def test_find_with_dataframe(self, jhu_data, population_data):
population = population_data.value("Italy")
clean_df = jhu_data.cleaned()
warnings.filterwarnings("ignore", category=DeprecationWarning)
change_finder = ChangeFinder(clean_df, population, country="Italy")
change_finder.run()
phase_series = change_finder.show()
assert isinstance(phase_series, PhaseSeries)
def test_find_with_small_min_size(self, jhu_data, population_data):
population = population_data.value("Italy")
clean_df = jhu_data.cleaned()
warnings.filterwarnings("ignore", category=DeprecationWarning)
with pytest.raises(ValueError):
min_size = 2
change_finder = ChangeFinder(clean_df,
population,
country="Italy",
min_size=min_size)
change_finder.run()
def test_find_with_few_records(self, jhu_data, population_data):
population = population_data.value("Italy")
min_size = 7
df = jhu_data.subset(country="Italy")
start_date = df.loc[df.index[0], Term.DATE].strftime(Term.DATE_FORMAT)
end_date = Term.date_change(start_date, days=min_size - 2)
sr_df = jhu_data.to_sr(country="Italy",
population=population,
end_date=end_date)
with pytest.raises(ValueError):
change_finder = ChangeFinder(sr_df, min_size=min_size)
change_finder.run()
def test_find_with_few_records(self, jhu_data, population_data):
min_size = 7
population = population_data.value("Italy")
clean_df = jhu_data.subset("Italy")
clean_df[Term.COUNTRY] = "Italy"
clean_df[Term.PROVINCE] = Term.UNKNOWN
clean_df = clean_df.iloc[:(min_size - 1), :]
warnings.filterwarnings("ignore", category=DeprecationWarning)
with pytest.raises(ValueError):
change_finder = ChangeFinder(clean_df,
population,
country="Italy",
min_size=min_size)
change_finder.run()
def test_find(self, jhu_data, population_data, country, max_rmsle=20.0):
# Setup
population = population_data.value(country)
sr_df = jhu_data.to_sr(country=country, population=population)
# Find change points
change_finder = ChangeFinder(sr_df, max_rmsle=max_rmsle)
change_finder.run()
# For all phases, check if RMSLE score is lower than max_rmsle=20.0
for (start_date, end_date) in zip(*change_finder.date_range()):
phase_df = jhu_data.to_sr(country=country,
population=population,
start_date=start_date,
end_date=end_date)
rmsle = Trend(sr_df=phase_df).rmsle()
assert rmsle < max_rmsle
def test_find(self, jhu_data, population_data, country):
warnings.simplefilter("ignore", category=DeprecationWarning)
warnings.simplefilter("ignore", category=UserWarning)
population = population_data.value(country)
sr_df = jhu_data.to_sr(country=country, population=population)
change_finder = ChangeFinder(sr_df)
change_finder.run()
change_finder.show(area=jhu_data.area_name(country=country))
assert isinstance(change_finder.date_range(), tuple)
def test_min_size(self, jhu_data, population_data, country):
# Dataset
population = population_data.value(country)
subset_df = jhu_data.subset(country=country, population=population)
# Too small min_size
with pytest.raises(ValueError):
TrendDetector(data=subset_df, min_size=2)
# Too large min_size
with pytest.raises(ValueError):
TrendDetector(data=subset_df, min_size=100000)
# Reset
detector = TrendDetector(data=subset_df)
detector.reset()
# Deprecated
warnings.filterwarnings("ignore", category=DeprecationWarning)
ChangeFinder(subset_df)
Trend(subset_df)
def test_find(self, jhu_data, population_data, country):
population = population_data.value(country)
change_finder = ChangeFinder(jhu_data, population, country=country)
change_finder.run()
phase_series = change_finder.show()
assert isinstance(phase_series, PhaseSeries)
def test_find_with_small_min_size(self, jhu_data, population_data):
population = population_data.value("Italy")
sr_df = jhu_data.to_sr(country="Italy", population=population)
with pytest.raises(ValueError):
change_finder = ChangeFinder(sr_df, min_size=2)
change_finder.run()