test_tsio.py 42.1 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
from tshistory.util import rename_series
12
13
from tshistory.testutil import (
    assert_df,
14
15
    assert_hist,
    assert_hist_equals,
16
17
18
19
    assert_group_equals,
    genserie,
    tempattr
)
20

21
DATADIR = Path(__file__).parent / 'data'
22

Aurélien Campéas's avatar
Aurélien Campéas committed
23

24
25
26
27
def utcdt(*dt):
    return pd.Timestamp(datetime(*dt), tz='UTC')


28
def test_tstamp_roundtrip(engine, tsh):
29
30
    ts = genserie(datetime(2017, 10, 28, 23),
                  'H', 4, tz='UTC')
31
32
33
34
35
36
37
38
39
40
    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)

41
42
    tsh.insert(engine, ts, 'tztest', 'Babar',
               _insertion_date=utcdt(2018, 1, 1))
43
    back = tsh.get(engine, 'tztest')
44
45
46

    # though un localized we understand it's been normalized to utc
    assert_df("""
47
48
49
50
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
51
52
53
""", back)

    assert (ts.index == back.index).all()
54
    assert str(back.index.dtype) == 'datetime64[ns, UTC]'
55

56
57
58
59
    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')

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
    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)

103
104
105
106
    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')

107

108
def test_differential(engine, tsh):
109
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 10)
110
    tsh.insert(engine, ts_begin, 'ts_test', 'test')
111

112
    id1 = tsh.last_id(engine, 'ts_test')
113
    assert tsh._previous_cset(engine, 'ts_test', id1) is None
114

115
116
    assert tsh.exists(engine, 'ts_test')
    assert not tsh.exists(engine, 'this_does_not_exist')
117

118
119
120
121
122
123
124
125
    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')

126
    assert_df("""
127
128
129
130
131
132
133
134
135
136
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
137
""", tsh.get(engine, 'ts_test'))
138
139

    # we should detect the emission of a message
140
    tsh.insert(engine, ts_begin, 'ts_test', 'babar')
141

142
    assert_df("""
143
144
145
146
147
148
149
150
151
152
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
153
""", tsh.get(engine, 'ts_test'))
154
155
156
157

    ts_slight_variation = ts_begin.copy()
    ts_slight_variation.iloc[3] = 0
    ts_slight_variation.iloc[6] = 0
158
    tsh.insert(engine, ts_slight_variation, 'ts_test', 'celeste')
159
    id2 = tsh.last_id(engine, 'ts_test')
160
    assert tsh._previous_cset(engine, 'ts_test', id2) == id1
161

162
    assert_df("""
163
164
165
166
167
168
169
170
171
172
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
173
""", tsh.get(engine, 'ts_test'))
174

175
    ts_longer = genserie(datetime(2010, 1, 3), 'D', 15)
176
177
178
179
    ts_longer.iloc[1] = 2.48
    ts_longer.iloc[3] = 3.14
    ts_longer.iloc[5] = ts_begin.iloc[7]

180
181
    with engine.connect() as cn:
        tsh.insert(cn, ts_longer, 'ts_test', 'test')
182
183
184
    id3 = tsh.last_id(engine, 'ts_test')

    assert id1 < id2 < id3
185

186
    assert_df("""
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
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
204
""", tsh.get(engine, 'ts_test'))
205

206
207
208
209
210
    assert tsh.interval(engine, 'ts_test') == pd.Interval(
        datetime(2010, 1, 1, 0, 0), datetime(2010, 1, 17, 0, 0),
        closed='both'
    )

211
    # start testing manual overrides
212
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 5, initval=[2])
213
    ts_begin.loc['2010-01-04'] = -1
214
    tsh.insert(engine, ts_begin, 'ts_mixte', 'test')
215
216

    # -1 represents bogus upstream data
217
    assert_df("""
218
219
220
221
222
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
223
""", tsh.get(engine, 'ts_mixte'))
224
225

    # refresh all the period + 1 extra data point
226
    ts_more = genserie(datetime(2010, 1, 2), 'D', 5, [2])
227
    ts_more.loc['2010-01-04'] = -1
228
    tsh.insert(engine, ts_more, 'ts_mixte', 'test')
229

230
    assert_df("""
231
232
233
234
235
236
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
237
""", tsh.get(engine, 'ts_mixte'))
238
239

    # just append an extra data point
240
241
    # with no intersection with the previous ts
    ts_one_more = genserie(datetime(2010, 1, 7), 'D', 1, [3])
242
    tsh.insert(engine, ts_one_more, 'ts_mixte', 'test')
243

