys())
ds_list = list()
for var in variables:
da = obs_fcs[var].set_index(time="time")
if var == 'fg10':
var = '10fg6'
elif var in ['mn2t', 'mx2t']:
var += '6'
if self._parameters_ufunc[var] == "sum":
ds_resampled = da.resample({'time': '6H'}, label='right', closed='right').sum()
elif self._parameters_ufunc[var] == "min":
ds_resampled = da.resample({'time': '6H'}, label='right', closed='right').min()
elif self._parameters_ufunc[var] == "max":
ds_resampled = da.resample({'time': '6H'}, label='right', closed='right').max()
else: # for debug, do nothing
ds_resampled = da
ds_list.append(ds_resampled.to_dataset())
obs_fcs = xr.merge(ds_list).assign_attrs(obs.attrs)
# reshape obs to fit fcs TODO: still messy, should be reworked
shape = list(fcs[list(fcs.keys())[0]].shape)
shape[0] = 1
obs_dict = obs_fcs.to_dict()
_, obs_fcs = xr.align(fcs, obs, join='left', exclude=['number'])
new_obs_dict = obs_fcs.to_dict()
new_obs_dict['coords']['valid_time']['data'] = [fcs_time_list]
for var in new_obs_dict['data_vars']:
new_obs_dict['data_vars'][var]['data'] = list(np.array(obs_dict['data_vars'][var]["data"]).swapaxes(-1, -2).swapaxes(-2, -3) .reshape(shape))
obs_fcs = obs_fcs.from_dict(new_obs_dict)
var_name = dict()
for var in obs_fcs.keys():
if var == 'fg10':
var_name[var] = "p10fg6"
elif var in ['mn2t', 'mx2t'] or var in self._not_6:
var_name[var] = var + '6'
obs_fcs = obs_fcs.rename_vars(var_name)
return obs_fcs