test_tsio.py 51.6 KB
Newer Older
1
from datetime import datetime, timedelta
2
3
from pathlib import Path
import pytz
4

5
from dateutil import parser
6
import pytest
7
8
import numpy as np
import pandas as pd
9

10
from tshistory.snapshot import Snapshot
11
12
from tshistory.util import rename_series, threadpool
from tshistory.tsio import TimeSerie
13
14
from tshistory.testutil import (
    assert_df,
15
16
    assert_hist,
    assert_hist_equals,
17
    assert_group_equals,
18
    assert_structures,
19
20
21
    genserie,
    tempattr
)
22

23
DATADIR = Path(__file__).parent / 'data'
24

Aurélien Campéas's avatar
Aurélien Campéas committed
25

26
27
28
29
def utcdt(*dt):
    return pd.Timestamp(datetime(*dt), tz='UTC')


30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
def test_in_tx(engine):
    tsh = TimeSerie()

    with pytest.raises(TypeError) as err:
        tsh.insert(engine, 0, 0, 0)
    assert err.value.args[0] == 'You must use a transaction object'

    with engine.connect() as cn:
        with pytest.raises(TypeError) as err:
            tsh.insert(cn, 0, 0, 0)
    assert err.value.args[0] == 'You must use a transaction object'

    ts = genserie(datetime(2017, 10, 28, 23),
                  'H', 4, tz='UTC')
    with engine.begin() as cn:
        tsh.insert(cn, ts, 'test_tx', 'Babar')


48
def test_tstamp_roundtrip(engine, tsh):
49
    assert_structures(engine, tsh)
50
51
    ts = genserie(datetime(2017, 10, 28, 23),
                  'H', 4, tz='UTC')
52
53
54
55
56
57
58
59
60
61
    ts.index = ts.index.tz_convert('Europe/Paris')

    assert_df("""
2017-10-29 01:00:00+02:00    0
2017-10-29 02:00:00+02:00    1
2017-10-29 02:00:00+01:00    2
2017-10-29 03:00:00+01:00    3
Freq: H
    """, ts)

62
63
    tsh.insert(engine, ts, 'tztest', 'Babar',
               _insertion_date=utcdt(2018, 1, 1))
64
    back = tsh.get(engine, 'tztest')
65
66
67

    # though un localized we understand it's been normalized to utc
    assert_df("""
68
69
70
71
2017-10-28 23:00:00+00:00    0.0
2017-10-29 00:00:00+00:00    1.0
2017-10-29 01:00:00+00:00    2.0
2017-10-29 02:00:00+00:00    3.0
72
73
74
""", back)

    assert (ts.index == back.index).all()
75
    assert str(back.index.dtype) == 'datetime64[ns, UTC]'
76

77
78
79
80
    ival = tsh.interval(engine, 'tztest')
    assert ival.left == pd.Timestamp('2017-10-28 23:00:00+0000', tz='UTC')
    assert ival.right == pd.Timestamp('2017-10-29 02:00:00+0000', tz='UTC')

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
    ts = genserie(datetime(2017, 10, 29, 1),
                  'H', 4, tz='UTC')
    ts.index = ts.index.tz_convert('Europe/Paris')
    tsh.insert(engine, ts, 'tztest', 'Celeste',
                   _insertion_date=utcdt(2018, 1, 3))

    ts = tsh.get(engine, 'tztest')
    assert_df("""
2017-10-28 23:00:00+00:00    0.0
2017-10-29 00:00:00+00:00    1.0
2017-10-29 01:00:00+00:00    0.0
2017-10-29 02:00:00+00:00    1.0
2017-10-29 03:00:00+00:00    2.0
2017-10-29 04:00:00+00:00    3.0
""", ts)

    hist = tsh.get_history(engine, 'tztest')
    assert_hist("""
insertion_date             value_date               
2018-01-01 00:00:00+00:00  2017-10-28 23:00:00+00:00    0.0
                           2017-10-29 00:00:00+00:00    1.0
                           2017-10-29 01:00:00+00:00    2.0
                           2017-10-29 02:00:00+00:00    3.0
2018-01-03 00:00:00+00:00  2017-10-28 23:00:00+00:00    0.0
                           2017-10-29 00:00:00+00:00    1.0
                           2017-10-29 01:00:00+00:00    0.0
                           2017-10-29 02:00:00+00:00    1.0
                           2017-10-29 03:00:00+00:00    2.0
                           2017-10-29 04:00:00+00:00    3.0
""", hist)

    hist = tsh.get_history(engine, 'tztest',
                           from_value_date=utcdt(2017, 10, 29, 1),
                           to_value_date=utcdt(2017, 10, 29, 3))
    assert_hist("""
insertion_date             value_date               
2018-01-01 00:00:00+00:00  2017-10-29 01:00:00+00:00    2.0
                           2017-10-29 02:00:00+00:00    3.0
2018-01-03 00:00:00+00:00  2017-10-29 01:00:00+00:00    0.0
                           2017-10-29 02:00:00+00:00    1.0
                           2017-10-29 03:00:00+00:00    2.0
""", hist)

