-
-
Notifications
You must be signed in to change notification settings - Fork 1k
/
times.py
762 lines (635 loc) · 27.2 KB
/
times.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
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
from __future__ import annotations
import re
import warnings
from collections.abc import Hashable
from datetime import datetime, timedelta
from functools import partial
from typing import TYPE_CHECKING, Callable, Union
import numpy as np
import pandas as pd
from pandas.errors import OutOfBoundsDatetime, OutOfBoundsTimedelta
from xarray.coding.variables import (
SerializationWarning,
VariableCoder,
lazy_elemwise_func,
pop_to,
safe_setitem,
unpack_for_decoding,
unpack_for_encoding,
)
from xarray.core import indexing
from xarray.core.common import contains_cftime_datetimes, is_np_datetime_like
from xarray.core.formatting import first_n_items, format_timestamp, last_item
from xarray.core.pdcompat import nanosecond_precision_timestamp
from xarray.core.pycompat import is_duck_dask_array
from xarray.core.variable import Variable
try:
import cftime
except ImportError:
cftime = None
if TYPE_CHECKING:
from xarray.core.types import CFCalendar
T_Name = Union[Hashable, None]
# standard calendars recognized by cftime
_STANDARD_CALENDARS = {"standard", "gregorian", "proleptic_gregorian"}
_NS_PER_TIME_DELTA = {
"ns": 1,
"us": int(1e3),
"ms": int(1e6),
"s": int(1e9),
"m": int(1e9) * 60,
"h": int(1e9) * 60 * 60,
"D": int(1e9) * 60 * 60 * 24,
}
_US_PER_TIME_DELTA = {
"microseconds": 1,
"milliseconds": 1_000,
"seconds": 1_000_000,
"minutes": 60 * 1_000_000,
"hours": 60 * 60 * 1_000_000,
"days": 24 * 60 * 60 * 1_000_000,
}
_NETCDF_TIME_UNITS_CFTIME = [
"days",
"hours",
"minutes",
"seconds",
"milliseconds",
"microseconds",
]
_NETCDF_TIME_UNITS_NUMPY = _NETCDF_TIME_UNITS_CFTIME + ["nanoseconds"]
TIME_UNITS = frozenset(
[
"days",
"hours",
"minutes",
"seconds",
"milliseconds",
"microseconds",
"nanoseconds",
]
)
def _is_standard_calendar(calendar: str) -> bool:
return calendar.lower() in _STANDARD_CALENDARS
def _is_numpy_compatible_time_range(times):
if is_np_datetime_like(times.dtype):
return True
# times array contains cftime objects
times = np.asarray(times)
tmin = times.min()
tmax = times.max()
try:
convert_time_or_go_back(tmin, pd.Timestamp)
convert_time_or_go_back(tmax, pd.Timestamp)
except pd.errors.OutOfBoundsDatetime:
return False
except ValueError as err:
if err.args[0] == "year 0 is out of range":
return False
raise
else:
return True
def _netcdf_to_numpy_timeunit(units: str) -> str:
units = units.lower()
if not units.endswith("s"):
units = f"{units}s"
return {
"nanoseconds": "ns",
"microseconds": "us",
"milliseconds": "ms",
"seconds": "s",
"minutes": "m",
"hours": "h",
"days": "D",
}[units]
def _ensure_padded_year(ref_date: str) -> str:
# Reference dates without a padded year (e.g. since 1-1-1 or since 2-3-4)
# are ambiguous (is it YMD or DMY?). This can lead to some very odd
# behaviour e.g. pandas (via dateutil) passes '1-1-1 00:00:0.0' as
# '2001-01-01 00:00:00' (because it assumes a) DMY and b) that year 1 is
# shorthand for 2001 (like 02 would be shorthand for year 2002)).
# Here we ensure that there is always a four-digit year, with the
# assumption being that year comes first if we get something ambiguous.
matches_year = re.match(r".*\d{4}.*", ref_date)
if matches_year:
# all good, return
return ref_date
# No four-digit strings, assume the first digits are the year and pad
# appropriately
matches_start_digits = re.match(r"(\d+)(.*)", ref_date)
if not matches_start_digits:
raise ValueError(f"invalid reference date for time units: {ref_date}")
ref_year, everything_else = (s for s in matches_start_digits.groups())
ref_date_padded = f"{int(ref_year):04d}{everything_else}"
warning_msg = (
f"Ambiguous reference date string: {ref_date}. The first value is "
"assumed to be the year hence will be padded with zeros to remove "
f"the ambiguity (the padded reference date string is: {ref_date_padded}). "
"To remove this message, remove the ambiguity by padding your reference "
"date strings with zeros."
