test_tsio.py 37.7 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
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')
    back = tsh.get(engine, 'tztest')
43
44
45

    # though un localized we understand it's been normalized to utc
    assert_df("""
46
47
48
49
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
50
51
52
""", back)

    assert (ts.index == back.index).all()
53
    assert str(back.index.dtype) == 'datetime64[ns, UTC]'
54
55


56
def test_differential(engine, tsh):
57
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 10)
58
    tsh.insert(engine, ts_begin, 'ts_test', 'test')
59

60
    id1 = tsh.last_id(engine, 'ts_test')
61
    assert tsh._previous_cset(engine, 'ts_test', id1) is None
62

63
64
    assert tsh.exists(engine, 'ts_test')
    assert not tsh.exists(engine, 'this_does_not_exist')
65

66
    assert_df("""
67
68
69
70
71
72
73
74
75
76
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
77
""", tsh.get(engine, 'ts_test'))
78
79

    # we should detect the emission of a message
80
    tsh.insert(engine, ts_begin, 'ts_test', 'babar')
81

82
    assert_df("""
83
84
85
86
87
88
89
90
91
92
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
93
""", tsh.get(engine, 'ts_test'))
94
95
96
97

    ts_slight_variation = ts_begin.copy()
    ts_slight_variation.iloc[3] = 0
    ts_slight_variation.iloc[6] = 0
98
    tsh.insert(engine, ts_slight_variation, 'ts_test', 'celeste')
99
    id2 = tsh.last_id(engine, 'ts_test')
100
    assert tsh._previous_cset(engine, 'ts_test', id2) == id1
101

102
    assert_df("""
103
104
105
106
107
108
109
110
111
112
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
113
""", tsh.get(engine, 'ts_test'))
114

115
    ts_longer = genserie(datetime(2010, 1, 3), 'D', 15)
116
117
118
119
    ts_longer.iloc[1] = 2.48
    ts_longer.iloc[3] = 3.14
    ts_longer.iloc[5] = ts_begin.iloc[7]

120
    tsh.insert(engine, ts_longer, 'ts_test', 'test')
121
122
123
    id3 = tsh.last_id(engine, 'ts_test')

    assert id1 < id2 < id3
124

125
    assert_df("""
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
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
143
""", tsh.get(engine, 'ts_test'))
144
145

    # start testing manual overrides
146
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 5, initval=[2])
147
    ts_begin.loc['2010-01-04'] = -1
148
    tsh.insert(engine, ts_begin, 'ts_mixte', 'test')
149
150

    # -1 represents bogus upstream data
151
    assert_df("""
152
153
154
155
156
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
157
""", tsh.get(engine, 'ts_mixte'))
158
159

    # refresh all the period + 1 extra data point
160
    ts_more = genserie(datetime(2010, 1, 2), 'D', 5, [2])
161
    ts_more.loc['2010-01-04'] = -1
162
    tsh.insert(engine, ts_more, 'ts_mixte', 'test')
163

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

    # just append an extra data point
174
175
    # with no intersection with the previous ts
    ts_one_more = genserie(datetime(2010, 1, 7), 'D', 1, [3])
176
    tsh.insert(engine, ts_one_more, 'ts_mixte', 'test')
177

178
    assert_df("""
179
180
181
182
183
184
185
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
186
""", tsh.get(engine, 'ts_mixte'))
187

188
    with engine.connect() as cn:
189
        cn.execute('set search_path to "{0}.timeserie", {0}, public'.format(tsh.namespace))
190
191
        allts = pd.read_sql("select seriename, table_name from registry "
                            "where seriename in ('ts_test', 'ts_mixte')",
192
                            cn)
193

194
        assert_df("""
195
196
197
seriename table_name
0   ts_test    ts_test
1  ts_mixte   ts_mixte
198
""".format(tsh.namespace), allts)
199

200
        assert_df("""
201
202
203
204
205
206
207
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
208
""", tsh.get(cn, 'ts_mixte',
209
             revision_date=datetime.now()))
210
211


212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
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)