124
125
126
    ival = tsh.interval(engine, 'tztest')
    assert ival.left == pd.Timestamp('2017-10-28 23:00:00+0000', tz='UTC')
    assert ival.right == pd.Timestamp('2017-10-29 04:00:00+0000', tz='UTC')
127
    assert_structures(engine, tsh)
128

129

130
def test_differential(engine, tsh):
131
    assert_structures(engine, tsh)
132
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 10)
133
    tsh.insert(engine, ts_begin, 'ts_test', 'test')
134

135
    id1 = tsh.last_id(engine, 'ts_test')
136
    assert tsh._previous_cset(engine, 'ts_test', id1) is None
137

138
139
    assert tsh.exists(engine, 'ts_test')
    assert not tsh.exists(engine, 'this_does_not_exist')
140

141
142
143
144
145
146
147
148
    assert tsh.interval(engine, 'ts_test') == pd.Interval(
        datetime(2010, 1, 1, 0, 0), datetime(2010, 1, 10, 0, 0),
        closed='both'
    )

    with pytest.raises(ValueError):
        assert tsh.interval(engine, 'nosuchts')

149
    fetched = tsh.get(engine, 'ts_test')
150
    assert_df("""
151
152
153
154
155
156
157
158
159
160
2010-01-01    0.0
2010-01-02    1.0
2010-01-03    2.0
2010-01-04    3.0
2010-01-05    4.0
2010-01-06    5.0
2010-01-07    6.0
2010-01-08    7.0
2010-01-09    8.0
2010-01-10    9.0
161
162
""", fetched)
    assert fetched.name == 'ts_test'
163
164

    # we should detect the emission of a message
165
    tsh.insert(engine, ts_begin, 'ts_test', 'babar')
166

167
    assert_df("""
168
169
170
171
172
173
174
175
176
177
2010-01-01    0.0
2010-01-02    1.0
2010-01-03    2.0
2010-01-04    3.0
2010-01-05    4.0
2010-01-06    5.0
2010-01-07    6.0
2010-01-08    7.0
2010-01-09    8.0
2010-01-10    9.0
178
""", tsh.get(engine, 'ts_test'))
179
180
181
182

    ts_slight_variation = ts_begin.copy()
    ts_slight_variation.iloc[3] = 0
    ts_slight_variation.iloc[6] = 0
183
    tsh.insert(engine, ts_slight_variation, 'ts_test', 'celeste')
184
    id2 = tsh.last_id(engine, 'ts_test')
185
    assert tsh._previous_cset(engine, 'ts_test', id2) == id1
186

187
    assert_df("""
188
189
190
191
192
193
194
195
196
197
2010-01-01    0.0
2010-01-02    1.0
2010-01-03    2.0
2010-01-04    0.0
2010-01-05    4.0
2010-01-06    5.0
2010-01-07    0.0
2010-01-08    7.0
2010-01-09    8.0
2010-01-10    9.0
198
""", tsh.get(engine, 'ts_test'))
199

200
    ts_longer = genserie(datetime(2010, 1, 3), 'D', 15)
201
202
203
204
    ts_longer.iloc[1] = 2.48
    ts_longer.iloc[3] = 3.14
    ts_longer.iloc[5] = ts_begin.iloc[7]

205
    with engine.begin() as cn:
206
        tsh.insert(cn, ts_longer, 'ts_test', 'test')
207
208
209
    id3 = tsh.last_id(engine, 'ts_test')

    assert id1 < id2 < id3
210

211
    assert_df("""
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
2010-01-01     0.00
2010-01-02     1.00
2010-01-03     0.00
2010-01-04     2.48
2010-01-05     2.00
2010-01-06     3.14
2010-01-07     4.00
2010-01-08     7.00
2010-01-09     6.00
2010-01-10     7.00
2010-01-11     8.00
2010-01-12     9.00
2010-01-13    10.00
2010-01-14    11.00
2010-01-15    12.00
2010-01-16    13.00
2010-01-17    14.00
229
""", tsh.get(engine, 'ts_test'))
230

231
232
233
234
235
    assert tsh.interval(engine, 'ts_test') == pd.Interval(
        datetime(2010, 1, 1, 0, 0), datetime(2010, 1, 17, 0, 0),
        closed='both'
    )

236
    # insert single data, in override of previous one
237
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 5, initval=[2])
238
    ts_begin.loc['2010-01-04'] = -1
239
    tsh.insert(engine, ts_begin, 'ts_mixte', 'test')
240

241
    assert_df("""
242
243
244
245
246
2010-01-01    2.0
2010-01-02    2.0
2010-01-03    2.0
2010-01-04   -1.0
2010-01-05    2.0
247
""", tsh.get(engine, 'ts_mixte'))
248

249
    # add new series with one additional values
250
    ts_more = genserie(datetime(2010, 1, 2), 'D', 5, [2])
251
    ts_more.loc['2010-01-04'] = -1
252
    tsh.insert(engine, ts_more, 'ts_mixte', 'test')
253

