test_tsio.py 50.7 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
    with engine.begin() as cn:
279
        cn.execute('set search_path to "{0}.timeserie", {0}, public'.format(tsh.namespace))
280
281
        allts = pd.read_sql("select seriename, table_name from registry "
                            "where seriename in ('ts_test', 'ts_mixte')",
282
                            cn)
283

284
        assert_df("""
285
286
287
seriename table_name
0   ts_test    ts_test
1  ts_mixte   ts_mixte
288
""".format(tsh.namespace), allts)
289

290
        assert_df("""
291
292
293
294
295
296
297
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
298
""", tsh.get(cn, 'ts_mixte',
299
             revision_date=datetime.now()))
300

301
302
    assert_structures(engine, tsh)

303

304
305
306
307
308
309
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 == {
310
311
        'index_dtype': '<M8[ns]',
        'index_names': [],
312
        'index_type': 'datetime64[ns]',
313
314
315
        'tzaware': False,
        'value_dtype': '<f8',
        'value_type': 'float64'
316
317
318
319
320
321
322
323
324
325
326
    }

    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') == {
327
        'index_dtype': '<M8[ns]',
328
        'index_names': [],
329
330
        'index_type': 'datetime64[ns]',
        'topic': 'banana spot price',
331
        'tzaware': False,
332
333
        'value_dtype': '<f8',
        'value_type': 'float64'
334
    }
335

336

337
338
def test_changeset_metadata(engine, tsh):
    serie = genserie(datetime(2010, 1, 1), 'D', 1, initval=[1])
339
    tsh.insert(engine, serie, 'ts-cs-metadata', 'babar',
340
341
               {'foo': 'A', 'bar': 42})

342
    log = tsh.log(engine, names=['ts-cs-metadata'])
343
344
345
346
    meta = tsh.changeset_metadata(engine, log[0]['rev'])
    assert meta == {'foo': 'A', 'bar': 42}


347
def test_revision_date(engine, tsh):
348
    for i in range(1, 5):
349
        with engine.begin() as cn:
350
351
            tsh.insert(cn, genserie(datetime(2017, 1, i), 'D', 3, [i]), 'revdate',
                       'test', _insertion_date=utcdt(2016, 1, i))
352
353

    # end of prologue, now some real meat
354
    idate0 = pd.Timestamp('2015-1-1 00:00:00', tz='UTC')
355
356
357
358
    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')
359

360
    idate1 = pd.Timestamp('2015-1-1 15:45:23', tz='UTC')
361
362
363
364
    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')
365

366
    idate2 = pd.Timestamp('2015-1-2 15:43:23', tz='UTC')
367
368
369
370
    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')
371

372
    idate3 = pd.Timestamp('2015-1-3', tz='UTC')
373
374
375
376
    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')
377

378
    ts = tsh.get(engine, 'ts_through_time')
379

380
    assert_df("""
381
382
383
384
2010-01-04    3.0
2010-01-05    3.0
2010-01-06    3.0
2010-01-07    3.0
385
""", ts)
386

387
    ts = tsh.get(engine, 'ts_through_time',
Aurélien Campéas's avatar
Aurélien Campéas committed
388
                 revision_date=datetime(2015, 1, 2, 18, 43, 23))
389

390
    assert_df("""
391
392
393
394
2010-01-04    2.0
2010-01-05    2.0
2010-01-06    2.0
2010-01-07    2.0
395
""", ts)
396

397
    ts = tsh.get(engine, 'ts_through_time',
398
                 revision_date=datetime(2015, 1, 1, 18, 43, 23))
399

400
    assert_df("""
401
402
403
404
2010-01-04    1.0
2010-01-05    1.0
2010-01-06    1.0
2010-01-07    1.0
405
""", ts)
406

407
    ts = tsh.get(engine, 'ts_through_time',
408
                 revision_date=datetime(2014, 1, 1, 18, 43, 23))
409
410
411

    assert ts is None

412
413
414
415
416
417
418
419
420
421
    # 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'))

422
    oldstate = tsh.get(engine, 'revdate', revision_date=datetime(2016, 1, 2))
423
424
425
    assert_df("""
2017-01-01    1.0
2017-01-02    2.0
426
427
428
2017-01-03    2.0
2017-01-04    2.0
""", oldstate)
429