244
    assert_df("""
245
246
247
248
249
250
251
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
252
""", tsh.get(engine, 'ts_mixte'))
253

254
    with engine.connect() as cn:
255
        cn.execute('set search_path to "{0}.timeserie", {0}, public'.format(tsh.namespace))
256
257
        allts = pd.read_sql("select seriename, table_name from registry "
                            "where seriename in ('ts_test', 'ts_mixte')",
258
                            cn)
259

260
        assert_df("""
261
262
263
seriename table_name
0   ts_test    ts_test
1  ts_mixte   ts_mixte
264
""".format(tsh.namespace), allts)
265

266
        assert_df("""
267
268
269
270
271
272
273
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
274
""", tsh.get(cn, 'ts_mixte',
275
             revision_date=datetime.now()))
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
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 == {
        'tzaware': False,
        'index_type': 'datetime64[ns]',
        'value_type': 'float64',
        'index_names': []
    }

    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})

    tsh.update_metadata(engine, 'ts-metadata', {'tzaware': True}, internal=True)
    assert tsh.metadata(engine, 'ts-metadata') == {
        'tzaware': True,
        'index_type': 'datetime64[ns]',
        'value_type': 'float64',
        'index_names': [],
        'topic': 'banana spot price'
    }
    # unbreak the serie for the second test pass :o
    tsh.update_metadata(engine, 'ts-metadata', initialmeta, internal=True)


310
311
def test_changeset_metadata(engine, tsh):
    serie = genserie(datetime(2010, 1, 1), 'D', 1, initval=[1])
312
    tsh.insert(engine, serie, 'ts-cs-metadata', 'babar',
313
314
               {'foo': 'A', 'bar': 42})

315
    log = tsh.log(engine, names=['ts-cs-metadata'])
316
317
318
319
    meta = tsh.changeset_metadata(engine, log[0]['rev'])
    assert meta == {'foo': 'A', 'bar': 42}


320
def test_bad_import(engine, tsh):
321
    # the data were parsed as date by pd.read_json()
Aurélien Campéas's avatar
Aurélien Campéas committed
322
    df_result = pd.read_csv(str(DATADIR / 'test_data.csv'))
323
324
325
    df_result['Gas Day'] = df_result['Gas Day'].apply(parser.parse, dayfirst=True, yearfirst=False)
    df_result.set_index('Gas Day', inplace=True)
    ts = df_result['SC']
326

327
328
    tsh.insert(engine, ts, 'SND_SC', 'test')
    result = tsh.get(engine, 'SND_SC')
329
    assert result.dtype == 'float64'
330
331
332

    # insertion of empty ts
    ts = pd.Series(name='truc', dtype='object')
333
334
    tsh.insert(engine, ts, 'empty_ts', 'test')
    assert tsh.get(engine, 'empty_ts') is None
335
336
337

    # nan in ts
    # all na
338
    ts = genserie(datetime(2010, 1, 10), 'D', 10, [np.nan], name='truc')
339
    tsh.insert(engine, ts, 'test_nan', 'test')
340
341
    assert len(tsh.get(engine, 'test_nan')) == 0
    assert len(tsh.get(engine, 'test_nan', _keep_nans=True)) == 10
342
343
344
345
346

    # mixe na
    ts = pd.Series([np.nan] * 5 + [3] * 5,
                   index=pd.date_range(start=datetime(2010, 1, 10),
                                       freq='D', periods=10), name='truc')
347
348
    tsh.insert(engine, ts, 'test_nan', 'test')
    result = tsh.get(engine, 'test_nan')
349

350
351
    tsh.insert(engine, ts, 'test_nan', 'test')
    result = tsh.get(engine, 'test_nan')
352
    assert_df("""
353
354
355
356
357
2010-01-15    3.0
2010-01-16    3.0
2010-01-17    3.0
2010-01-18    3.0
2010-01-19    3.0
358
""", result)
359
360

    # get_ts with name not in database
361
    assert tsh.get(engine, 'inexisting_name', 'test') is None
362
363


364
def test_revision_date(engine, tsh):
365
366
    for i in range(1, 5):
        with engine.connect() as cn:
367
368
            tsh.insert(cn, genserie(datetime(2017, 1, i), 'D', 3, [i]), 'revdate',
                       'test', _insertion_date=utcdt(2016, 1, i))
369
370

    # end of prologue, now some real meat
371
    idate0 = pd.Timestamp('2015-1-1 00:00:00', tz='UTC')
372
373
374
375
    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')
376