254
    assert_df("""
255
256
257
258
259
260
2010-01-01    2.0
2010-01-02    2.0
2010-01-03    2.0
2010-01-04   -1.0
2010-01-05    2.0
2010-01-06    2.0
261
""", tsh.get(engine, 'ts_mixte'))
262
263

    # just append an extra data point
264
265
    # with no intersection with the previous ts
    ts_one_more = genserie(datetime(2010, 1, 7), 'D', 1, [3])
266
    tsh.insert(engine, ts_one_more, 'ts_mixte', 'test')
267

268
    assert_df("""
269
270
271
272
273
274
275
2010-01-01    2.0
2010-01-02    2.0
2010-01-03    2.0
2010-01-04   -1.0
2010-01-05    2.0
2010-01-06    2.0
2010-01-07    3.0
276
""", tsh.get(engine, 'ts_mixte'))
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
    assert_df("""
2010-01-03    2.0
2010-01-04   -1.0
2010-01-05    2.0
2010-01-06    2.0
    """, tsh.get(engine, 'ts_mixte',
                 from_value_date=datetime(2010, 1, 3),
                 to_value_date=datetime(2010, 1, 6)
    ))

    assert_df("""
2010-01-04   -1.0
2010-01-05    2.0
2010-01-06    2.0
2010-01-07    3.0
        """, tsh.get(engine, 'ts_mixte',
                     from_value_date=datetime(2010, 1, 4)
    ))

    assert_df("""
2010-01-01    2.0
2010-01-02    2.0
2010-01-03    2.0
2010-01-04   -1.0
2010-01-05    2.0
            """, tsh.get(engine, 'ts_mixte',
                         to_value_date=datetime(2010, 1, 5)
        ))

307
    with engine.begin() as cn:
308
        cn.execute('set search_path to "{0}.timeserie", {0}, public'.format(tsh.namespace))
309
310
        allts = pd.read_sql("select seriename, table_name from registry "
                            "where seriename in ('ts_test', 'ts_mixte')",
311
                            cn)
312

313
        assert_df("""
314
315
316
seriename table_name
0   ts_test    ts_test
1  ts_mixte   ts_mixte
317
""".format(tsh.namespace), allts)
318

319
        assert_df("""
320
321
322
323
324
325
326
2010-01-01    2.0
2010-01-02    2.0
2010-01-03    2.0
2010-01-04   -1.0
2010-01-05    2.0
2010-01-06    2.0
2010-01-07    3.0
327
""", tsh.get(cn, 'ts_mixte',
328
             revision_date=datetime.now()))
329

330
331
    assert_structures(engine, tsh)

332

333
334
335
336
337
338
def test_serie_metadata(engine, tsh):
    serie = genserie(datetime(2010, 1, 1), 'D', 1, initval=[1])
    tsh.insert(engine, serie, 'ts-metadata', 'babar')

    initialmeta = tsh.metadata(engine, 'ts-metadata')
    assert initialmeta == {
339
340
        'index_dtype': '<M8[ns]',
        'index_names': [],
341
        'index_type': 'datetime64[ns]',
342
343
344
        'tzaware': False,
        'value_dtype': '<f8',
        'value_type': 'float64'
345
346
347
348
349
350
351
352
353
354
355
    }

    tsh.update_metadata(engine, 'ts-metadata',
                        {'topic': 'banana spot price'}
    )
    assert tsh.metadata(engine, 'ts-metadata')['topic'] == 'banana spot price'

    with pytest.raises(AssertionError):
        tsh.update_metadata(engine, 'ts-metadata', {'tzaware': True})

    assert tsh.metadata(engine, 'ts-metadata') == {
356
        'index_dtype': '<M8[ns]',
357
        'index_names': [],
358
359
        'index_type': 'datetime64[ns]',
        'topic': 'banana spot price',
360
        'tzaware': False,
361
362
        'value_dtype': '<f8',
        'value_type': 'float64'
363
    }
364

365

366
367
def test_changeset_metadata(engine, tsh):
    serie = genserie(datetime(2010, 1, 1), 'D', 1, initval=[1])
368
    tsh.insert(engine, serie, 'ts-cs-metadata', 'babar',
369
370
               {'foo': 'A', 'bar': 42})

371
    log = tsh.log(engine, names=['ts-cs-metadata'])
372
373
374
375
    meta = tsh.changeset_metadata(engine, log[0]['rev'])
    assert meta == {'foo': 'A', 'bar': 42}


376
def test_revision_date(engine, tsh):
377
    for i in range(1, 5):
378
        with engine.begin() as cn:
379
380
            tsh.insert(cn, genserie(datetime(2017, 1, i), 'D', 3, [i]), 'revdate',
                       'test', _insertion_date=utcdt(2016, 1, i))
381
382

    # end of prologue, now some real meat
383
    idate0 = pd.Timestamp('2015-1-1 00:00:00', tz='UTC')
384
385
386
387
    ts = genserie(datetime(2010, 1, 4), 'D', 4, [0], name='truc')
    tsh.insert(engine, ts, 'ts_through_time',
               'test', _insertion_date=idate0)
    assert idate0 == tsh.latest_insertion_date(engine, 'ts_through_time')
388

