def main():
parser = ArgumentParser()
parser.add_argument('--group',
choices=['first', 'second', 'everything'],
default='everything')
parser.add_argument('--area')
parser.add_argument('-t', '--type', choices=['first', 'second'])
add_date_arg(parser,
help='first release to use',
default=earliest_vaccination)
add_date_arg(parser,
'--to-date',
default=find_latest('vaccination_cum_*')[1])
parser.add_argument('--duration', type=float, default=0.2)
parser.add_argument('--raise-errors', action='store_true')
args = parser.parse_args()
if args.type:
to_show = tuple_product_array(
[' '.join(w.capitalize() for w in args.area.split('-'))],
args.type.capitalize())
name = f'{args.area}_{args.type}'
else:
to_show = selection_mapping(args.from_date)[args.group]
name = args.group
dates = pd.date_range(args.from_date, args.to_date)
parallel_render(f'animated_vaccinations_{name}',
partial(render_plots, to_show, args.to_date),
dates,
duration=args.duration,
raise_errors=args.raise_errors)
def summary_data(series,
data_date=None,
start=None,
end=None,
nation='england'):
if data_date in (None, '*'):
data_path, data_date = find_latest(f'{nation}_*.csv')
else:
data_path = base_path / f'{nation}_{pd.to_datetime(data_date).date()}.csv'
data = read_csv(data_path,
start,
end, [s_.metric for s_ in series],
index_col=[date_col]) / 7
return data, data_date
def weekly_data():
raw_path, _ = find_latest('vaccination_old_style_2021-04-08.csv')
raw = read_csv(raw_path)
raw.sort_values([date_col, area_code], inplace=True)
weekly = raw[[
date_col, area_code,
'weeklyPeopleVaccinatedFirstDoseByVaccinationDate',
'weeklyPeopleVaccinatedSecondDoseByVaccinationDate'
]].dropna()
weekly.rename(errors='raise',
inplace=True,
columns={
'weeklyPeopleVaccinatedFirstDoseByVaccinationDate':
any_cov,
'weeklyPeopleVaccinatedSecondDoseByVaccinationDate':
full_cov
})
return weekly.set_index([date_col, area_code]).groupby(level=-1).cumsum()
def raw_vaccination_data(dt='*', sanity_checks: bool = True):
if dt == '*':
dt = '????-*'
else:
dt = pd.to_datetime(dt).date()
data_path, data_date = find_latest(f'vaccination_{dt}.csv')
raw = read_csv(data_path)
raw.sort_values([date_col, area_code], inplace=True)
if sanity_checks:
complete = raw[[
complete_dose_publish_cum, second_dose_publish_cum,
complete_dose_publish_new, second_dose_publish_new
]].dropna(how='any')
cum_equal = (complete[complete_dose_publish_cum] ==
complete[second_dose_publish_cum]).all()
new_equal = ((complete[complete_dose_publish_new] ==
complete[second_dose_publish_new]).all())
assert raw[complete_dose_publish_cum].isnull().all() or (cum_equal
and new_equal)
return raw, data_date
def best_data(dt='*',
area_type=ltla,
areas=None,
earliest=None,
days=None,
metric=new_cases_by_specimen_date,
file_prefix: str = None,
metrics=(),
date_index=False):
metrics = list(metrics) if metrics else [metric]
if file_prefix is None:
file_prefix = area_type
if area_type == msoa:
assert dt == '*'
data_path = base_path / 'msoa_composite.csv'
data = read_csv(data_path)
data_date = pd.to_datetime(data.iloc[-1][release_timestamp])
else:
try:
data_path, data_date = find_latest(f'{file_prefix}_{dt}.csv')
except FileNotFoundError:
if metric != [new_cases_by_specimen_date]:
raise
area_type_filter = area_type_filters.get(area_type)
if area_type_filter is None:
raise
data_path, data_date = find_latest(f'coronavirus-cases_{dt}.csv')
data = pd.read_csv(data_path, parse_dates=[specimen_date])
data = data[data['Area type'].isin(area_type_filter)]
data.rename(inplace=True,
errors='raise',
columns={
area:
area_name,
code:
area_code,
specimen_date:
date_col,
'Daily lab-confirmed cases':
new_cases_by_specimen_date,
})
else:
data = read_csv(data_path)
if days:
earliest = datetime.combine(data_date - timedelta(days=days),
datetime.min.time())
if earliest:
data = data[data[date_col] >= pd.to_datetime(earliest)]
if areas:
data = data[data[area_code].isin(areas)]
missing = []
if data.empty:
missing = metrics
else:
for metric in metrics:
series = data.get(metric)
if series is None or series.empty:
missing.append(metric)
if missing:
missing = ', '.join(missing)
raise NoData(
f'No {missing} for {file_prefix} in {areas} available in {data_path}'
)
if date_index:
data = data.set_index(date_col).sort_index()
return data, data_date
def latest_map_data():
path, dt = find_latest('zoe_prevalence_map_*.pickle', date_index=-2)
df = read_pickle(path)
gdf = convert_df(df, 'the_geom_webmercator')
return dt, gdf
def check_path(path):
path, dt = find_latest(Path(path).name)
checker = Checker(dt, path)
for row in add_blank_rows(tqdm_dict_reader(path)):
checker.add_row(row)
checker.check()
def load_prevalence(source, dates, data_date):
data_path, data_date = find_latest(f'{source}_*.csv', on_or_before=data_date)
data = pd.read_csv(data_path, parse_dates=[dates], index_col=[dates])
return data.sort_index(), data_date