def test_get_baseline_data_with_ignored_gap(
il_electricity_cdd_hdd_billing_monthly):
meter_data = il_electricity_cdd_hdd_billing_monthly["meter_data"]
baseline_data, warnings = get_baseline_data(
meter_data,
end=datetime(2016, 11, 9, tzinfo=pytz.UTC),
max_days=45,
ignore_billing_period_gap_for_day_count=True,
)
assert baseline_data.shape == (2, 1)
assert round(baseline_data.value.sum(), 2) == 632.31
assert len(warnings) == 0
baseline_data, warnings = get_baseline_data(
meter_data,
end=datetime(2016, 11, 9, tzinfo=pytz.UTC),
max_days=45,
ignore_billing_period_gap_for_day_count=False,
)
assert baseline_data.shape == (1, 1)
assert round(baseline_data.value.sum(), 2) == 0
assert len(warnings) == 0
baseline_data, warnings = get_baseline_data(
meter_data,
end=datetime(2016, 11, 9, tzinfo=pytz.UTC),
max_days=25,
ignore_billing_period_gap_for_day_count=True,
)
assert baseline_data.shape == (1, 1)
assert round(baseline_data.value.sum(), 2) == 0
assert len(warnings) == 0
def test_get_baseline_data_with_timezones(il_electricity_cdd_hdd_hourly):
meter_data = il_electricity_cdd_hdd_hourly["meter_data"]
baseline_data, warnings = get_baseline_data(
meter_data.tz_convert("America/New_York"))
assert len(warnings) == 0
baseline_data, warnings = get_baseline_data(
meter_data.tz_convert("Australia/Sydney"))
assert len(warnings) == 0
def test_get_baseline_data_with_end(il_electricity_cdd_hdd_hourly):
meter_data = il_electricity_cdd_hdd_hourly["meter_data"]
blackout_start_date = il_electricity_cdd_hdd_hourly["blackout_start_date"]
baseline_data, warnings = get_baseline_data(meter_data,
end=blackout_start_date)
assert meter_data.shape != baseline_data.shape == (8761, 1)
assert len(warnings) == 0
def baseline_model_hourly(il_electricity_cdd_hdd_hourly):
meter_data = il_electricity_cdd_hdd_hourly["meter_data"]
temperature_data = il_electricity_cdd_hdd_hourly["temperature_data"]
blackout_start_date = il_electricity_cdd_hdd_hourly["blackout_start_date"]
baseline_meter_data, warnings = get_baseline_data(meter_data,
end=blackout_start_date)
preliminary_hourly_design_matrix = create_caltrack_hourly_preliminary_design_matrix(
baseline_meter_data, temperature_data)
segmentation = segment_time_series(preliminary_hourly_design_matrix.index,
"three_month_weighted")
occupancy_lookup = estimate_hour_of_week_occupancy(
preliminary_hourly_design_matrix, segmentation=segmentation)
occupied_temperature_bins, unoccupied_temperature_bins = fit_temperature_bins(
preliminary_hourly_design_matrix,
segmentation=segmentation,
occupancy_lookup=occupancy_lookup,
)
design_matrices = create_caltrack_hourly_segmented_design_matrices(
preliminary_hourly_design_matrix,
segmentation,
occupancy_lookup,
occupied_temperature_bins,
unoccupied_temperature_bins,
)
segmented_model = fit_caltrack_hourly_model(
design_matrices,
occupancy_lookup,
occupied_temperature_bins,
unoccupied_temperature_bins,
)
return segmented_model
def baseline_data_daily(il_electricity_cdd_hdd_daily):
meter_data = il_electricity_cdd_hdd_daily["meter_data"]
temperature_data = il_electricity_cdd_hdd_daily["temperature_data"]
blackout_start_date = il_electricity_cdd_hdd_daily["blackout_start_date"]
baseline_meter_data, warnings = get_baseline_data(meter_data,
end=blackout_start_date)
baseline_data = create_caltrack_daily_design_matrix(
baseline_meter_data, temperature_data)
return baseline_data
def baseline_model_billing(il_electricity_cdd_hdd_billing_monthly):
meter_data = il_electricity_cdd_hdd_billing_monthly["meter_data"]
temperature_data = il_electricity_cdd_hdd_billing_monthly[
"temperature_data"]
blackout_start_date = il_electricity_cdd_hdd_billing_monthly[
"blackout_start_date"]
baseline_meter_data, warnings = get_baseline_data(meter_data,
end=blackout_start_date)
baseline_data = create_caltrack_billing_design_matrix(
baseline_meter_data, temperature_data)
model_results = fit_caltrack_usage_per_day_model(baseline_data,
use_billing_presets=True,
weights_col="n_days_kept")
return model_results
def test_get_baseline_data_start_gap(il_electricity_cdd_hdd_hourly):
meter_data = il_electricity_cdd_hdd_hourly["meter_data"]
start = meter_data.index.min() - timedelta(days=1)
baseline_data, warnings = get_baseline_data(meter_data, start=start)
assert meter_data.shape == baseline_data.shape == (19417, 1)
assert len(warnings) == 1
warning = warnings[0]
assert warning.qualified_name == "eemeter.get_baseline_data.gap_at_baseline_start"
assert (warning.description ==
"Data does not have coverage at requested baseline start date.")