389
    idate1 = pd.Timestamp('2015-1-1 15:45:23', tz='UTC')
390
391
392
393
    ts = genserie(datetime(2010, 1, 4), 'D', 4, [1], name='truc')
    tsh.insert(engine, ts, 'ts_through_time',
               'test', _insertion_date=idate1)
    assert idate1 == tsh.latest_insertion_date(engine, 'ts_through_time')
394

395
    idate2 = pd.Timestamp('2015-1-2 15:43:23', tz='UTC')
396
397
398
399
    ts = genserie(datetime(2010, 1, 4), 'D', 4, [2], name='truc')
    tsh.insert(engine, ts, 'ts_through_time',
               'test', _insertion_date=idate2)
    assert idate2 == tsh.latest_insertion_date(engine, 'ts_through_time')
400

401
    idate3 = pd.Timestamp('2015-1-3', tz='UTC')
402
403
404
405
    ts = genserie(datetime(2010, 1, 4), 'D', 4, [3], name='truc')
    tsh.insert(engine, ts, 'ts_through_time',
               'test', _insertion_date=idate3)
    assert idate3 == tsh.latest_insertion_date(engine, 'ts_through_time')
406

407
    ts = tsh.get(engine, 'ts_through_time')
408

409
    assert_df("""
410
411
412
413
2010-01-04    3.0
2010-01-05    3.0
2010-01-06    3.0
2010-01-07    3.0
414
""", ts)
415

416
    ts = tsh.get(engine, 'ts_through_time',
Aurélien Campéas's avatar
Aurélien Campéas committed
417
                 revision_date=datetime(2015, 1, 2, 18, 43, 23))
418

419
    assert_df("""
420
421
422
423
2010-01-04    2.0
2010-01-05    2.0
2010-01-06    2.0
2010-01-07    2.0
424
""", ts)
425

426
    ts = tsh.get(engine, 'ts_through_time',
427
                 revision_date=datetime(2015, 1, 1, 18, 43, 23))
428

429
    assert_df("""
430
431
432
433
2010-01-04    1.0
2010-01-05    1.0
2010-01-06    1.0
2010-01-07    1.0
434
""", ts)
435

436
    ts = tsh.get(engine, 'ts_through_time',
437
                 revision_date=datetime(2014, 1, 1, 18, 43, 23))
438
439
440

    assert ts is None

441
442
443
444
445
446
447
448
449
450
    # epilogue: back to the revdate issue
    assert_df("""
2017-01-01    1.0
2017-01-02    2.0
2017-01-03    3.0
2017-01-04    4.0
2017-01-05    4.0
2017-01-06    4.0
""", tsh.get(engine, 'revdate'))

451
    oldstate = tsh.get(engine, 'revdate', revision_date=datetime(2016, 1, 2))
452
453
454
    assert_df("""
2017-01-01    1.0
2017-01-02    2.0
455
456
457
2017-01-03    2.0
2017-01-04    2.0
""", oldstate)
458

459

460
def test_point_deletion(engine, tsh):
461
462
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 11)
    ts_begin.iloc[-1] = np.nan
463
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
464

465
    _, ts = Snapshot(engine, tsh, 'ts_del').find()
466
    assert ts.iloc[-2] == 8.0
467

468
    ts_begin.iloc[0] = np.nan
469
    ts_begin.iloc[3] = np.nan
470

471
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
472

473
    assert_df("""
474
475
476
477
478
479
480
2010-01-02    1.0
2010-01-03    2.0
2010-01-05    4.0
2010-01-06    5.0
2010-01-07    6.0
2010-01-08    7.0
2010-01-09    8.0
481
2010-01-10    9.0
482
""", tsh.get(engine, 'ts_del'))
483

484
    ts2 = tsh.get(engine, 'ts_del',
485
486
                 # force snapshot reconstruction feature
                 revision_date=datetime(2038, 1, 1))
487
    assert (tsh.get(engine, 'ts_del') == ts2).all()
488

489
490
491
    ts_begin.iloc[0] = 42
    ts_begin.iloc[3] = 23

492
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
493

494
    assert_df("""
495
496
497
498
499
500
501
502
503
2010-01-01    42.0
2010-01-02     1.0
2010-01-03     2.0
2010-01-04    23.0
2010-01-05     4.0
2010-01-06     5.0
2010-01-07     6.0
2010-01-08     7.0
2010-01-09     8.0
504
2010-01-10     9.0
505
""", tsh.get(engine, 'ts_del'))
506
507
508

    # now with string!

509
    ts_string = genserie(datetime(2010, 1, 1), 'D', 10, ['machin'])
510
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
511
512
513
514

    ts_string[4] = None
    ts_string[5] = None

515
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
516
    assert_df("""
517
518
519
520
521
522
523
524
2010-01-01    machin
2010-01-02    machin
2010-01-03    machin
2010-01-04    machin
2010-01-07    machin
2010-01-08    machin
2010-01-09    machin
2010-01-10    machin
525
""", tsh.get(engine, 'ts_string_del'))
526
527
528
529

    ts_string[4] = 'truc'
    ts_string[6] = 'truc'