377
    idate1 = pd.Timestamp('2015-1-1 15:45:23', tz='UTC')
378
379
380
381
    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')
382

383
    idate2 = pd.Timestamp('2015-1-2 15:43:23', tz='UTC')
384
385
386
387
    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')
388

389
    idate3 = pd.Timestamp('2015-1-3', tz='UTC')
390
391
392
393
    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')
394

395
    ts = tsh.get(engine, 'ts_through_time')
396

397
    assert_df("""
398
399
400
401
2010-01-04    3.0
2010-01-05    3.0
2010-01-06    3.0
2010-01-07    3.0
402
""", ts)
403

404
    ts = tsh.get(engine, 'ts_through_time',
Aurélien Campéas's avatar
Aurélien Campéas committed
405
                 revision_date=datetime(2015, 1, 2, 18, 43, 23))
406

407
    assert_df("""
408
409
410
411
2010-01-04    2.0
2010-01-05    2.0
2010-01-06    2.0
2010-01-07    2.0
412
""", ts)
413

414
    ts = tsh.get(engine, 'ts_through_time',
415
                 revision_date=datetime(2015, 1, 1, 18, 43, 23))
416

417
    assert_df("""
418
419
420
421
2010-01-04    1.0
2010-01-05    1.0
2010-01-06    1.0
2010-01-07    1.0
422
""", ts)
423

424
    ts = tsh.get(engine, 'ts_through_time',
425
                 revision_date=datetime(2014, 1, 1, 18, 43, 23))
426
427
428

    assert ts is None

429
430
431
432
433
434
435
436
437
438
    # 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'))

439
    oldstate = tsh.get(engine, 'revdate', revision_date=datetime(2016, 1, 2))
440
441
442
    assert_df("""
2017-01-01    1.0
2017-01-02    2.0
443
444
445
2017-01-03    2.0
2017-01-04    2.0
""", oldstate)
446

447

448
def test_point_deletion(engine, tsh):
449
450
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 11)
    ts_begin.iloc[-1] = np.nan
451
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
452

453
    _, ts = Snapshot(engine, tsh, 'ts_del').find()
454
    assert ts.iloc[-2] == 9.0
455

456
    ts_begin.iloc[0] = np.nan
457
    ts_begin.iloc[3] = np.nan
458

459
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
460

461
    assert_df("""
462
463
464
465
466
467
468
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
469
2010-01-10    9.0
470
""", tsh.get(engine, 'ts_del'))
471

472
    ts2 = tsh.get(engine, 'ts_del',
473
474
                 # force snapshot reconstruction feature
                 revision_date=datetime(2038, 1, 1))
475
    assert (tsh.get(engine, 'ts_del') == ts2).all()
476

477
478
479
    ts_begin.iloc[0] = 42
    ts_begin.iloc[3] = 23

480
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
481

482
    assert_df("""
483
484
485
486
487
488
489
490
491
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
492
2010-01-10     9.0
493
""", tsh.get(engine, 'ts_del'))
494
495
496

    # now with string!

497
    ts_string = genserie(datetime(2010, 1, 1), 'D', 10, ['machin'])
498
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
499
500
501
502

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

503
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
504
    assert_df("""
505
506
507
508
509
510
511
512
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
513
""", tsh.get(engine, 'ts_string_del'))
514
515
516
517

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

518
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
519
    assert_df("""
520
521
522
523
524
525
526
527
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
528
2010-01-10    machin
529
""", tsh.get(engine, 'ts_string_del'))
530

531
    ts_string[ts_string.index] = np.nan
532
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
533

534
    erased = tsh.get(engine, 'ts_string_del')
535
536
    assert len(erased) == 0

537
538
    # first insertion with only nan

539
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 10, [np.nan])
540
    tsh.insert(engine, ts_begin, 'ts_null', 'test')
541

542
    assert len(tsh.get(engine, 'ts_null')) == 0
543

544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
    # exhibit issue with nans handling
    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)

563
564
    tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
    diff = tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
565
566
    assert diff is None

567
    # there is no difference
568
    assert 0 == len(tsh.diff(ts_repushed, ts_repushed))
569
570
571
572
573

    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
574
    diff = tsh.diff(ts_repushed, ts_add)
575
576
577
578
579
580

    assert_df("""
2010-01-02     1.0
2010-01-03     2.0
2010-01-09     NaN
2010-01-12    11.0
581
2010-01-13    12.0""", diff.sort_index())
582
583
584
585
    # value on nan => value
    # nan on value => nan
    # nan on nan => Nothing
    # nan on nothing=> Nothing
586

