def build_timeseries_for_fips(
intervention,
us_latest,
us_timeseries,
model_output_dir,
fips,
) -> Optional[CovidActNowAreaTimeseries]:
fips_latest = us_latest.get_record_for_fips(fips)
if intervention is Intervention.SELECTED_INTERVENTION:
state = fips_latest[CommonFields.STATE]
intervention = get_can_projection.get_intervention_for_state(state)
model_output = CANPyseirLocationOutput.load_from_model_output_if_exists(
fips, intervention, model_output_dir)
if not model_output and intervention is not Intervention.OBSERVED_INTERVENTION:
# All model output is currently tied to a specific intervention. However,
# we want to generate results for areas that don't have a fit result, but we're not
# duplicating non-model outputs.
return None
try:
area_summary = api.generate_area_summary(fips_latest, model_output)
fips_timeseries = us_timeseries.get_subset(None, fips=fips)
area_timeseries = api.generate_area_timeseries(area_summary,
fips_timeseries,
model_output)
except Exception:
logger.error(f"failed to run output", fips=fips)
return None
return area_timeseries
def test_generate_timeseries_for_fips(
include_projections,
nyc_model_output_path,
nyc_region,
nyc_rt_dataset,
nyc_icu_dataset,
):
us_latest = combined_datasets.load_us_latest_dataset()
us_timeseries = combined_datasets.load_us_timeseries_dataset()
nyc_latest = us_latest.get_record_for_fips(nyc_region.fips)
nyc_timeseries = us_timeseries.get_one_region(nyc_region)
intervention = Intervention.OBSERVED_INTERVENTION
model_output = CANPyseirLocationOutput.load_from_path(
nyc_model_output_path)
metrics_series, latest_metric = api_pipeline.generate_metrics_and_latest(
nyc_timeseries, nyc_rt_dataset, nyc_icu_dataset)
region_summary = generate_api.generate_region_summary(
nyc_latest, latest_metric, model_output)
region_timeseries = generate_api.generate_region_timeseries(
region_summary, nyc_timeseries, metrics_series, model_output)
summary = generate_api.generate_region_summary(nyc_latest, latest_metric,
model_output)
assert summary.dict() == region_timeseries.region_summary.dict()
# Double checking that serialized json does not contain NaNs, all values should
# be serialized using the simplejson wrapper.
assert "NaN" not in region_timeseries.json()
def test_load_from_path(nyc_model_output_path):
output = CANPyseirLocationOutput.load_from_path(nyc_model_output_path)
assert output.fips == "36061"
assert output.intervention == Intervention.STRONG_INTERVENTION
# manually checked values
assert output.peak_hospitalizations_date == datetime.datetime(2020, 4, 15)
assert not output.hospitals_shortfall_date
assert output.latest_rt == pytest.approx(0.972686)
def test_hospitalization_date():
rows = [
_build_row(date="2020-12-11", **{schema.ALL_HOSPITALIZED: 10}),
_build_row(date="2020-12-12", **{schema.ALL_HOSPITALIZED: 17}),
_build_row(date="2020-12-13", **{schema.ALL_HOSPITALIZED: 17}),
_build_row(date="2020-12-14", **{schema.ALL_HOSPITALIZED: 8}),
]
data = _build_input_df(rows)
model_output = CANPyseirLocationOutput(data)
# Check that it picks first date of max.
expected_date = datetime.datetime(year=2020, month=12, day=12)
assert model_output.peak_hospitalizations_date == expected_date
def test_no_rt_val():
rows = [
_build_row(date="2020-12-13",
**{
schema.RT_INDICATOR: None,
schema.RT_INDICATOR_CI90: None
}),
]
data = _build_input_df(rows)
model_output = CANPyseirLocationOutput(data)
# Check that it picks first date of max.
assert not model_output.latest_rt
assert not model_output.latest_rt_ci90
def test_shortfall():
rows = [
_build_row(date="2020-12-13", all_hospitalized=10, beds=11),
_build_row(date="2020-12-14", all_hospitalized=12, beds=11),
_build_row(date="2020-12-15", all_hospitalized=13, beds=11),
]
data = _build_input_df(rows)
model_output = CANPyseirLocationOutput(data)
# Check that it picks first date of max.
expected_date = datetime.datetime.fromisoformat("2020-12-14")
assert model_output.hospitals_shortfall_date == expected_date
assert model_output.peak_hospitalizations_shortfall == 2
# No shortfall
rows = [
_build_row(date="2020-12-13", all_hospitalized=10, beds=11),
_build_row(date="2020-12-14", all_hospitalized=12, beds=12),
]
data = _build_input_df(rows)
model_output = CANPyseirLocationOutput(data)
# Check that it picks first date of max.
assert not model_output.hospitals_shortfall_date
def test_pyseir_end_to_end_idaho():