430

431
def test_point_deletion(engine, tsh):
432
433
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 11)
    ts_begin.iloc[-1] = np.nan
434
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
435

436
    _, ts = Snapshot(engine, tsh, 'ts_del').find()
437
    assert ts.iloc[-2] == 8.0
438

439
    ts_begin.iloc[0] = np.nan
440
    ts_begin.iloc[3] = np.nan
441

442
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
443

444
    assert_df("""
445
446
447
448
449
450
451
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
452
2010-01-10    9.0
453
""", tsh.get(engine, 'ts_del'))
454

455
    ts2 = tsh.get(engine, 'ts_del',
456
457
                 # force snapshot reconstruction feature
                 revision_date=datetime(2038, 1, 1))
458
    assert (tsh.get(engine, 'ts_del') == ts2).all()
459

460
461
462
    ts_begin.iloc[0] = 42
    ts_begin.iloc[3] = 23

463
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
464

465
    assert_df("""
466
467
468
469
470
471
472
473
474
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
475
2010-01-10     9.0
476
""", tsh.get(engine, 'ts_del'))
477
478
479

    # now with string!

480
    ts_string = genserie(datetime(2010, 1, 1), 'D', 10, ['machin'])
481
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
482
483
484
485

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

486
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
487
    assert_df("""
488
489
490
491
492
493
494
495
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
496
""", tsh.get(engine, 'ts_string_del'))
497
498
499
500

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

501
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
502
    assert_df("""
503
504
505
506
507
508
509
510
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
511
2010-01-10    machin
512
""", tsh.get(engine, 'ts_string_del'))
513

514
    ts_string[ts_string.index] = np.nan
515
516
    with pytest.raises(ValueError):
        tsh.insert(engine, ts_string, 'ts_string_del', 'test')
517

518

519
520
def test_nan_first(engine, tsh):
    # first insertion with only nan
521
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 10, [np.nan])
522
    assert tsh.insert(engine, ts_begin, 'ts_null', 'test') is None
523

524

525
def test_more_point_deletion(engine, tsh):
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
    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)

544
545
    tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
    diff = tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
546
547
    assert diff is None

548
    # there is no difference
549
    assert 0 == len(tsh.diff(ts_repushed, ts_repushed))
550
551
552
553
554

    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
555
    diff = tsh.diff(ts_repushed, ts_add)
556
557
558
559
560
561

    assert_df("""
2010-01-02     1.0
2010-01-03     2.0
2010-01-09     NaN
2010-01-12    11.0
562
2010-01-13    12.0""", diff.sort_index())
563
564
565
566
    # value on nan => value
    # nan on value => nan
    # nan on nan => Nothing
    # nan on nothing=> Nothing
567

Aurélien Campéas's avatar
Aurélien Campéas committed
568
    # full erasing
569
570
    # numeric
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4)
571
    tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
572

Aurélien Campéas's avatar
Aurélien Campéas committed
573
    ts_begin.iloc[:] = np.nan
574
575
    with pytest.raises(ValueError):
        tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
576
577

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4)
578
    tsh.insert(engine, ts_end, 'ts_full_del', 'test')
579
580
581
582

    # string

    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
583
    tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
584

585
586
    ts_begin = pd.Series([np.nan] * 4, name='ts_full_del_str',
                         index=ts_begin.index)
587
588
589

    with pytest.raises(ValueError):
        tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
590
591

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
592
    tsh.insert(engine, ts_end, 'ts_full_del_str', 'test')
593

Aurélien Campéas's avatar
Aurélien Campéas committed
594

595
def test_deletion_over_horizon(engine, tsh):
596
    idate = utcdt(2018, 2, 1)
597
598
599
600
601
602
    ts = pd.Series(
        [1, 2, 3],
        index=pd.date_range(datetime(2018, 1, 1), freq='D', periods=3)
    )

    name = 'delete_over_hz'
603
604
    tsh.insert(engine, ts, name, 'Babar',
               _insertion_date=idate)
605
606
607
608
609
610