244
245
def test_changeset_metadata(engine, tsh):
    serie = genserie(datetime(2010, 1, 1), 'D', 1, initval=[1])
246
    tsh.insert(engine, serie, 'ts-cs-metadata', 'babar',
247
248
               {'foo': 'A', 'bar': 42})

249
    log = tsh.log(engine, names=['ts-cs-metadata'])
250
251
252
253
    meta = tsh.changeset_metadata(engine, log[0]['rev'])
    assert meta == {'foo': 'A', 'bar': 42}


254
def test_bad_import(engine, tsh):
255
    # the data were parsed as date by pd.read_json()
Aurélien Campéas's avatar
Aurélien Campéas committed
256
    df_result = pd.read_csv(str(DATADIR / 'test_data.csv'))
257
258
259
    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']
260

261
262
    tsh.insert(engine, ts, 'SND_SC', 'test')
    result = tsh.get(engine, 'SND_SC')
263
    assert result.dtype == 'float64'
264
265
266

    # insertion of empty ts
    ts = pd.Series(name='truc', dtype='object')
267
268
    tsh.insert(engine, ts, 'empty_ts', 'test')
    assert tsh.get(engine, 'empty_ts') is None
269
270
271

    # nan in ts
    # all na
272
    ts = genserie(datetime(2010, 1, 10), 'D', 10, [np.nan], name='truc')
273
    tsh.insert(engine, ts, 'test_nan', 'test')
274
275
    assert len(tsh.get(engine, 'test_nan')) == 0
    assert len(tsh.get(engine, 'test_nan', _keep_nans=True)) == 10
276
277
278
279
280

    # 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')
281
282
    tsh.insert(engine, ts, 'test_nan', 'test')
    result = tsh.get(engine, 'test_nan')
283

284
285
    tsh.insert(engine, ts, 'test_nan', 'test')
    result = tsh.get(engine, 'test_nan')
286
    assert_df("""
287
288
289
290
291
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
292
""", result)
293
294

    # get_ts with name not in database
295
    assert tsh.get(engine, 'inexisting_name', 'test') is None
296
297


298
def test_revision_date(engine, tsh):
299
    # we prepare a good joke for the end of the test
300
301
    # ival = Snapshot._interval
    # Snapshot._interval = 3
302
303
304

    for i in range(1, 5):
        with engine.connect() as cn:
305
306
            tsh.insert(cn, genserie(datetime(2017, 1, i), 'D', 3, [i]), 'revdate',
                       'test', _insertion_date=utcdt(2016, 1, i))
307
308

    # end of prologue, now some real meat
309
    idate0 = pd.Timestamp('2015-1-1 00:00:00', tz='UTC')
310
311
312
313
    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')
314

315
    idate1 = pd.Timestamp('2015-1-1 15:45:23', tz='UTC')
316
317
318
319
    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')
320

321
    idate2 = pd.Timestamp('2015-1-2 15:43:23', tz='UTC')
322
323
324
325
    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')
326

327
    idate3 = pd.Timestamp('2015-1-3', tz='UTC')
328
329
330
331
    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')
332

333
    ts = tsh.get(engine, 'ts_through_time')
334

335
    assert_df("""
336
337
338
339
2010-01-04    3.0
2010-01-05    3.0
2010-01-06    3.0
2010-01-07    3.0
340
""", ts)
341

342
    ts = tsh.get(engine, 'ts_through_time',
Aurélien Campéas's avatar
Aurélien Campéas committed
343
                 revision_date=datetime(2015, 1, 2, 18, 43, 23))
344

345
    assert_df("""
346
347
348
349
2010-01-04    2.0
2010-01-05    2.0
2010-01-06    2.0
2010-01-07    2.0
350
""", ts)
351

352
    ts = tsh.get(engine, 'ts_through_time',
353
                 revision_date=datetime(2015, 1, 1, 18, 43, 23))
354

355
    assert_df("""
356
357
358
359
2010-01-04    1.0
2010-01-05    1.0
2010-01-06    1.0
2010-01-07    1.0
360
""", ts)
361