# This covers a lot of edge cases.
# cli._run_all(state='Idaho')
cli._build_all_for_states(states=["Idaho"], generate_reports=False, fips="16001")
path = get_run_artifact_path("16001", RunArtifact.WEB_UI_RESULT).replace(
"__INTERVENTION_IDX__", "2"
)
path = pathlib.Path(path)
assert path.exists()
output = CANPyseirLocationOutput.load_from_path(path)
data = output.data
with_values = data[schema.RT_INDICATOR].dropna()
assert len(with_values) > 10
assert (with_values > 0).all()
assert (with_values < 6).all()
def test_generate_timeseries_for_fips(include_projections,
nyc_model_output_path, nyc_fips):
us_latest = combined_datasets.build_us_latest_with_all_fields()
us_timeseries = combined_datasets.build_us_timeseries_with_all_fields()
nyc_latest = us_latest.get_record_for_fips(nyc_fips)
nyc_timeseries = us_timeseries.get_subset(None, fips=nyc_fips)
intervention = Intervention.OBSERVED_INTERVENTION
model_output = CANPyseirLocationOutput.load_from_path(
nyc_model_output_path)
area_summary = generate_api.generate_area_summary(nyc_latest, model_output)
area_timeseries = generate_api.generate_area_timeseries(
area_summary, nyc_timeseries, model_output)
summary = generate_api.generate_area_summary(nyc_latest, model_output)
assert summary.dict() == area_timeseries.area_summary.dict()
# Double checking that serialized json does not contain NaNs, all values should
# be serialized using the simplejson wrapper.
assert "NaN" not in area_timeseries.json()
def test_build_summary_for_fips(include_projections, nyc_model_output_path,
nyc_fips):
us_latest = combined_datasets.build_us_latest_with_all_fields()
nyc_latest = us_latest.get_record_for_fips(nyc_fips)
model_output = None
expected_projections = None
intervention = Intervention.OBSERVED_INTERVENTION
if include_projections:
model_output = CANPyseirLocationOutput.load_from_path(
nyc_model_output_path)
expected_projections = Projections(
totalHospitalBeds=ResourceUsageProjection(peakShortfall=0,
peakDate=datetime.date(
2020, 4, 15),
shortageStartDate=None),
ICUBeds=None,
Rt=model_output.latest_rt,
RtCI90=model_output.latest_rt_ci90,
)
intervention = Intervention.STRONG_INTERVENTION
summary = generate_api.generate_region_summary(nyc_latest, model_output)
expected = RegionSummary(
population=nyc_latest["population"],
stateName="New York",
countyName="New York County",
fips="36061",
lat=None,
long=None,
actuals=Actuals(
population=nyc_latest["population"],
intervention="STRONG_INTERVENTION",
cumulativeConfirmedCases=nyc_latest["cases"],
cumulativeDeaths=nyc_latest["deaths"],
cumulativePositiveTests=nyc_latest["positive_tests"],
cumulativeNegativeTests=nyc_latest["negative_tests"],
hospitalBeds={
# Manually calculated from capacity calculation in generate_api.py
"capacity": 12763,
"totalCapacity": nyc_latest["max_bed_count"],
"currentUsageCovid": 0,
"currentUsageTotal": None,
"typicalUsageRate": nyc_latest["all_beds_occupancy_rate"],
},
ICUBeds={
"capacity": nyc_latest["icu_beds"],
"totalCapacity": nyc_latest["icu_beds"],
"currentUsageCovid": 0,
"currentUsageTotal": 0,
"typicalUsageRate": nyc_latest["icu_occupancy_rate"],
},
contactTracers=nyc_latest["contact_tracers_count"],
),
lastUpdatedDate=datetime.datetime.utcnow(),
projections=expected_projections,
)
import pprint
pprint.pprint(expected.actuals.ICUBeds.dict())
pprint.pprint(summary.actuals.ICUBeds.dict())
assert expected.dict() == summary.dict()