)
warnings.warn(warning_msg, SerializationWarning)
return ref_date_padded
def _unpack_netcdf_time_units(units: str) -> tuple[str, str]:
# CF datetime units follow the format: "UNIT since DATE"
# this parses out the unit and date allowing for extraneous
# whitespace. It also ensures that the year is padded with zeros
# so it will be correctly understood by pandas (via dateutil).
matches = re.match(r"(.+) since (.+)", units)
if not matches:
raise ValueError(f"invalid time units: {units}")
delta_units, ref_date = (s.strip() for s in matches.groups())
ref_date = _ensure_padded_year(ref_date)
return delta_units, ref_date
def _decode_cf_datetime_dtype(
data, units: str, calendar: str, use_cftime: bool | None
) -> np.dtype:
# Verify that at least the first and last date can be decoded
# successfully. Otherwise, tracebacks end up swallowed by
# Dataset.__repr__ when users try to view their lazily decoded array.
values = indexing.ImplicitToExplicitIndexingAdapter(indexing.as_indexable(data))
example_value = np.concatenate(
[first_n_items(values, 1) or [0], last_item(values) or [0]]
)
try:
result = decode_cf_datetime(example_value, units, calendar, use_cftime)
except Exception:
calendar_msg = (
"the default calendar" if calendar is None else f"calendar {calendar!r}"
)
msg = (
f"unable to decode time units {units!r} with {calendar_msg!r}. Try "
"opening your dataset with decode_times=False or installing cftime "
"if it is not installed."
)
raise ValueError(msg)
else:
dtype = getattr(result, "dtype", np.dtype("object"))
return dtype
def _decode_datetime_with_cftime(
num_dates: np.ndarray, units: str, calendar: str
) -> np.ndarray:
if cftime is None:
raise ModuleNotFoundError("No module named 'cftime'")
if num_dates.size > 0:
return np.asarray(
cftime.num2date(num_dates, units, calendar, only_use_cftime_datetimes=True)
)
else:
return np.array([], dtype=object)
def _decode_datetime_with_pandas(
flat_num_dates: np.ndarray, units: str, calendar: str
) -> np.ndarray:
if not _is_standard_calendar(calendar):
raise OutOfBoundsDatetime(
"Cannot decode times from a non-standard calendar, {!r}, using "
"pandas.".format(calendar)
)
delta, ref_date = _unpack_netcdf_time_units(units)
delta = _netcdf_to_numpy_timeunit(delta)
try:
# TODO: the strict enforcement of nanosecond precision Timestamps can be
# relaxed when addressing GitHub issue #7493.
ref_date = nanosecond_precision_timestamp(ref_date)
except ValueError:
# ValueError is raised by pd.Timestamp for non-ISO timestamp
# strings, in which case we fall back to using cftime
raise OutOfBoundsDatetime
with warnings.catch_warnings():
warnings.filterwarnings("ignore", "invalid value encountered", RuntimeWarning)
if flat_num_dates.size > 0:
# avoid size 0 datetimes GH1329
pd.to_timedelta(flat_num_dates.min(), delta) + ref_date
pd.to_timedelta(flat_num_dates.max(), delta) + ref_date
# To avoid integer overflow when converting to nanosecond units for integer
# dtypes smaller than np.int64 cast all integer and unsigned integer dtype
# arrays to np.int64 (GH 2002, GH 6589). Note this is safe even in the case
# of np.uint64 values, because any np.uint64 value that would lead to
# overflow when converting to np.int64 would not be representable with a
# timedelta64 value, and therefore would raise an error in the lines above.
if flat_num_dates.dtype.kind in "iu":
flat_num_dates = flat_num_dates.astype(np.int64)
# Cast input ordinals to integers of nanoseconds because pd.to_timedelta
# works much faster when dealing with integers (GH 1399).
flat_num_dates_ns_int = (flat_num_dates * _NS_PER_TIME_DELTA[delta]).astype(
np.int64
)
# Use pd.to_timedelta to safely cast integer values to timedeltas,
# and add those to a Timestamp to safely produce a DatetimeIndex. This
# ensures that we do not encounter integer overflow at any point in the
# process without raising OutOfBoundsDatetime.
return (pd.to_timedelta(flat_num_dates_ns_int, "ns") + ref_date).values
def decode_cf_datetime(
num_dates, units: str, calendar: str | None = None, use_cftime: bool | None = None
) -> np.ndarray:
"""Given an array of numeric dates in netCDF format, convert it into a
numpy array of date time objects.