362
    ts = tsh.get(engine, 'ts_through_time',
363
                 revision_date=datetime(2014, 1, 1, 18, 43, 23))
364
365
366

    assert ts is None

367
368
369
370
371
372
373
374
375
376
    # 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'))

377
    oldstate = tsh.get(engine, 'revdate', revision_date=datetime(2016, 1, 2))
378
379
380
    assert_df("""
2017-01-01    1.0
2017-01-02    2.0
381
382
383
2017-01-03    2.0
2017-01-04    2.0
""", oldstate)
384

385

386
def test_deletion(engine, tsh):
387
388
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 11)
    ts_begin.iloc[-1] = np.nan
389
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
390

391
    _, ts = Snapshot(engine, tsh, 'ts_del').find()
392
    assert ts.iloc[-2] == 9.0
393

394
    ts_begin.iloc[0] = np.nan
395
    ts_begin.iloc[3] = np.nan
396

397
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
398

399
    assert_df("""
400
401
402
403
404
405
406
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
407
2010-01-10    9.0
408
""", tsh.get(engine, 'ts_del'))
409

410
    ts2 = tsh.get(engine, 'ts_del',
411
412
                 # force snapshot reconstruction feature
                 revision_date=datetime(2038, 1, 1))
413
    assert (tsh.get(engine, 'ts_del') == ts2).all()
414

415
416
417
    ts_begin.iloc[0] = 42
    ts_begin.iloc[3] = 23

418
    tsh.insert(engine, ts_begin, 'ts_del', 'test')
419

420
    assert_df("""
421
422
423
424
425
426
427
428
429
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
430
2010-01-10     9.0
431
""", tsh.get(engine, 'ts_del'))
432
433
434

    # now with string!

435
    ts_string = genserie(datetime(2010, 1, 1), 'D', 10, ['machin'])
436
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
437
438
439
440

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

441
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
442
    assert_df("""
443
444
445
446
447
448
449
450
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
451
""", tsh.get(engine, 'ts_string_del'))
452
453
454
455

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

456
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
457
    assert_df("""
458
459
460
461
462
463
464
465
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
466
2010-01-10    machin
467
""", tsh.get(engine, 'ts_string_del'))
468

469
    ts_string[ts_string.index] = np.nan
470
    tsh.insert(engine, ts_string, 'ts_string_del', 'test')
471

472
    erased = tsh.get(engine, 'ts_string_del')
473
474
    assert len(erased) == 0

475
476
    # first insertion with only nan

477
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 10, [np.nan])
478
    tsh.insert(engine, ts_begin, 'ts_null', 'test')
479

480
    assert len(tsh.get(engine, 'ts_null')) == 0
481

482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
    # 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)

501
502
    tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
    diff = tsh.insert(engine, ts_repushed, 'ts_repushed', 'test')
503
504
    assert diff is None

505
    # there is no difference
506
    assert 0 == len(tsh.diff(ts_repushed, ts_repushed))
507
508
509
510
511

    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
512
    diff = tsh.diff(ts_repushed, ts_add)
513
514
515
516
517
518

    assert_df("""
2010-01-02     1.0
2010-01-03     2.0
2010-01-09     NaN
2010-01-12    11.0
519
2010-01-13    12.0""", diff.sort_index())
520
521
522
523
    # value on nan => value
    # nan on value => nan
    # nan on nan => Nothing
    # nan on nothing=> Nothing
524

Aurélien Campéas's avatar
Aurélien Campéas committed
525
    # full erasing
526
527
    # numeric
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4)
528
    tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
529

Aurélien Campéas's avatar
Aurélien Campéas committed
530
    ts_begin.iloc[:] = np.nan
531
    tsh.insert(engine, ts_begin, 'ts_full_del', 'test')
532
533

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4)
534
    tsh.insert(engine, ts_end, 'ts_full_del', 'test')
535
536
537
538

