def test_pandas_inferable_temporal_frequencies(case):
input_series = case["dates"]
expected_infer_freq = case["expected_infer_freq"]
inferred_freq = infer_frequency(observed_ts=input_series, debug=False)
assert inferred_freq == expected_infer_freq
def test_error_messages(case):
input_series = case["dates"]
expected_debug_obj = case["expected_debug_obj"]
inferred_freq, actual_debug_obj = infer_frequency(observed_ts=input_series,
debug=True)
assert inferred_freq is None
assert actual_debug_obj == expected_debug_obj
def test_inferable_temporal_frequencies_cases(case):
input_series = case["dates"]
expected_debug_obj = case["expected_debug_obj"]
inferred_freq, actual_debug_obj = infer_frequency(observed_ts=input_series,
debug=True)
assert inferred_freq is None
assert actual_debug_obj == expected_debug_obj
def test_inferable_temporal_frequencies_duplicates(freq, error_range):
# strip off first element, since it probably doesn't agree with freq
dates = ((pd.date_range("2005-01-01", periods=1001,
freq=freq)[1:]).to_series().reset_index(drop=True))
actual_range_start = dates.loc[0].isoformat()
actual_range_end = dates.loc[len(dates) - 1].isoformat()
idx = len(dates) // 2
d = dates[idx].isoformat()
dates_observed = pd.concat([
dates[:idx + 1],
pd.Series(np.full((error_range, ), dates[idx])),
dates[idx + 1:],
])
dates_observed = dates_observed.reset_index(drop=True)
expected_debug_obj = {
"actual_range_start":
actual_range_start,
"actual_range_end":
actual_range_end,
"message":
None,
"estimated_freq":
freq,
"estimated_range_start":
dates_observed.loc[0].isoformat(),
"estimated_range_end":
dates_observed.loc[len(dates_observed) - 1].isoformat(),
"duplicate_values": [
{
"dt": d,
"idx": idx + 1,
"range": error_range
},
],
"missing_values": [],
"extra_values": [],
"nan_values": [],
}
inferred_freq, actual_debug_obj = infer_frequency(
observed_ts=dates_observed, debug=True)
assert inferred_freq is None
assert actual_debug_obj == expected_debug_obj