For standard (Gregorian) calendars, this function uses vectorized
operations, which makes it much faster than cftime.num2date. In such a
case, the returned array will be of type np.datetime64.
Note that time unit in `units` must not be smaller than microseconds and
not larger than days.
See Also
--------
cftime.num2date
"""
num_dates = np.asarray(num_dates)
flat_num_dates = num_dates.ravel()
if calendar is None:
calendar = "standard"
if use_cftime is None:
try:
dates = _decode_datetime_with_pandas(flat_num_dates, units, calendar)
except (KeyError, OutOfBoundsDatetime, OutOfBoundsTimedelta, OverflowError):
dates = _decode_datetime_with_cftime(
flat_num_dates.astype(float), units, calendar
)
if (
dates[np.nanargmin(num_dates)].year < 1678
or dates[np.nanargmax(num_dates)].year >= 2262
):
if _is_standard_calendar(calendar):
warnings.warn(
"Unable to decode time axis into full "
"numpy.datetime64 objects, continuing using "
"cftime.datetime objects instead, reason: dates out "
"of range",
SerializationWarning,
stacklevel=3,
)
else:
if _is_standard_calendar(calendar):
dates = cftime_to_nptime(dates)
elif use_cftime:
dates = _decode_datetime_with_cftime(flat_num_dates, units, calendar)
else:
dates = _decode_datetime_with_pandas(flat_num_dates, units, calendar)
return dates.reshape(num_dates.shape)
def to_timedelta_unboxed(value, **kwargs):
result = pd.to_timedelta(value, **kwargs).to_numpy()
assert result.dtype == "timedelta64[ns]"
return result
def to_datetime_unboxed(value, **kwargs):
result = pd.to_datetime(value, **kwargs).to_numpy()
assert result.dtype == "datetime64[ns]"
return result
def decode_cf_timedelta(num_timedeltas, units: str) -> np.ndarray:
"""Given an array of numeric timedeltas in netCDF format, convert it into a
numpy timedelta64[ns] array.
"""
num_timedeltas = np.asarray(num_timedeltas)
units = _netcdf_to_numpy_timeunit(units)
result = to_timedelta_unboxed(num_timedeltas.ravel(), unit=units)
return result.reshape(num_timedeltas.shape)
def _unit_timedelta_cftime(units: str) -> timedelta:
return timedelta(microseconds=_US_PER_TIME_DELTA[units])
def _unit_timedelta_numpy(units: str) -> np.timedelta64:
numpy_units = _netcdf_to_numpy_timeunit(units)
return np.timedelta64(_NS_PER_TIME_DELTA[numpy_units], "ns")
def _infer_time_units_from_diff(unique_timedeltas) -> str:
unit_timedelta: Callable[[str], timedelta] | Callable[[str], np.timedelta64]
zero_timedelta: timedelta | np.timedelta64
if unique_timedeltas.dtype == np.dtype("O"):
time_units = _NETCDF_TIME_UNITS_CFTIME
unit_timedelta = _unit_timedelta_cftime
zero_timedelta = timedelta(microseconds=0)
else:
time_units = _NETCDF_TIME_UNITS_NUMPY
unit_timedelta = _unit_timedelta_numpy
zero_timedelta = np.timedelta64(0, "ns")
for time_unit in time_units:
if np.all(unique_timedeltas % unit_timedelta(time_unit) == zero_timedelta):
return time_unit
return "seconds"
def infer_calendar_name(dates) -> CFCalendar:
"""Given an array of datetimes, infer the CF calendar name"""
if is_np_datetime_like(dates.dtype):
return "proleptic_gregorian"
elif dates.dtype == np.dtype("O") and dates.size > 0:
# Logic copied from core.common.contains_cftime_datetimes.
if cftime is not None:
sample = np.asarray(dates).flat[0]
if is_duck_dask_array(sample):
sample = sample.compute()
if isinstance(sample, np.ndarray):
sample = sample.item()
if isinstance(sample, cftime.datetime):
return sample.calendar
# Error raise if dtype is neither datetime or "O", if cftime is not importable, and if element of 'O' dtype is not cftime.