    # string

    ts_begin = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
539
    tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
540
541

    ts_begin.iloc[:] = np.nan
542
    tsh.insert(engine, ts_begin, 'ts_full_del_str', 'test')
543
544

    ts_end = genserie(datetime(2010, 1, 1), 'D', 4, ['text'])
545
    tsh.insert(engine, ts_end, 'ts_full_del_str', 'test')
546

Aurélien Campéas's avatar
Aurélien Campéas committed
547

548
def test_get_history(engine, tsh):
549
550
    for numserie in (1, 2, 3):
        with engine.connect() as cn:
551
552
553
            tsh.insert(cn, genserie(datetime(2017, 1, 1), 'D', numserie), 'smallserie',
                       'aurelien.campeas@pythonian.fr',
                       _insertion_date=utcdt(2017, 2, numserie))
554
555
556
557
558
559
560
561
562
563
564

    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',
565
         'meta': {},
566
         'date': pd.Timestamp('2017-02-01 00:00:00+0000', tz='UTC'),
567
568
569
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
570
         'meta': {},
571
         'date': pd.Timestamp('2017-02-02 00:00:00+0000', tz='UTC'),
572
573
574
         'names': ['smallserie']
        },
        {'author': 'aurelien.campeas@pythonian.fr',
575
         'meta': {},
576
         'date': pd.Timestamp('2017-02-03 00:00:00+0000', tz='UTC'),
577
578
579
580
581
582
         'names': ['smallserie']
        }
    ] == [{k: v for k, v in log.items() if k != 'rev'}
          for log in logs]
    histts = tsh.get_history(engine, 'smallserie')

583
    assert_hist("""
584
585
586
587
588
589
590
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
591
592
""", histts)

593
    diffs = tsh.get_history(engine, 'smallserie', diffmode=True)
594
    assert_hist("""
595
596
597
598
599
600
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)

601
    for idate in histts:
602
        with engine.connect() as cn:
603
            idate = idate.replace(tzinfo=pytz.timezone('UTC'))
604
605
            tsh.insert(cn, histts[idate], 'smallserie2',
                       'aurelien.campeas@pythonian.f', _insertion_date=idate)
606
607
608

    # this is perfectly round-tripable
    assert (tsh.get(engine, 'smallserie2') == ts).all()
609
    assert_hist_equals(tsh.get_history(engine, 'smallserie2'), histts)
610
611
612
613

    # get history ranges
    tsa = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2))
614
    assert_hist("""
615
616
617
618
619
620
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
621
622
623
624
""", tsa)

    tsb = tsh.get_history(engine, 'smallserie',
                          to_insertion_date=datetime(2017, 2, 2))
625
    assert_hist("""
626
627
628
629
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
630
631
632
633
634
""", tsb)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_insertion_date=datetime(2017, 2, 2),
                          to_insertion_date=datetime(2017, 2, 2))
635
    assert_hist("""
636
637
638
insertion_date             value_date
2017-02-02 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
639
640
""", tsc)

641
642
643
644
645
646
647
648
    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))
649
    assert_hist("""
650
651
652
653
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
654
655
656
657
658
659
660
""", 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

661
662
663
664
    # restrictions on value dates
    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 1),
                          to_value_date=datetime(2017, 1, 2))
665
    assert_hist("""
666
667
668
669
670
671
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
672
673
674
675
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          from_value_date=datetime(2017, 1, 2))
676
    assert_hist("""
677
678
679
680
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
681
682
683
684
""", tsc)

    tsc = tsh.get_history(engine, 'smallserie',
                          to_value_date=datetime(2017, 1, 2))
685
    assert_hist("""
686
687
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
2017-02-03 00:00:00+00:00  2017-01-01    0.0
                           2017-01-02    1.0
692
693
""", tsc)

694

695
696
697
698
699
700
701
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)

702
    assert_hist("""
703
insertion_date             value_date               
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
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'))

724
    assert_hist("""
725
insertion_date             value_date               
726
727
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
728
729
                           2018-01-01 01:00:00+00:00    1.0
                           2018-01-01 02:00:00+00:00    1.0
730
731
732
733
734
735
736
737
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
738
739
740
741
                           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)))

