def testFindFrequency(self):
dates = list(
pd.date_range(start='2020-01-01', end='2020-02-01', freq='D'))
geos = [1, 2, 3, 4]
df = pd.DataFrame({
'date': dates * len(geos),
'geo': sorted(geos * len(dates))
})
df.set_index(['geo', 'date'], inplace=True)
frequency = util.infer_frequency(df, 'date', 'geo')
self.assertEqual(frequency, 'D')
weeks = list(
pd.date_range(start='2020-01-01', end='2020-02-01', freq='W'))
df = pd.DataFrame({
'date': weeks * len(geos),
'geo': sorted(geos * len(weeks))
})
df.set_index(['geo', 'date'], inplace=True)
frequency = util.infer_frequency(df, 'date', 'geo')
self.assertEqual(frequency, 'W')
def testUnknownFrequency(self):
dates = list(pd.to_datetime(['2020-10-10', '2020-10-13',
'2020-10-16']))
geos = [1, 2]
df = pd.DataFrame({
'date': dates * len(geos),
'geo': sorted(geos * len(dates))
})
df.set_index(['geo', 'date'], inplace=True)
with self.assertRaises(ValueError) as cm:
_ = util.infer_frequency(df, 'date', 'geo')
self.assertEqual(str(cm.exception),
'Frequency could not be identified. Got 3 days.')
def testFindFrequencyDataNotSorted(self):
dates = list(
pd.date_range(start='2020-01-01', end='2020-02-01', freq='D'))
geos = [1, 2, 3, 4]
df = pd.DataFrame({
'date': dates * len(geos),
'geo': sorted(geos * len(dates))
})
# permute the order of the rows, so that the dataset is not sorted by date
df = df.sample(frac=1, replace=False)
df.set_index(['geo', 'date'], inplace=True)
frequency = util.infer_frequency(df, 'date', 'geo')
self.assertEqual(frequency, 'D')
def testDifferentFrequencies(self):
dates = list(
pd.date_range(start='2020-01-01', end='2020-02-01', freq='D'))
weeks = list(
pd.date_range(start='2020-01-01', end='2020-02-01', freq='W'))
geos = [1] * len(dates) + [2] * len(weeks)
df = pd.DataFrame({'date': dates + weeks, 'geo': geos})
df.set_index(['geo', 'date'], inplace=True)
with self.assertRaises(ValueError) as cm:
_ = util.infer_frequency(df, 'date', 'geo')
self.assertEqual(
str(cm.exception),
'The provided time series seem to have irregular frequencies.')
def testInsufficientData(self):
dates = list(
pd.date_range(start='2020-01-01', end='2020-01-01', freq='D'))
geos = [1, 2]
df = pd.DataFrame({
'date': dates * len(geos),
'geo': sorted(geos * len(dates))
})
df.set_index(['geo', 'date'], inplace=True)
with self.assertRaises(ValueError) as cm:
_ = util.infer_frequency(df, 'date', 'geo')
self.assertEqual(
str(cm.exception),
'At least one series with more than one observation must be provided.'
)