assert warning.data == {
"data_start": "2015-11-22T06:00:00+00:00",
"requested_start": "2015-11-21T06:00:00+00:00",
}
def test_get_baseline_data_n_days_billing_period_overshoot(
il_electricity_cdd_hdd_billing_monthly):
meter_data = il_electricity_cdd_hdd_billing_monthly["meter_data"]
baseline_data, warnings = get_baseline_data(
meter_data,
end=datetime(2017, 11, 9, tzinfo=pytz.UTC),
max_days=45,
allow_billing_period_overshoot=True,
n_days_billing_period_overshoot=45,
ignore_billing_period_gap_for_day_count=True,
)
assert baseline_data.shape == (2, 1)
assert round(baseline_data.value.sum(), 2) == 526.25
assert len(warnings) == 0
def test_get_baseline_data_end_gap(il_electricity_cdd_hdd_hourly):
meter_data = il_electricity_cdd_hdd_hourly["meter_data"]
end = meter_data.index.max() + timedelta(days=1)
baseline_data, warnings = get_baseline_data(meter_data,
end=end,
max_days=None)
assert meter_data.shape == baseline_data.shape == (19417, 1)
assert len(warnings) == 1
warning = warnings[0]
assert warning.qualified_name == "eemeter.get_baseline_data.gap_at_baseline_end"
assert (warning.description ==
"Data does not have coverage at requested baseline end date.")
assert warning.data == {
"data_end": "2018-02-08T06:00:00+00:00",
"requested_end": "2018-02-09T06:00:00+00:00",
}
def test_get_baseline_data_too_far_from_date(
il_electricity_cdd_hdd_billing_monthly):
meter_data = il_electricity_cdd_hdd_billing_monthly["meter_data"]
end_date = datetime(2020, 11, 9, tzinfo=pytz.UTC)
max_days = 45
baseline_data, warnings = get_baseline_data(
meter_data,
end=end_date,
max_days=max_days,
ignore_billing_period_gap_for_day_count=True,
)
assert baseline_data.shape == (2, 1)
assert round(baseline_data.value.sum(), 2) == 1393.4
assert len(warnings) == 0
with pytest.raises(NoBaselineDataError):
get_baseline_data(
meter_data,
end=end_date,
max_days=max_days,
n_days_billing_period_overshoot=45,
ignore_billing_period_gap_for_day_count=True,
)
baseline_data, warnings = get_baseline_data(
meter_data,
end=end_date,
max_days=max_days,
allow_billing_period_overshoot=True,
ignore_billing_period_gap_for_day_count=True,
)
assert baseline_data.shape == (3, 1)
assert round(baseline_data.value.sum(), 2) == 2043.92
assert len(warnings) == 0
# Includes 3 data points because data at index -3 is closer to start target
# then data at index -2
start_target = baseline_data.index[-1] - timedelta(days=max_days)
assert abs((baseline_data.index[0] - start_target).days) < abs(
(baseline_data.index[1] - start_target).days)
with pytest.raises(NoBaselineDataError):
get_baseline_data(
meter_data,
end=end_date,
max_days=max_days,
allow_billing_period_overshoot=True,
n_days_billing_period_overshoot=45,
ignore_billing_period_gap_for_day_count=True,
)
def test_get_baseline_data_empty(il_electricity_cdd_hdd_hourly):
meter_data = il_electricity_cdd_hdd_hourly["meter_data"]
blackout_start_date = il_electricity_cdd_hdd_hourly["blackout_start_date"]
with pytest.raises(NoBaselineDataError):
get_baseline_data(meter_data,
end=pd.Timestamp("2000").tz_localize("UTC"))
def test_get_baseline_data(il_electricity_cdd_hdd_hourly):
meter_data = il_electricity_cdd_hdd_hourly["meter_data"]
baseline_data, warnings = get_baseline_data(meter_data)
assert meter_data.shape == baseline_data.shape == (19417, 1)
assert len(warnings) == 0