742
    assert_hist("""
743
insertion_date             value_date               
744
2018-01-01 00:00:00+00:00  2018-01-01 00:00:00+00:00    1.0
745
                           2018-01-01 01:00:00+00:00    1.0
746
2018-01-02 00:00:00+00:00  2018-01-02 00:00:00+00:00    2.0
747
748
                           2018-01-02 01:00:00+00:00    2.0
""",  tsh.get_history(engine, 'hd',
749
                      deltabefore=timedelta(hours=0),
750
751
752
                      deltaafter=timedelta(hours=1)))


753
754
755
756
757
758
759
760
761
762
763
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'))
764
    idate = utcdt(2016, 1, 1)
765
766
    for i in range(5):
        with engine.connect() as cn:
767
768
769
            tsh.insert(cn, s1 * i, 'foo',
                       'aurelien.campeas@pythonian.f',
                       _insertion_date=idate + timedelta(days=i))
770
771
772
773
774
775
776

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

777
    assert_hist("""
778
779
780
781
782
783
784
785
786
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
787
788
789
""", df)


790
def test_add_na(engine, tsh):
791
792
793
794
795
    # 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

796
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
797
    assert len(diff) == 5
798
    result = tsh.get(engine, 'ts_add_na')
799
800
801
802
803
804
805
806
807
808
    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)
809
810
811

    # in case of insertion in existing data
    ts_begin = genserie(datetime(2010, 1, 1), 'D', 5)
812
    tsh.insert(engine, ts_begin, 'ts_add_na', 'test')
813
814
815
816
817

    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])

818
    diff = tsh.insert(engine, ts_nan, 'ts_add_na', 'test')
819
820
    assert diff is None

821
    result = tsh.get(engine, 'ts_add_na')
822
    assert len(result) == 5
823
824


825
def test_dtype_mismatch(engine, tsh):
826
    tsh.insert(engine,
827
828
829
830
831
               genserie(datetime(2015, 1, 1), 'D', 11).astype('str'),
               'error1',
               'test')

    with pytest.raises(Exception) as excinfo:
832
        tsh.insert(engine,
833
834
835
836
837
838
                   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)

839
    tsh.insert(engine,
840
841
842
843
844
               genserie(datetime(2015, 1, 1), 'D', 11),
               'error2',
               'test')

    with pytest.raises(Exception) as excinfo:
845
        tsh.insert(engine,
846
847
848
849
850
                   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)
851
852


853
854
855
856
857
858
859
860
861
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
862
    assert diff is None  # the roundtriped series does not produce a diff when reinserted
863

Aurélien Campéas's avatar
Aurélien Campéas committed
864
    diff = tsh.insert(engine, ts, 'precision', 'test')  # neither does the original series
865
866
867
    assert diff is None


868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
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')
896
    assert_hist("""
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
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)

918
    assert_hist("""
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
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'])
939
940
941
942
943
944
    for l in log:
        if l['meta']:
            meta = l['meta']
            stripinfo = meta.get('tshistory.info')
            if stripinfo:
                assert stripinfo.startswith('got stripped from')
945
946


947
948
949
def test_long_name(engine, tsh):
    serie = genserie(datetime(2010, 1, 1), 'D', 40)

950
951
952
    name = 'a' * 64
    tsh.insert(engine, serie, name, 'babar')
    assert tsh.get(engine, name) is not None
953
954
955
956
957
958
959
960
961
962
963


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')
964
    assert_hist("""
965
966
967
968
969
970
971
972
973
974
975
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
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
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=timedelta(hours=3))
1026
    assert_hist("""
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
insertion_date             value_date               
2015-01-01 00:00:00+00:00  2015-01-01 03:00:00+00:00    3.0
2015-01-01 01:00:00+00:00  2015-01-01 04:00:00+00:00    3.0
2015-01-01 02:00:00+00:00  2015-01-01 05:00:00+00:00    3.0
2015-01-01 03:00:00+00:00  2015-01-01 06:00:00+00:00    3.0
""", hist)

    hist = tsh.get_history(engine, 'republication',
                           deltabefore=-timedelta(hours=5),
                           deltaafter=timedelta(hours=5))