raise ValueError("Array does not contain datetime objects.")
def infer_datetime_units(dates) -> str:
"""Given an array of datetimes, returns a CF compatible time-unit string of
the form "{time_unit} since {date[0]}", where `time_unit` is 'days',
'hours', 'minutes' or 'seconds' (the first one that can evenly divide all
unique time deltas in `dates`)
"""
dates = np.asarray(dates).ravel()
if np.asarray(dates).dtype == "datetime64[ns]":
dates = to_datetime_unboxed(dates)
dates = dates[pd.notnull(dates)]
reference_date = dates[0] if len(dates) > 0 else "1970-01-01"
# TODO: the strict enforcement of nanosecond precision Timestamps can be
# relaxed when addressing GitHub issue #7493.
reference_date = nanosecond_precision_timestamp(reference_date)
else:
reference_date = dates[0] if len(dates) > 0 else "1970-01-01"
reference_date = format_cftime_datetime(reference_date)
unique_timedeltas = np.unique(np.diff(dates))
units = _infer_time_units_from_diff(unique_timedeltas)
return f"{units} since {reference_date}"
def format_cftime_datetime(date) -> str:
"""Converts a cftime.datetime object to a string with the format:
YYYY-MM-DD HH:MM:SS.UUUUUU
"""
return "{:04d}-{:02d}-{:02d} {:02d}:{:02d}:{:02d}.{:06d}".format(
date.year,
date.month,
date.day,
date.hour,
date.minute,
date.second,
date.microsecond,
)
def infer_timedelta_units(deltas) -> str:
"""Given an array of timedeltas, returns a CF compatible time-unit from
{'days', 'hours', 'minutes' 'seconds'} (the first one that can evenly
divide all unique time deltas in `deltas`)
"""
deltas = to_timedelta_unboxed(np.asarray(deltas).ravel())
unique_timedeltas = np.unique(deltas[pd.notnull(deltas)])
return _infer_time_units_from_diff(unique_timedeltas)
def cftime_to_nptime(times, raise_on_invalid: bool = True) -> np.ndarray:
"""Given an array of cftime.datetime objects, return an array of
numpy.datetime64 objects of the same size
If raise_on_invalid is True (default), invalid dates trigger a ValueError.
Otherwise, the invalid element is replaced by np.NaT."""
times = np.asarray(times)
# TODO: the strict enforcement of nanosecond precision datetime values can
# be relaxed when addressing GitHub issue #7493.
new = np.empty(times.shape, dtype="M8[ns]")
for i, t in np.ndenumerate(times):
try:
# Use pandas.Timestamp in place of datetime.datetime, because
# NumPy casts it safely it np.datetime64[ns] for dates outside
# 1678 to 2262 (this is not currently the case for
# datetime.datetime).
dt = nanosecond_precision_timestamp(
t.year, t.month, t.day, t.hour, t.minute, t.second, t.microsecond
)
except ValueError as e:
if raise_on_invalid:
raise ValueError(
"Cannot convert date {} to a date in the "
"standard calendar. Reason: {}.".format(t, e)
)
else:
dt = "NaT"
new[i] = np.datetime64(dt)
return new
def convert_times(times, date_type, raise_on_invalid: bool = True) -> np.ndarray:
"""Given an array of datetimes, return the same dates in another cftime or numpy date type.
Useful to convert between calendars in numpy and cftime or between cftime calendars.
If raise_on_valid is True (default), invalid dates trigger a ValueError.
Otherwise, the invalid element is replaced by np.NaN for cftime types and np.NaT for np.datetime64.
"""
if date_type in (pd.Timestamp, np.datetime64) and not is_np_datetime_like(
times.dtype
):
return cftime_to_nptime(times, raise_on_invalid=raise_on_invalid)
if is_np_datetime_like(times.dtype):
# Convert datetime64 objects to Timestamps since those have year, month, day, etc. attributes
times = pd.DatetimeIndex(times)
new = np.empty(times.shape, dtype="O")
for i, t in enumerate(times):
try:
dt = date_type(
t.year, t.month, t.day, t.hour, t.minute, t.second, t.microsecond
)
except ValueError as e:
if raise_on_invalid:
raise ValueError(
"Cannot convert date {} to a date in the "
"{} calendar. Reason: {}.".format(
t, date_type(2000, 1, 1).calendar, e
)
)
else:
dt = np.NaN
new[i] = dt
return new
def convert_time_or_go_back(date, date_type):
"""Convert a single date to a new date_type (cftime.datetime or pd.Timestamp).