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

611
612
    tsh.insert(engine, ts, name, 'Celeste',
               _insertion_date=idate.replace(day=2))
613
614
    ival = tsh.interval(engine, name)
    assert ival.left == datetime(2018, 1, 1)
615
    assert ival.right == datetime(2018, 1, 2)
616
617
618
619
620

    ts = pd.Series(
        [np.nan, np.nan, np.nan],
        index=pd.date_range(datetime(2017, 12, 30), freq='D', periods=3)
    )
621
622
    tsh.insert(engine, ts, name, 'Arthur',
               _insertion_date=idate.replace(day=3))
623
    ival = tsh.interval(engine, name)
624
625
    assert ival.left == datetime(2018, 1, 2)
    assert ival.right == datetime(2018, 1, 2)
626
627


628
def test_get_history(engine, tsh):
629
    for numserie in (1, 2, 3):
630
        with engine.begin() as cn:
631
632
633
            tsh.insert(cn, genserie(datetime(2017, 1, 1), 'D', numserie), 'smallserie',
                       'aurelien.campeas@pythonian.fr',
                       _insertion_date=utcdt(2017, 2, numserie))
634
635
636
637
638
639
640
641
642
643
644

    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',
645
         'meta': {},
646
         'date': pd.Timestamp('2017-02-01 00:00:00+0000', tz='UTC'),
647
648
649
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
650
         'meta': {},
651
         'date': pd.Timestamp('2017-02-02 00:00:00+0000', tz='UTC'),
652
653
654
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
655
         'meta': {},
656
         'date': pd.Timestamp('2017-02-03 00:00:00+0000', tz='UTC'),
657
658
659
660
661
662
         'names': ['smallserie']
        }
    ] == [{k: v for k, v in log.items() if k != 'rev'}
          for log in logs]
    histts = tsh.get_history(engine, 'smallserie')

663
    assert_hist("""
664
665
666
667
668
669
670
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
671
672
""", histts)

673
    diffs = tsh.get_history(engine, 'smallserie', diffmode=True)
674
    assert_hist("""
675
676
677
678
679
680
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)

681
    for idate in histts:
682
        with engine.begin() as cn:
683
            idate = idate.replace(tzinfo=pytz.timezone('UTC'))
684
685
            tsh.insert(cn, histts[idate], 'smallserie2',
                       'aurelien.campeas@pythonian.f', _insertion_date=idate)
686
687
688

    # this is perfectly round-tripable
    assert (tsh.get(engine, 'smallserie2') == ts).all()
689
    assert_hist_equals(tsh.get_history(engine, 'smallserie2'), histts)
690
691
692
693

    # get history ranges
    tsa = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2))
694
    assert_hist("""
695
696
697
698
699
700
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
701
702
703
704
""", tsa)

    tsb = tsh.get_history(engine, 'smallserie',
                          to_insertion_date=datetime(2017, 2, 2))
705
    assert_hist("""
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  2017-01-01    0.0
                           2017-01-02    1.0
710
711
712
713
714
""", tsb)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2),
                          to_insertion_date=datetime(2017, 2, 2))
715
    assert_hist("""
716
717
718
insertion_date             value_date
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
719
720
""", tsc)

721
722
723
    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 4),
                          to_insertion_date=datetime(2017, 2, 4))
724
    assert tsc == {}
725
726
727
728

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2016, 2, 1),
                          to_insertion_date=datetime(2017, 2, 2))
729
    assert_hist("""
730
731
732
733
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
734
735
736
737
738
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2016, 2, 1),
                          to_insertion_date=datetime(2016, 12, 31))
739
    assert tsc == {}
740

741
742
743
744
    # restrictions on value dates
    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 1),
                          to_value_date=datetime(2017, 1, 2))
745
    assert_hist("""
746
747
748
749
750
751
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
752
753
754
755
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 2))
756
    assert_hist("""
757
758
759
760
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
761
762
763
764
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          to_value_date=datetime(2017, 1, 2))
765
    assert_hist("""
766
767
768
769
770
771
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
772
773
""", tsc)

774
775
776
    tsc = tsh.get_history(engine, 'no-such-series')
    assert tsc is None

777
778
779
780
781
782
783
784
    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')
    ]


785

786
787
788
789
790
791
792
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)

793
    assert_hist("""
794
insertion_date             value_date               
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
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'))

815
    assert_hist("""
816
insertion_date             value_date               
817
818
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
819
820
                           2018-01-01 01:00:00+00:00    1.0
                           2018-01-01 02:00:00+00:00    1.0
821
822
823
824
825
826
827
828
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
829
830
831
832
                           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)))