530
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
531
    assert_df("""
532
533
534
535
536
537
538
539
2010-01-01    machin
2010-01-02    machin
2010-01-03    machin
2010-01-04    machin
2010-01-05      truc
2010-01-07      truc
2010-01-08    machin
2010-01-09    machin
540
2010-01-10    machin
541
""", tsh.get(engine, 'ts_string_del'))
542

543
    ts_string[ts_string.index] = np.nan
544
545
    with pytest.raises(ValueError):
        tsh.insert(engine, ts_string, 'ts_string_del', 'test')
546

547

548
549
def test_nan_first(engine, tsh):
    # first insertion with only nan
550
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 10, [np.nan])
551
    assert tsh.insert(engine, ts_begin, 'ts_null', 'test') is None
552

553

554
def test_more_point_deletion(engine, tsh):
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
    ts_repushed = genserie(datetime(2010, 1, 1), 'D', 11)
    ts_repushed[0:3] = np.nan

    assert_df("""
2010-01-01     NaN
2010-01-02     NaN
2010-01-03     NaN
2010-01-04     3.0
2010-01-05     4.0
2010-01-06     5.0
2010-01-07     6.0
2010-01-08     7.0
2010-01-09     8.0
2010-01-10     9.0
2010-01-11    10.0
Freq: D
""", ts_repushed)

573
574
    tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
    diff = tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
575
576
    assert diff is None

577
    # there is no difference
578
    assert 0 == len(tsh.diff(ts_repushed, ts_repushed))
579
580
581
582
583

    ts_add = genserie(datetime(2010, 1, 1), 'D', 15)
    ts_add.iloc[0] = np.nan
    ts_add.iloc[13:] = np.nan
    ts_add.iloc[8] = np.nan
584
    diff = tsh.diff(ts_repushed, ts_add)
585
586
587
588
589
590

    assert_df("""
2010-01-02     1.0
2010-01-03     2.0
2010-01-09     NaN
2010-01-12    11.0
591
2010-01-13    12.0""", diff.sort_index())
592
593
594
595
    # value on nan => value
    # nan on value => nan
    # nan on nan => Nothing
    # nan on nothing=> Nothing
596

Aurélien Campéas's avatar
Aurélien Campéas committed
597
    # full erasing
598
599
    # numeric
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4)
600
    tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
601

Aurélien Campéas's avatar
Aurélien Campéas committed
602
    ts_begin.iloc[:] = np.nan
603
604
    with pytest.raises(ValueError):
        tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
605
606

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4)
607
    tsh.insert(engine, ts_end, 'ts_full_del', 'test')
608
609
610
611

    # string

    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
612
    tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
613

614
615
    ts_begin = pd.Series([np.nan] * 4, name='ts_full_del_str',
                         index=ts_begin.index)
616
617
618

    with pytest.raises(ValueError):
        tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
619
620

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
621
    tsh.insert(engine, ts_end, 'ts_full_del_str', 'test')
622

Aurélien Campéas's avatar
Aurélien Campéas committed
623

624
def test_deletion_over_horizon(engine, tsh):
625
    idate = utcdt(2018, 2, 1)
626
627
628
629
630
631
    ts = pd.Series(
        [1, 2, 3],
        index=pd.date_range(datetime(2018, 1, 1), freq='D', periods=3)
    )

    name = 'delete_over_hz'
632
633
    tsh.insert(engine, ts, name, 'Babar',
               _insertion_date=idate)
634
635
636
637
638
639

    ts = pd.Series(
        [np.nan, np.nan, np.nan],
        index=pd.date_range(datetime(2018, 1, 3), freq='D', periods=3)
    )

640
641
    tsh.insert(engine, ts, name, 'Celeste',
               _insertion_date=idate.replace(day=2))
642
643
    ival = tsh.interval(engine, name)
    assert ival.left == datetime(2018, 1, 1)
644
    assert ival.right == datetime(2018, 1, 2)
645
646
647
648
649

    ts = pd.Series(
        [np.nan, np.nan, np.nan],
        index=pd.date_range(datetime(2017, 12, 30), freq='D', periods=3)
    )
650
651
    tsh.insert(engine, ts, name, 'Arthur',
               _insertion_date=idate.replace(day=3))
652
    ival = tsh.interval(engine, name)
653
654
    assert ival.left == datetime(2018, 1, 2)
    assert ival.right == datetime(2018, 1, 2)
655
656