Aurélien Campéas's avatar
Aurélien Campéas committed
587
    # full erasing
588
589
    # numeric
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4)
590
    tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
591

Aurélien Campéas's avatar
Aurélien Campéas committed
592
    ts_begin.iloc[:] = np.nan
593
    tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
594
595

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4)
596
    tsh.insert(engine, ts_end, 'ts_full_del', 'test')
597
598
599
600

    # string

    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
601
    tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
602
603

    ts_begin.iloc[:] = np.nan
604
    tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
605
606

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
607
    tsh.insert(engine, ts_end, 'ts_full_del_str', 'test')
608

Aurélien Campéas's avatar
Aurélien Campéas committed
609

610
def test_get_history(engine, tsh):
611
612
    for numserie in (1, 2, 3):
        with engine.connect() as cn:
613
614
615
            tsh.insert(cn, genserie(datetime(2017, 1, 1), 'D', numserie), 'smallserie',
                       'aurelien.campeas@pythonian.fr',
                       _insertion_date=utcdt(2017, 2, numserie))
616
617
618
619
620
621
622
623
624
625
626

    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',
627
         'meta': {},
628
         'date': pd.Timestamp('2017-02-01 00:00:00+0000', tz='UTC'),
629
630
631
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
632
         'meta': {},
633
         'date': pd.Timestamp('2017-02-02 00:00:00+0000', tz='UTC'),
634
635
636
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
637
         'meta': {},
638
         'date': pd.Timestamp('2017-02-03 00:00:00+0000', tz='UTC'),
639
640
641
642
643
644
         'names': ['smallserie']
        }
    ] == [{k: v for k, v in log.items() if k != 'rev'}
          for log in logs]
    histts = tsh.get_history(engine, 'smallserie')

645
    assert_hist("""
646
647
648
649
650
651
652
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
653
654
""", histts)

655
    diffs = tsh.get_history(engine, 'smallserie', diffmode=True)
656
    assert_hist("""
657
658
659
660
661
662
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)

663
    for idate in histts:
664
        with engine.connect() as cn:
665
            idate = idate.replace(tzinfo=pytz.timezone('UTC'))
666
667
            tsh.insert(cn, histts[idate], 'smallserie2',
                       'aurelien.campeas@pythonian.f', _insertion_date=idate)
668
669
670

    # this is perfectly round-tripable
    assert (tsh.get(engine, 'smallserie2') == ts).all()
671
    assert_hist_equals(tsh.get_history(engine, 'smallserie2'), histts)
672
673
674
675

    # get history ranges
    tsa = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2))
676
    assert_hist("""
677
678
679
680
681
682
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
683
684
685
686
""", tsa)

    tsb = tsh.get_history(engine, 'smallserie',
                          to_insertion_date=datetime(2017, 2, 2))
687
    assert_hist("""
688
689
690
691
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
692
693
694
695
696
""", tsb)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2),
                          to_insertion_date=datetime(2017, 2, 2))
697
    assert_hist("""
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
701
702
""", tsc)

703
704
705
706
707
708
709
710
    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 4),
                          to_insertion_date=datetime(2017, 2, 4))
    assert tsc is None

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2016, 2, 1),
                          to_insertion_date=datetime(2017, 2, 2))
711
    assert_hist("""
712
713
714
715
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
716
717
718
719
720
721
722
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2016, 2, 1),
                          to_insertion_date=datetime(2016, 12, 31))
    assert tsc is None

723
724
725
726
    # restrictions on value dates
    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 1),
                          to_value_date=datetime(2017, 1, 2))
727
    assert_hist("""
728
729
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
2017-02-03 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
734
735
736
737
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 2))
738
    assert_hist("""
739
740
741
742
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
743
744
745
746
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          to_value_date=datetime(2017, 1, 2))
747
    assert_hist("""
748
749
750
751
752
753
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
754
755
""", tsc)

756

757
758
759
760
761
762
763
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)

764
    assert_hist("""
765
insertion_date             value_date               
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
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'))

786
    assert_hist("""
787
insertion_date             value_date               
788
789
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
790
791
                           2018-01-01 01:00:00+00:00    1.0
                           2018-01-01 02:00:00+00:00    1.0
792
793
794
795
796
797
798
799
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
800
801
802
803
                           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)))