If the new date is invalid, it goes back a day and tries again. If it is still
invalid, goes back a second day.
This is meant to convert end-of-month dates into a new calendar.
"""
# TODO: the strict enforcement of nanosecond precision Timestamps can be
# relaxed when addressing GitHub issue #7493.
if date_type == pd.Timestamp:
date_type = nanosecond_precision_timestamp
try:
return date_type(
date.year,
date.month,
date.day,
date.hour,
date.minute,
date.second,
date.microsecond,
)
except OutOfBoundsDatetime:
raise
except ValueError:
# Day is invalid, happens at the end of months, try again the day before
try:
return date_type(
date.year,
date.month,
date.day - 1,
date.hour,
date.minute,
date.second,
date.microsecond,
)
except ValueError:
# Still invalid, happens for 360_day to non-leap february. Try again 2 days before date.
return date_type(
date.year,
date.month,
date.day - 2,
date.hour,
date.minute,
date.second,
date.microsecond,
)
def _should_cftime_be_used(
source, target_calendar: str, use_cftime: bool | None
) -> bool:
"""Return whether conversion of the source to the target calendar should
result in a cftime-backed array.
Source is a 1D datetime array, target_cal a string (calendar name) and
use_cftime is a boolean or None. If use_cftime is None, this returns True
if the source's range and target calendar are convertible to np.datetime64 objects.
"""
# Arguments Checks for target
if use_cftime is not True:
if _is_standard_calendar(target_calendar):
if _is_numpy_compatible_time_range(source):
# Conversion is possible with pandas, force False if it was None
return False
elif use_cftime is False:
raise ValueError(
"Source time range is not valid for numpy datetimes. Try using `use_cftime=True`."
)
elif use_cftime is False:
raise ValueError(
f"Calendar '{target_calendar}' is only valid with cftime. Try using `use_cftime=True`."
)
return True
def _cleanup_netcdf_time_units(units: str) -> str:
delta, ref_date = _unpack_netcdf_time_units(units)
try:
units = f"{delta} since {format_timestamp(ref_date)}"
except (OutOfBoundsDatetime, ValueError):
# don't worry about reifying the units if they're out of bounds or
# formatted badly
pass
return units
def _encode_datetime_with_cftime(dates, units: str, calendar: str) -> np.ndarray:
"""Fallback method for encoding dates using cftime.
This method is more flexible than xarray's parsing using datetime64[ns]
arrays but also slower because it loops over each element.
"""
if cftime is None:
raise ModuleNotFoundError("No module named 'cftime'")
if np.issubdtype(dates.dtype, np.datetime64):
# numpy's broken datetime conversion only works for us precision
dates = dates.astype("M8[us]").astype(datetime)
def encode_datetime(d):
# Since netCDF files do not support storing float128 values, we ensure
# that float64 values are used by setting longdouble=False in num2date.
# This try except logic can be removed when xarray's minimum version of
# cftime is at least 1.6.2.
try:
return (
np.nan
if d is None
else cftime.date2num(d, units, calendar, longdouble=False)
)
except TypeError:
return np.nan if d is None else cftime.date2num(d, units, calendar)
return np.array([encode_datetime(d) for d in dates.ravel()]).reshape(dates.shape)
def cast_to_int_if_safe(num) -> np.ndarray:
int_num = np.asarray(num, dtype=np.int64)
if (num == int_num).all():
num = int_num
return num
def encode_cf_datetime(
dates, units: str | None = None, calendar: str | None = None
) -> tuple[np.ndarray, str, str]:
"""Given an array of datetime objects, returns the tuple `(num, units,
calendar)` suitable for a CF compliant time variable.
Unlike `date2num`, this function can handle datetime64 arrays.