1038
    assert_hist("""
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
insertion_date             value_date               
2015-01-01 00:00:00+00:00  2015-01-01 05:00:00+00:00    5.0
2015-01-01 01:00:00+00:00  2015-01-01 06:00:00+00:00    5.0
2015-01-01 02:00:00+00:00  2015-01-01 07:00:00+00:00    5.0
2015-01-01 03:00:00+00:00  2015-01-01 08:00:00+00:00    5.0
""", hist)


    # maybe a more interesting example, each days we insert 7 data points
    for idx, idate in enumerate(pd.DatetimeIndex(start=utcdt(2015, 1, 1),
                                                 end=utcdt(2015, 1, 4),
                                                 freq='D')):
        ts = genserie(start=idate, freq='H', repeat=7)
        tsh.insert(engine, ts, 'repu2', 'test', _insertion_date=idate)

    deltas = tsh.get_delta(engine, 'repu2', delta=timedelta(hours=3))
    assert_df("""
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
2015-01-02 03:00:00+00:00    3.0
2015-01-02 04:00:00+00:00    4.0
2015-01-02 05:00:00+00:00    5.0
2015-01-02 06:00:00+00:00    6.0
2015-01-03 03:00:00+00:00    3.0
2015-01-03 04:00:00+00:00    4.0
2015-01-03 05:00:00+00:00    5.0
2015-01-03 06:00:00+00:00    6.0
2015-01-04 03:00:00+00:00    3.0
2015-01-04 04:00:00+00:00    4.0
2015-01-04 05:00:00+00:00    5.0
2015-01-04 06:00:00+00:00    6.0
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
""", deltas)

    deltas = tsh.get_delta(engine, 'repu2', delta=timedelta(hours=3),
                           from_value_date=datetime(2015,1,2),
                           to_value_date=datetime(2015, 1,3))
    assert_df("""
2015-01-02 03:00:00+00:00    3.0
2015-01-02 04:00:00+00:00    4.0
2015-01-02 05:00:00+00:00    5.0
2015-01-02 06:00:00+00:00    6.0
1082
1083
1084
1085
1086
""", deltas)

    # which is basically the same as below
    hist = tsh.get_history(engine, 'repu2',
                           deltabefore=-timedelta(hours=3))
1087
    assert_hist("""
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
insertion_date             value_date               
2015-01-01 00:00:00+00:00  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
2015-01-02 00:00:00+00:00  2015-01-02 03:00:00+00:00    3.0
                           2015-01-02 04:00:00+00:00    4.0
                           2015-01-02 05:00:00+00:00    5.0
                           2015-01-02 06:00:00+00:00    6.0
2015-01-03 00:00:00+00:00  2015-01-03 03:00:00+00:00    3.0
                           2015-01-03 04:00:00+00:00    4.0
                           2015-01-03 05:00:00+00:00    5.0
                           2015-01-03 06:00:00+00:00    6.0
2015-01-04 00:00:00+00:00  2015-01-04 03:00:00+00:00    3.0
                           2015-01-04 04:00:00+00:00    4.0
                           2015-01-04 05:00:00+00:00    5.0
                           2015-01-04 06:00:00+00:00    6.0
""", hist)
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129


def test_rename(engine, tsh):
    if tsh.namespace == 'zzz':
        return  # this test can only run once

    serie = genserie(datetime(2020, 1, 1), 'D', 3)

    tsh.insert(engine, serie, 'foo', 'Babar')
    tsh.insert(engine, serie, 'bar', 'Babar')
    tsh.insert(engine, serie, 'quux', 'Babar')

    rename_series(engine, {
        'foo': 'new-foo',
        'bar': 'new-bar'
    })

    tsh.resetcaches()

    assert tsh.get(engine, 'foo') is None
    assert tsh.get(engine, 'bar') is None

    for name in ('quux', 'new-foo', 'new-bar'):
        assert tsh.get(engine, name) is not None