804
    assert_hist("""
805
insertion_date             value_date               
806
2018-01-01 00:00:00+00:00  2018-01-01 00:00:00+00:00    1.0
807
                           2018-01-01 01:00:00+00:00    1.0
808
2018-01-02 00:00:00+00:00  2018-01-02 00:00:00+00:00    2.0
809
810
                           2018-01-02 01:00:00+00:00    2.0
""",  tsh.get_history(engine, 'hd',
811
                      deltabefore=timedelta(hours=0),
812
813
814
                      deltaafter=timedelta(hours=1)))


815
816
817
818
819
820
821
822
823
824
825
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'))
826
    idate = utcdt(2016, 1, 1)
827
828
    for i in range(5):
        with engine.connect() as cn:
829
830
831
            tsh.insert(cn, s1 * i, 'foo',
                       'aurelien.campeas@pythonian.f',
                       _insertion_date=idate + timedelta(days=i))
832
833
834
835
836
837
838

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

839
    assert_hist("""
840
841
842
843
844
845
846
847
848
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
849
850
851
""", df)


852
def test_add_na(engine, tsh):
853
854
855
856
857
    # 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

858
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
859
    assert len(diff) == 5
860
    result = tsh.get(engine, 'ts_add_na')
861
862
863
864
865
866
867
868
869
870
    assert len(result) == 0

    result = tsh.get(engine, 'ts_add_na', _keep_nans=True)
    assert_df("""
2010-01-01   NaN
2010-01-02   NaN
2010-01-03   NaN
2010-01-04   NaN
2010-01-05   NaN
""", result)
871
872
873

    # in case of insertion in existing data
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 5)
874
    tsh.insert(engine, ts_begin, 'ts_add_na', 'test')
875
876
877
878
879

    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])

880
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
881
882
    assert diff is None

883
    result = tsh.get(engine, 'ts_add_na')
884
    assert len(result) == 5
885
886


887
def test_dtype_mismatch(engine, tsh):
888
    tsh.insert(engine,
889
890
891
892
893
               genserie(datetime(2015, 1, 1), 'D', 11).astype('str'),
               'error1',
               'test')

    with pytest.raises(Exception) as excinfo:
894
        tsh.insert(engine,
895
896
897
898
899
900
                   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)

901
    tsh.insert(engine,
902
903
904
905
906
               genserie(datetime(2015, 1, 1), 'D', 11),
               'error2',
               'test')

    with pytest.raises(Exception) as excinfo:
907
        tsh.insert(engine,
908
909
910
911
912
                   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)
913
914


915
916
917
918
919
920
921
922
923
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')
    assert 0.12345678912346 == ts_round.iloc[0]

    diff = tsh.insert(engine, ts_round, 'precision', 'test')
Aurélien Campéas's avatar
Aurélien Campéas committed
924
    assert diff is None  # the roundtriped series does not produce a diff when reinserted
925

Aurélien Campéas's avatar
Aurélien Campéas committed
926
    diff = tsh.insert(engine, ts, 'precision', 'test')  # neither does the original series
927
928
929
    assert diff is None


930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
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')

    seriecount = engine.execute(
        'select count(*) from {}.registry'.format(tsh.namespace)
    ).scalar()
    csetcount = engine.execute(
        'select count(*) from {}.changeset'.format(tsh.namespace)
    ).scalar()
    csetcount2 = engine.execute(
        'select count(*) from {}.changeset_series'.format(tsh.namespace)
    ).scalar()
    assert csetcount == csetcount2

    with engine.connect() as cn:
        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']

    csetcount3 = engine.execute(
        'select count(*) from {}.changeset'.format(tsh.namespace)
    ).scalar()
    csetcount4 = engine.execute(
        'select count(*) from {}.changeset_series'.format(tsh.namespace)
    ).scalar()
    seriecount2 = engine.execute(
        'select count (*) from {}.registry'.format(tsh.namespace)
    ).scalar()

    assert csetcount - csetcount3  == 2
    assert csetcount2 - csetcount4 == 2
    assert seriecount - seriecount2 == 1

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

975
    tsh.insert(engine, ts, 'deleteme', 'Celeste')
976

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
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')
1006
    assert_hist("""
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
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))
    with engine.connect() as cn:
        tsh.strip(cn, 'xserie', csid)