See Also
--------
cftime.date2num
"""
dates = np.asarray(dates)
if units is None:
units = infer_datetime_units(dates)
else:
units = _cleanup_netcdf_time_units(units)
if calendar is None:
calendar = infer_calendar_name(dates)
delta, _ref_date = _unpack_netcdf_time_units(units)
try:
if not _is_standard_calendar(calendar) or dates.dtype.kind == "O":
# parse with cftime instead
raise OutOfBoundsDatetime
assert dates.dtype == "datetime64[ns]"
delta_units = _netcdf_to_numpy_timeunit(delta)
time_delta = np.timedelta64(1, delta_units).astype("timedelta64[ns]")
# TODO: the strict enforcement of nanosecond precision Timestamps can be
# relaxed when addressing GitHub issue #7493.
ref_date = nanosecond_precision_timestamp(_ref_date)
# If the ref_date Timestamp is timezone-aware, convert to UTC and
# make it timezone-naive (GH 2649).
if ref_date.tz is not None:
ref_date = ref_date.tz_convert(None)
# Wrap the dates in a DatetimeIndex to do the subtraction to ensure
# an OverflowError is raised if the ref_date is too far away from
# dates to be encoded (GH 2272).
dates_as_index = pd.DatetimeIndex(dates.ravel())
time_deltas = dates_as_index - ref_date
# Use floor division if time_delta evenly divides all differences
# to preserve integer dtype if possible (GH 4045).
if np.all(time_deltas % time_delta == np.timedelta64(0, "ns")):
num = time_deltas // time_delta
else:
num = time_deltas / time_delta
num = num.values.reshape(dates.shape)
except (OutOfBoundsDatetime, OverflowError, ValueError):
num = _encode_datetime_with_cftime(dates, units, calendar)
num = cast_to_int_if_safe(num)
return (num, units, calendar)
def encode_cf_timedelta(timedeltas, units: str | None = None) -> tuple[np.ndarray, str]:
if units is None:
units = infer_timedelta_units(timedeltas)
np_unit = _netcdf_to_numpy_timeunit(units)
num = 1.0 * timedeltas / np.timedelta64(1, np_unit)
num = np.where(pd.isnull(timedeltas), np.nan, num)
num = cast_to_int_if_safe(num)
return (num, units)
class CFDatetimeCoder(VariableCoder):
def __init__(self, use_cftime: bool | None = None) -> None:
self.use_cftime = use_cftime
def encode(self, variable: Variable, name: T_Name = None) -> Variable:
if np.issubdtype(
variable.data.dtype, np.datetime64
) or contains_cftime_datetimes(variable):
dims, data, attrs, encoding = unpack_for_encoding(variable)
(data, units, calendar) = encode_cf_datetime(
data, encoding.pop("units", None), encoding.pop("calendar", None)
)
safe_setitem(attrs, "units", units, name=name)
safe_setitem(attrs, "calendar", calendar, name=name)
return Variable(dims, data, attrs, encoding, fastpath=True)
else:
return variable
def decode(self, variable: Variable, name: T_Name = None) -> Variable:
units = variable.attrs.get("units", None)
if isinstance(units, str) and "since" in units:
dims, data, attrs, encoding = unpack_for_decoding(variable)
units = pop_to(attrs, encoding, "units")
calendar = pop_to(attrs, encoding, "calendar")
dtype = _decode_cf_datetime_dtype(data, units, calendar, self.use_cftime)
transform = partial(
decode_cf_datetime,
units=units,
calendar=calendar,
use_cftime=self.use_cftime,
)
data = lazy_elemwise_func(data, transform, dtype)
return Variable(dims, data, attrs, encoding, fastpath=True)
else:
return variable
class CFTimedeltaCoder(VariableCoder):
def encode(self, variable: Variable, name: T_Name = None) -> Variable:
if np.issubdtype(variable.data.dtype, np.timedelta64):
dims, data, attrs, encoding = unpack_for_encoding(variable)
data, units = encode_cf_timedelta(data, encoding.pop("units", None))
safe_setitem(attrs, "units", units, name=name)
return Variable(dims, data, attrs, encoding, fastpath=True)
else:
return variable
def decode(self, variable: Variable, name: T_Name = None) -> Variable:
units = variable.attrs.get("units", None)
if isinstance(units, str) and units in TIME_UNITS:
dims, data, attrs, encoding = unpack_for_decoding(variable)
units = pop_to(attrs, encoding, "units")
transform = partial(decode_cf_timedelta, units=units)
dtype = np.dtype("timedelta64[ns]")
data = lazy_elemwise_func(data, transform, dtype=dtype)
return Variable(dims, data, attrs, encoding, fastpath=True)
else:
return variable