657
def test_get_history(engine, tsh):
658
    for numserie in (1, 2, 3):
659
        with engine.begin() as cn:
660
661
662
            tsh.insert(cn, genserie(datetime(2017, 1, 1), 'D', numserie), 'smallserie',
                       'aurelien.campeas@pythonian.fr',
                       _insertion_date=utcdt(2017, 2, numserie))
663
664
665
666
667
668
669
670
671
672
673

    ts = tsh.get(engine, 'smallserie')
    assert_df("""
2017-01-01    0.0
2017-01-02    1.0
2017-01-03    2.0
""", ts)

    logs = tsh.log(engine, names=['smallserie'])
    assert [
        {'author': 'aurelien.campeas@pythonian.fr',
674
         'meta': {},
675
         'date': pd.Timestamp('2017-02-01 00:00:00+0000', tz='UTC'),
676
677
678
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
679
         'meta': {},
680
         'date': pd.Timestamp('2017-02-02 00:00:00+0000', tz='UTC'),
681
682
683
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
684
         'meta': {},
685
         'date': pd.Timestamp('2017-02-03 00:00:00+0000', tz='UTC'),
686
687
688
689
690
691
         'names': ['smallserie']
        }
    ] == [{k: v for k, v in log.items() if k != 'rev'}
          for log in logs]
    histts = tsh.get_history(engine, 'smallserie')

692
    assert_hist("""
693
694
695
696
697
698
699
insertion_date             value_date
2017-02-01 00:00:00+00:00  2017-01-01    0.0
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
2017-02-03 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
                           2017-01-03    2.0
700
701
""", histts)

702
    diffs = tsh.get_history(engine, 'smallserie', diffmode=True)
703
    assert_hist("""
704
705
706
707
708
709
insertion_date             value_date
2017-02-01 00:00:00+00:00  2017-01-01    0.0
2017-02-02 00:00:00+00:00  2017-01-02    1.0
2017-02-03 00:00:00+00:00  2017-01-03    2.0
""", diffs)

710
    for idate in histts:
711
        with engine.begin() as cn:
712
            idate = idate.replace(tzinfo=pytz.timezone('UTC'))
713
714
            tsh.insert(cn, histts[idate], 'smallserie2',
                       'aurelien.campeas@pythonian.f', _insertion_date=idate)
715
716
717

    # this is perfectly round-tripable
    assert (tsh.get(engine, 'smallserie2') == ts).all()
718
    assert_hist_equals(tsh.get_history(engine, 'smallserie2'), histts)
719
720
721
722

    # get history ranges
    tsa = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2))
723
    assert_hist("""
724
725
726
727
728
729
insertion_date             value_date
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
2017-02-03 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
                           2017-01-03    2.0
730
731
732
733
""", tsa)

    tsb = tsh.get_history(engine, 'smallserie',
                          to_insertion_date=datetime(2017, 2, 2))
734
    assert_hist("""
735
736
737
738
insertion_date             value_date
2017-02-01 00:00:00+00:00  2017-01-01    0.0
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
739
740
741
742
743
""", tsb)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2),
                          to_insertion_date=datetime(2017, 2, 2))
744
    assert_hist("""
745
746
747
insertion_date             value_date
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
748
749
""", tsc)

750
751
752
    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 4),
                          to_insertion_date=datetime(2017, 2, 4))
753
    assert tsc == {}
754
755
756
757

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2016, 2, 1),
                          to_insertion_date=datetime(2017, 2, 2))
758
    assert_hist("""
759
760
761
762
insertion_date             value_date
2017-02-01 00:00:00+00:00  2017-01-01    0.0
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
763
764
765
766
767
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2016, 2, 1),
                          to_insertion_date=datetime(2016, 12, 31))
768
    assert tsc == {}
769

770
771
772
773
    # restrictions on value dates
    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 1),
                          to_value_date=datetime(2017, 1, 2))
774
    assert_hist("""
775
776
777
778
779
780
insertion_date             value_date
2017-02-01 00:00:00+00:00  2017-01-01    0.0
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
2017-02-03 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
781
782
783
784
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 2))
785
    assert_hist("""
786
787
788
789
insertion_date             value_date
2017-02-02 00:00:00+00:00  2017-01-02    1.0
2017-02-03 00:00:00+00:00  2017-01-02    1.0
                           2017-01-03    2.0
790
791
792
793
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          to_value_date=datetime(2017, 1, 2))
794
    assert_hist("""
795
796
797
798
799
800
insertion_date             value_date
2017-02-01 00:00:00+00:00  2017-01-01    0.0
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
2017-02-03 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
801
802
""", tsc)

803
804
805
    tsc = tsh.get_history(engine, 'no-such-series')
    assert tsc is None

806
807
808
809
810
811
812
813
    idates = tsh.insertion_dates(engine, 'smallserie')
    assert idates == [
        pd.Timestamp('2017-02-01 00:00:00+0000', tz='UTC'),
        pd.Timestamp('2017-02-02 00:00:00+0000', tz='UTC'),
        pd.Timestamp('2017-02-03 00:00:00+0000', tz='UTC')
    ]


814

815
816
817
818
819
820
821
def test_history_delta(engine, tsh):
    for d in range(1, 3):
        idate = utcdt(2018, 1, d)
        serie = genserie(idate - timedelta(hours=1), 'H', 6, initval=[d])
        tsh.insert(engine, serie, 'hd', 'aurelien.campeas@pythonian.fr',
                   _insertion_date=idate)