1028
    assert_hist("""
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
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
""", tsh.get_history(engine, 'xserie'))

    assert_df("""
2017-01-10 00:00:00    0.0
2017-01-10 01:00:00    1.0
2017-01-10 02:00:00    2.0
""", tsh.get(engine, 'xserie'))

    log = tsh.log(engine, names=['xserie', 'yserie'])
    # 5 and 7 have disappeared
    assert [l['author'] for l in log
    ] == ['babar', 'celeste', 'babar', 'celeste', 'celeste', 'celeste']

    log = tsh.log(engine, stripped=True, names=['xserie', 'yserie'])
1049
1050
1051
1052
1053
1054
    for l in log:
        if l['meta']:
            meta = l['meta']
            stripinfo = meta.get('tshistory.info')
            if stripinfo:
                assert stripinfo.startswith('got stripped from')
1055
1056


1057
1058
1059
def test_long_name(engine, tsh):
    serie = genserie(datetime(2010, 1, 1), 'D', 40)

1060
1061
1062
    name = 'a' * 64
    tsh.insert(engine, serie, name, 'babar')
    assert tsh.get(engine, name) is not None
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073


def test_get_delta(engine, tsh):
    for idate in pd.DatetimeIndex(start=utcdt(2015, 1, 1),
                                  end=utcdt(2015, 1, 1, 3),
                                  freq='H'):
        ts = genserie(start=idate, freq='H', repeat=7)
        tsh.insert(engine, ts, 'republication', 'test',
                   _insertion_date=idate)

    hist = tsh.get_history(engine, 'republication')
1074
    assert_hist("""
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
insertion_date             value_date               
2015-01-01 00:00:00+00:00  2015-01-01 00:00:00+00:00    0.0
                           2015-01-01 01:00:00+00:00    1.0
                           2015-01-01 02:00:00+00:00    2.0
                           2015-01-01 03:00:00+00:00    3.0
                           2015-01-01 04:00:00+00:00    4.0
                           2015-01-01 05:00:00+00:00    5.0
                           2015-01-01 06:00:00+00:00    6.0
2015-01-01 01:00:00+00:00  2015-01-01 00:00:00+00:00    0.0
                           2015-01-01 01:00:00+00:00    0.0
                           2015-01-01 02:00:00+00:00    1.0
                           2015-01-01 03:00:00+00:00    2.0
                           2015-01-01 04:00:00+00:00    3.0
                           2015-01-01 05:00:00+00:00    4.0
                           2015-01-01 06:00:00+00:00    5.0
                           2015-01-01 07:00:00+00:00    6.0
2015-01-01 02:00:00+00:00  2015-01-01 00:00:00+00:00    0.0
                           2015-01-01 01:00:00+00:00    0.0
                           2015-01-01 02:00:00+00:00    0.0
                           2015-01-01 03:00:00+00:00    1.0
                           2015-01-01 04:00:00+00:00    2.0
                           2015-01-01 05:00:00+00:00    3.0
                           2015-01-01 06:00:00+00:00    4.0
                           2015-01-01 07:00:00+00:00    5.0
                           2015-01-01 08:00:00+00:00    6.0
2015-01-01 03:00:00+00:00  2015-01-01 00:00:00+00:00    0.0
                           2015-01-01 01:00:00+00:00    0.0
                           2015-01-01 02:00:00+00:00    0.0
                           2015-01-01 03:00:00+00:00    0.0
                           2015-01-01 04:00:00+00:00    1.0
                           2015-01-01 05:00:00+00:00    2.0
                           2015-01-01 06:00:00+00:00    3.0
                           2015-01-01 07:00:00+00:00    4.0
                           2015-01-01 08:00:00+00:00    5.0
                           2015-01-01 09:00:00+00:00    6.0
""", hist)

    deltas = tsh.get_delta(engine,  'republication', delta=timedelta(hours=3))

    assert_df("""
2015-01-01 03:00:00+00:00    3.0
2015-01-01 04:00:00+00:00    3.0
2015-01-01 05:00:00+00:00    3.0
2015-01-01 06:00:00+00:00    3.0
2015-01-01 07:00:00+00:00    4.0
2015-01-01 08:00:00+00:00    5.0
2015-01-01 09:00:00+00:00    6.0
""", deltas)

    deltas = tsh.get_delta(engine,  'republication', delta=timedelta(hours=5))
    assert_df("""
2015-01-01 05:00:00+00:00    5.0
2015-01-01 06:00:00+00:00    5.0
2015-01-01 07:00:00+00:00    5.0
2015-01-01 08:00:00+00:00    5.0
2015-01-01 09:00:00+00:00    6.0
""", deltas)

    hist = tsh.get_history(engine, 'republication',
                           deltabefore=-timedelta(hours=3),
                           deltaafter