833
    assert_hist("""
834
insertion_date             value_date               
835
2018-01-01 00:00:00+00:00  2018-01-01 00:00:00+00:00    1.0
836
                           2018-01-01 01:00:00+00:00    1.0
837
2018-01-02 00:00:00+00:00  2018-01-02 00:00:00+00:00    2.0
838
839
                           2018-01-02 01:00:00+00:00    2.0
""",  tsh.get_history(engine, 'hd',
840
                      deltabefore=timedelta(hours=0),
841
842
843
                      deltaafter=timedelta(hours=1)))


844
845
846
847
848
849
850
851
852
853
854
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'))
855
    idate = utcdt(2016, 1, 1)
856
    for i in range(5):
857
        with engine.begin() as cn:
858
859
860
            tsh.insert(cn, s1 * i, 'foo',
                       'aurelien.campeas@pythonian.f',
                       _insertion_date=idate + timedelta(days=i))
861
862
863
864
865
866
867

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

868
    assert_hist("""
869
870
871
872
873
874
875
876
877
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
878
879
880
""", df)


881
def test_add_na(engine, tsh):
882
883
884
885
886
    # 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

887
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
888
    assert diff is None
889
890
891

    # in case of insertion in existing data
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 5)
892
    tsh.insert(engine, ts_begin, 'ts_add_na', 'test')
893
894
895
896
897

    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])

898
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
899
900
    assert diff is None

901
    result = tsh.get(engine, 'ts_add_na')
902
    assert len(result) == 5
903
904


905
def test_dtype_mismatch(engine, tsh):
906
    tsh.insert(engine,
907
908
909
910
911
               genserie(datetime(2015, 1, 1), 'D', 11).astype('str'),
               'error1',
               'test')

    with pytest.raises(Exception) as excinfo:
912
        tsh.insert(engine,
913
914
915
916
917
918
                   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)

919
    tsh.insert(engine,
920
921
922
923
924
               genserie(datetime(2015, 1, 1), 'D', 11),
               'error2',
               'test')

    with pytest.raises(Exception) as excinfo:
925
        tsh.insert(engine,
926
927
928
929
930
                   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)
931
932


933
934
935
936
937
938
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')
939
    assert 0.12345678912345678 == ts_round.iloc[0]
940
941

    diff = tsh.insert(engine, ts_round, 'precision', 'test')
Aurélien Campéas's avatar
Aurélien Campéas committed
942
    assert diff is None  # the roundtriped series does not produce a diff when reinserted
943

Aurélien Campéas's avatar
Aurélien Campéas committed
944
    diff = tsh.insert(engine, ts, 'precision', 'test')  # neither does the original series
945
946
947
    assert diff is None


948
949
950
951
952
953
954
955
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')

956
    seriecount, csetcount, csetseriecount = assert_structures(engine, tsh)
957

958
    with engine.begin() as cn:
959
960
961
962
963
964
965
        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']

966
967
968
969
    seriecount2, csetcount2, csetseriecount2 = assert_structures(engine, tsh)

    assert csetcount - csetcount2  == 2
    assert csetseriecount - csetseriecount2 == 2
970
971
972
973
974
975
    assert seriecount - seriecount2 == 1

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

976
    tsh.insert(engine, ts, 'deleteme', 'Celeste')
977

978

979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
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')
1007
    assert_hist("""
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
insertion_date             value_date         
2017-01-01 00:00:00+00:00  2017-01-10 00:00:00    0.0
                           2017-01-10 01:00:00    1.0
2017-01-02 00:00:00+00:00  2017-01-10 00:00:00    0.0
                           2017-01-10 01:00:00    1.0
                           2017-01-10 02:00:00    2.0
2017-01-03 00:00:00+00:00  2017-01-10 00:00:00    0.0
                           2017-01-10 01:00:00    1.0
                           2017-01-10 02:00:00    2.0
                           2017-01-10 03:00:00    3.0
2017-01-04 00:00:00+00:00  2017-01-10 00:00:00    0.0
                           2017-01-10 01:00:00    1.0
                           2017-01-10 02:00:00    2.0
                           2017-01-10 03:00:00    3.0
                           2017-01-10 04:00:00    4.0
""", h)

    csid = tsh.changeset_at(engine, 'xserie', datetime(2017, 1, 3))
1026
    with engine.begin() as cn: