def main():
# Create output directory in example directory
code_path = Path(__file__)
output_dir = code_path.with_name("output").joinpath(code_path.stem)
output_dir.mkdir(exist_ok=True, parents=True)
# Read JHU dataset
jhu_file = "input/covid_19_data.csv"
jhu_data = JHUData(jhu_file)
ncov_df = jhu_data.cleaned()
ncov_df.to_csv(output_dir.joinpath("jhu_cleaned.csv"), index=False)
# Read Japan datset
jpn_file = "input/covid_jpn_total.csv"
jpn_data = CountryData(jpn_file, country="Japan")
jpn_data.set_variables(date="Date",
confirmed="Positive",
fatal="Fatal",
recovered="Discharged",
province=None)
jpn_df = jpn_data.cleaned()
jpn_df.to_csv(output_dir.joinpath("jpn_cleaned.csv"), index=False)
# Replace data in Japan with Japan-specific dataset
jhu_data.replace(jpn_data)
ncov_df = jhu_data.cleaned()
ncov_df.to_csv(output_dir.joinpath("jhu_cleaned_replaced.csv"),
index=False)
return ncov_df
def main():
# Create output directory in example directory
code_path = Path(__file__)
output_dir = code_path.with_name("output").joinpath(code_path.stem)
output_dir.mkdir(exist_ok=True, parents=True)
# Read Japan dataset
jpn_file = "input/covid_jpn_total.csv"
jpn_data = CountryData(jpn_file, country="Japan")
jpn_data.set_variables(date="Date",
confirmed="Positive",
fatal="Fatal",
recovered="Discharged",
province=None)
# Show the cleaned data as a CSV file
jpn_df = jpn_data.cleaned()
jpn_df.to_csv(output_dir.joinpath("cleaned.csv"), index=False)
def test_create(self, japan_data):
country_data = CountryData(filename=None, country="Moon")
with pytest.raises(ValueError):
country_data.cleaned()
country_data.raw = japan_data.raw
country_data.set_variables(date="Date",
confirmed="Positive",
fatal="Fatal",
recovered="Discharged",
province="Area")
df = country_data.cleaned()
assert set(df.columns) == set(Term.COLUMNS)
def test_register_total(self):
# Directly download province level data from COVID-19 Data Hub
raw_df, *_ = covid19dh.covid19("Italy", level=2, verbose=False)
filename = "input/italy_raw.csv"
raw_df.to_csv(filename)
# Create CountryData instance
country_data = CountryData(filename=filename, country="Italy")
country_data.set_variables(
date="date", confirmed="confirmed", recovered="recovered", fatal="deaths",
province="administrative_area_level_2"
)
# Register total value of all provinces as country level data
country_data.register_total()
provinces = country_data.cleaned()[Term.PROVINCE].unique()
assert Term.UNKNOWN in provinces
def test_create_province(self, japan_data):
country_data = CountryData(
filename=None, country="Moon", province="Reiner Gamma")
country_data.raw = japan_data.raw
country_data.set_variables(
date="Date",
confirmed="Positive",
fatal="Fatal",
recovered="Discharged",
province=None)
df = country_data.cleaned()
assert set(df.columns) == set(Term.COLUMNS)