822
    assert_hist("""
823
insertion_date             value_date               
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
2018-01-01 00:00:00+00:00  2017-12-31 23:00:00+00:00    1.0
                           2018-01-01 00:00:00+00:00    1.0
                           2018-01-01 01:00:00+00:00    1.0
                           2018-01-01 02:00:00+00:00    1.0
                           2018-01-01 03:00:00+00:00    1.0
                           2018-01-01 04:00:00+00:00    1.0
2018-01-02 00:00:00+00:00  2017-12-31 23:00:00+00:00    1.0
                           2018-01-01 00:00:00+00:00    1.0
                           2018-01-01 01:00:00+00:00    1.0
                           2018-01-01 02:00:00+00:00    1.0
                           2018-01-01 03:00:00+00:00    1.0
                           2018-01-01 04:00:00+00:00    1.0
                           2018-01-01 23:00:00+00:00    2.0
                           2018-01-02 00:00:00+00:00    2.0
                           2018-01-02 01:00:00+00:00    2.0
                           2018-01-02 02:00:00+00:00    2.0
                           2018-01-02 03:00:00+00:00    2.0
                           2018-01-02 04:00:00+00:00    2.0
    """, tsh.get_history(engine, 'hd'))

844
    assert_hist("""
845
insertion_date             value_date               
846
847
2018-01-01 00:00:00+00:00  2017-12-31 23:00:00+00:00    1.0
                           2018-01-01 00:00:00+00:00    1.0
848
849
                           2018-01-01 01:00:00+00:00    1.0
                           2018-01-01 02:00:00+00:00    1.0
850
851
852
853
854
855
856
857
2018-01-02 00:00:00+00:00  2017-12-31 23:00:00+00:00    1.0
                           2018-01-01 00:00:00+00:00    1.0
                           2018-01-01 01:00:00+00:00    1.0
                           2018-01-01 02:00:00+00:00    1.0
                           2018-01-01 03:00:00+00:00    1.0
                           2018-01-01 04:00:00+00:00    1.0
                           2018-01-01 23:00:00+00:00    2.0
                           2018-01-02 00:00:00+00:00    2.0
858
859
860
861
                           2018-01-02 01:00:00+00:00    2.0
                           2018-01-02 02:00:00+00:00    2.0
""",  tsh.get_history(engine, 'hd', deltaafter=timedelta(hours=2)))

862
    assert_hist("""
863
insertion_date             value_date               
864
2018-01-01 00:00:00+00:00  2018-01-01 00:00:00+00:00    1.0
865
                           2018-01-01 01:00:00+00:00    1.0
866
2018-01-02 00:00:00+00:00  2018-01-02 00:00:00+00:00    2.0
867
868
                           2018-01-02 01:00:00+00:00    2.0
""",  tsh.get_history(engine, 'hd',
869
                      deltabefore=timedelta(hours=0),
870
871
872
                      deltaafter=timedelta(hours=1)))


873
874
875
876
877
878
879
880
881
882
883
def test_nr_gethistory(engine, tsh):
    s0 = pd.Series([-1, 0, 0, -1],
                   index=pd.DatetimeIndex(start=datetime(2016, 12, 29),
                                          end=datetime(2017, 1, 1),
                                          freq='D'))
    tsh.insert(engine, s0, 'foo', 'zogzog')

    s1 = pd.Series([1, 0, 0, 1],
                   index=pd.DatetimeIndex(start=datetime(2017, 1, 1),
                                          end=datetime(2017, 1, 4),
                                          freq='D'))
884
    idate = utcdt(2016, 1, 1)
885
    for i in range(5):
886
        with engine.begin() as cn:
887
888
889
            tsh.insert(cn, s1 * i, 'foo',
                       'aurelien.campeas@pythonian.f',
                       _insertion_date=idate + timedelta(days=i))
890
891
892
893
894
895
896

    df = tsh.get_history(engine, 'foo',
                         datetime(2016, 1, 3),
                         datetime(2016, 1, 4),
                         datetime(2017, 1, 1),
                         datetime(2017, 1, 4))

897
    assert_hist("""
898
899
900
901
902
903
904
905
906
insertion_date             value_date
2016-01-03 00:00:00+00:00  2017-01-01    2.0
                           2017-01-02    0.0
                           2017-01-03    0.0
                           2017-01-04    2.0
2016-01-04 00:00:00+00:00  2017-01-01    3.0
                           2017-01-02    0.0
                           2017-01-03    0.0
                           2017-01-04    3.0
907
908
909
""", df)


910
def test_add_na(engine, tsh):
911
912
913
914
915
    # a serie of NaNs won't be insert in base
    # in case of first insertion
    ts_nan = genserie(datetime(2010, 1, 1), 'D', 5)
    ts_nan[[True] * len(ts_nan)] = np.nan

916
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
917
    assert diff is None
918
919
920

