def reporting_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_end_date = il_electricity_cdd_hdd_hourly["blackout_end_date"]
reporting_meter_data, warnings = get_reporting_data(
meter_data, start=blackout_end_date)
preliminary_hourly_design_matrix = create_caltrack_hourly_preliminary_design_matrix(
reporting_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 test_serialize_caltrack_hourly_model(segmented_design_matrices,
occupancy_lookup, temperature_bins,
temps):
segmented_model = fit_caltrack_hourly_model(segmented_design_matrices,
occupancy_lookup,
temperature_bins)
assert json.dumps(segmented_model.json())
def test_fit_caltrack_hourly_model(segmented_design_matrices, occupancy_lookup,
temperature_bins, temps):
segmented_model = fit_caltrack_hourly_model(segmented_design_matrices,
occupancy_lookup,
temperature_bins)
assert segmented_model.segment_models is not None
prediction = segmented_model.predict(temps.index, temps).result
def test_fit_caltrack_hourly_model(
segmented_design_matrices, occupancy_lookup, temperature_bins, temps
):
segmented_model_results = fit_caltrack_hourly_model(
segmented_design_matrices, occupancy_lookup, temperature_bins
)
assert segmented_model_results.model.segment_models is not None
assert str(segmented_model_results).startswith("CalTRACKHourlyModelResults")
prediction = segmented_model_results.predict(temps.index, temps).result
def test_predict_caltrack_hourly_model_empty_models(
temps, segmented_design_matrices_empty_models, occupancy_lookup, temperature_bins
):
segmented_model_results = fit_caltrack_hourly_model(
segmented_design_matrices_empty_models, occupancy_lookup, temperature_bins
)
assert segmented_model_results.model.segment_models is not None
prediction = segmented_model_results.predict(temps.index, temps).result
assert prediction.shape[0] == 24
assert prediction.dropna().shape[0] == 0
def test_fit_caltrack_hourly_model_nans_less_than_week_fit(
segmented_design_matrices_nans_less_than_week,
occupancy_lookup_nans_less_than_week,
temperature_bins_nans_less_than_week,
temps_extended,
):
segmented_model = fit_caltrack_hourly_model(
segmented_design_matrices_nans_less_than_week,
occupancy_lookup_nans_less_than_week,
temperature_bins_nans_less_than_week,
)
assert segmented_model.segment_models is not None
prediction = segmented_model.predict(temps_extended.index,
temps_extended).result
assert prediction.shape[0] == 168
assert prediction.dropna().shape[0] == 4
def test_fit_caltrack_hourly_model_nans_less_than_week_predict(
segmented_design_matrices_nans,
occupancy_lookup_nans,
temperature_bins_nans,
temps_extended,
temps,
):
segmented_model = fit_caltrack_hourly_model(segmented_design_matrices_nans,
occupancy_lookup_nans,
temperature_bins_nans)
assert segmented_model.segment_models is not None
assert segmented_model.model_lookup["jan"].model is not None
assert segmented_model.model_lookup["may"].model is None
assert (segmented_model.model_lookup["may"].warnings[0].qualified_name ==
"eemeter.fit_caltrack_hourly_model_segment.no_nonnull_data")
prediction = segmented_model.predict(temps.index, temps).result
assert prediction.shape[0] == 24
assert prediction["predicted_usage"].sum().round() == 955.0
def test_fit_caltrack_hourly_model_nans_less_than_week_predict(
segmented_design_matrices_nans,
occupancy_lookup_nans,
temperature_bins_nans,
temps_extended,
temps,
):
segmented_model_results = fit_caltrack_hourly_model(
segmented_design_matrices_nans,
occupancy_lookup_nans,
temperature_bins_nans,
temperature_bins_nans,
)
assert segmented_model_results.model.segment_models is not None
assert segmented_model_results.model.model_lookup["jan"].model is not None
assert segmented_model_results.model.model_lookup["may"].model is not None
assert segmented_model_results.model.model_lookup["may"].warnings == []
prediction = segmented_model_results.predict(temps.index, temps).result
assert prediction.shape[0] == 24
assert prediction["predicted_usage"].sum().round() == 955.0
