def __init__(
self,
*,
start_date: str,
end_date: str,
aggregation_unit: AnySpatialUnit,
label_a: str,
label_b: str,
labels: Dict[str, Dict[str, dict]],
tower_day_of_week_scores: Dict[str, float],
tower_hour_of_day_scores: List[float],
tower_cluster_radius: float = 1.0,
tower_cluster_call_threshold: int = 0,
event_types: Optional[Union[str, List[str]]],
subscriber_subset: Union[dict, None] = None,
):
# Note: all input parameters need to be defined as attributes on `self`
# so that marshmallow can serialise the object correctly.
self.start_date = start_date
self.end_date = end_date
self.aggregation_unit = aggregation_unit
self.event_types = event_types
self.label_a = label_a
self.label_b = label_b
self.labels = labels
self.tower_day_of_week_scores = tower_day_of_week_scores
self.tower_hour_of_day_scores = tower_hour_of_day_scores
self.tower_cluster_radius = tower_cluster_radius
self.tower_cluster_call_threshold = tower_cluster_call_threshold
self.subscriber_subset = subscriber_subset
common_params = dict(
labels=labels,
start_date=start_date,
end_date=end_date,
event_types=event_types,
subscriber_subset=subscriber_subset,
tower_cluster_call_threshold=tower_cluster_call_threshold,
tower_cluster_radius=tower_cluster_radius,
tower_day_of_week_scores=tower_day_of_week_scores,
tower_hour_of_day_scores=tower_hour_of_day_scores,
)
locs_a = _make_meaningful_locations_object(label=label_a,
**common_params)
locs_b = _make_meaningful_locations_object(label=label_b,
**common_params)
self.q_meaningful_locations_od = RedactedMeaningfulLocationsOD(
meaningful_locations_od=MeaningfulLocationsOD(
meaningful_locations_a=locs_a,
meaningful_locations_b=locs_b,
spatial_unit=aggregation_unit,
))
def test_meaningful_locations_od_results(get_dataframe,
meaningful_locations_labels):
"""
Test that OD on MeaningfulLocations returns expected results and counts clusters per subscriber correctly.
"""
mfl_a = MeaningfulLocations(
clusters=HartiganCluster(
calldays=CallDays(subscriber_locations=SubscriberLocations(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
)),
radius=1,
),
scores=EventScore(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
),
labels=meaningful_locations_labels,
label="unknown",
)
mfl_b = MeaningfulLocations(
clusters=HartiganCluster(
calldays=CallDays(subscriber_locations=SubscriberLocations(
start="2016-01-02",
stop="2016-01-03",
spatial_unit=make_spatial_unit("versioned-site"),
)),
radius=1,
),
scores=EventScore(
start="2016-01-02",
stop="2016-01-03",
spatial_unit=make_spatial_unit("versioned-site"),
),
labels=meaningful_locations_labels,
label="unknown",
)
mfl_od = MeaningfulLocationsOD(
meaningful_locations_a=mfl_a,
meaningful_locations_b=mfl_b,
spatial_unit=make_spatial_unit("admin", level=1),
)
mfl_od_df = get_dataframe(mfl_od)
# Smoke test one admin1 region gets the expected result
regional_flow = mfl_od_df[(mfl_od_df.pcod_from == "524 1") &
(mfl_od_df.pcod_to == "524 4")].value.tolist()[0]
assert regional_flow == pytest.approx(16.490_807)
assert mfl_od_df.value.sum() == pytest.approx(454.0)
def test_column_names_meaningful_locations_od(exemplar_spatial_unit_param,
get_column_names_from_run,
meaningful_locations_labels):
"""Test that column_names property matches head(0) for an od matrix between meaningful locations"""
if not exemplar_spatial_unit_param.is_polygon:
pytest.xfail(
f"The spatial unit {exemplar_spatial_unit_param} is not supported as an aggregation unit for ODs between MeaningfulLocations."
)
mfl_a = MeaningfulLocations(
clusters=HartiganCluster(
calldays=CallDays(subscriber_locations=SubscriberLocations(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
)),
radius=1,
),
scores=EventScore(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
),
labels=meaningful_locations_labels,
label="evening",
)
mfl_b = MeaningfulLocations(
clusters=HartiganCluster(
calldays=CallDays(subscriber_locations=SubscriberLocations(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
)),
radius=1,
),
scores=EventScore(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
),
labels=meaningful_locations_labels,
label="unknown",
)
mfl_od = MeaningfulLocationsOD(
meaningful_locations_a=mfl_a,
meaningful_locations_b=mfl_b,
spatial_unit=exemplar_spatial_unit_param,
)
assert get_column_names_from_run(mfl_od) == mfl_od.column_names
def test_meaningful_locations_od_redaction(get_dataframe,
meaningful_locations_labels):
"""
Test that OD on MeaningfulLocations is redacted to >15.
"""
mfl_a = MeaningfulLocations(
clusters=HartiganCluster(
calldays=CallDays(subscriber_locations=SubscriberLocations(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
)),
radius=1,
),
scores=EventScore(
start="2016-01-01",
stop="2016-01-02",
spatial_unit=make_spatial_unit("versioned-site"),
),
labels=meaningful_locations_labels,
label="unknown",
)
mfl_b = MeaningfulLocations(
clusters=HartiganCluster(
calldays=CallDays(subscriber_locations=SubscriberLocations(
start="2016-01-02",
stop="2016-01-03",
spatial_unit=make_spatial_unit("versioned-site"),
)),
radius=1,
),
scores=EventScore(
start="2016-01-02",
stop="2016-01-03",
spatial_unit=make_spatial_unit("versioned-site"),
),
labels=meaningful_locations_labels,
label="unknown",
)
mfl_od = RedactedMeaningfulLocationsOD(
meaningful_locations_od=MeaningfulLocationsOD(
meaningful_locations_a=mfl_a,
meaningful_locations_b=mfl_b,
spatial_unit=make_spatial_unit("admin", level=1),
))
mfl_od_df = get_dataframe(mfl_od)
# Aggregate should not include any counts below 15
assert all(mfl_od_df.value > 15)