    # in case of insertion in existing data
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 5)
921
    tsh.insert(engine, ts_begin, 'ts_add_na', 'test')
922
923
924
925
926

    ts_nan = genserie(datetime(2010, 1, 6), 'D', 5)
    ts_nan[[True] * len(ts_nan)] = np.nan
    ts_nan = pd.concat([ts_begin, ts_nan])

927
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
928
929
    assert diff is None

930
    result = tsh.get(engine, 'ts_add_na')
931
    assert len(result) == 5
932
933


934
def test_dtype_mismatch(engine, tsh):
935
    tsh.insert(engine,
936
937
938
939
940
               genserie(datetime(2015, 1, 1), 'D', 11).astype('str'),
               'error1',
               'test')

    with pytest.raises(Exception) as excinfo:
941
        tsh.insert(engine,
942
943
944
945
946
947
                   genserie(datetime(2015, 1, 1), 'D', 11),
                   'error1',
                   'test')

    assert 'Type error when inserting error1, new type is float64, type in base is object' == str(excinfo.value)

948
    tsh.insert(engine,
949
950
951
952
953
               genserie(datetime(2015, 1, 1), 'D', 11),
               'error2',
               'test')

    with pytest.raises(Exception) as excinfo:
954
        tsh.insert(engine,
955
956
957
958
959
                   genserie(datetime(2015, 1, 1), 'D', 11).astype('str'),
                   'error2',
                   'test')

    assert 'Type error when inserting error2, new type is object, type in base is float64' == str(excinfo.value)
960
961


962
963
964
965
966
967
def test_precision(engine, tsh):
    floaty = 0.123456789123456789
    ts = genserie(datetime(2015, 1, 1), 'D', 5, initval=[floaty])

    tsh.insert(engine, ts, 'precision', 'test')
    ts_round = tsh.get(engine, 'precision')
968
    assert 0.12345678912345678 == ts_round.iloc[0]
969
970

    diff = tsh.insert(engine, ts_round, 'precision', 'test')
Aurélien Campéas's avatar
Aurélien Campéas committed
971
    assert diff is None  # the roundtriped series does not produce a diff when reinserted
972

Aurélien Campéas's avatar
Aurélien Campéas committed
973
    diff = tsh.insert(engine, ts, 'precision', 'test')  # neither does the original series
974
975
976
    assert diff is None


977
978
979
980
981
982
983
984
def test_serie_deletion(engine, tsh):
    ts = genserie(datetime(2018, 1, 10), 'H', 10)
    tsh.insert(engine, ts, 'keepme', 'Babar')
    tsh.insert(engine, ts, 'deleteme', 'Celeste')
    ts = genserie(datetime(2018, 1, 12), 'H', 10)
    tsh.insert(engine, ts, 'keepme', 'Babar')
    tsh.insert(engine, ts, 'deleteme', 'Celeste')

985
    seriecount, csetcount, csetseriecount = assert_structures(engine, tsh)
986

987
    with engine.begin() as cn:
988
989
990
991
992
993
994
        tsh.delete(cn, 'deleteme')

    assert not tsh.exists(engine, 'deleteme')
    log = [entry['author']
           for entry in tsh.log(engine, names=('keepme', 'deleteme'))]
    assert log == ['Babar', 'Babar']

995
996
997
998
    seriecount2, csetcount2, csetseriecount2 = assert_structures(engine, tsh)

    assert csetcount - csetcount2  == 2
    assert csetseriecount - csetseriecount2 == 2
999
1000
1001
1002
1003
1004
    assert seriecount - seriecount2 == 1

    with pytest.raises(AssertionError) as werr:
        tsh.delete(engine, 'keepme')
    assert werr.value.args[0] == 'use a transaction object'

1005
    tsh.insert(engine, ts, 'deleteme', 'Celeste')
1006

1007

1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
def test_strip(engine, tsh):
    for i in range(1, 5):
        pubdate = utcdt(2017, 1, i)
        ts = genserie(datetime(2017, 1, 10), 'H', 1 + i)
        tsh.insert(engine, ts, 'xserie', 'babar', _insertion_date=pubdate)
        # also insert something completely unrelated
        tsh.insert(engine, genserie(datetime(2018, 1, 1), 'D', 1 + i), 'yserie', 'celeste')

    csida = tsh.changeset_at(engine, 'xserie', datetime(2017, 1, 3))
    assert csida is not None
    csidb = tsh.changeset_at(engine, 'xserie', datetime(2017, 1, 3, 1), mode='before')
    csidc = tsh.changeset_at(engine, 'xserie', datetime(2017, 1, 3, 1), mode='after')
    assert csidb < csida < csidc

    log = tsh.log(engine, names=['xserie', 'yserie'])
    assert [(idx, l['author']) for idx, l in enumerate(log, start=1)
    ] == [
        (1, 'babar'),
        (2, 'celeste'),
        (3, 'babar'),
        (4, 'celeste'),
        (5, 'babar'),
        (6, 'celeste'),
        (7, 'babar'),
        (8, 'celeste')
    ]

    h = tsh.get_history(engine, 'xserie')