def get(self, request):
self.request = request
output_format = 'json'
try:
epidate_s = self._get('epidate',
error_message='epidate sent is empty.')
epidate = datetime.strptime(epidate_s, '%Y-%m-%d')
epi_year_week = episem(epidate, sep='')
if output_format == 'json':
result = json.dumps(
dict(
epi_year_week=epi_year_week,
epi_year=epi_year_week[:4],
epi_week=epi_year_week[4:],
))
else:
result = '' % epi_year_week
except Exception as e:
if output_format == 'json':
result = '{"error_message": "%s"}' % e
else:
result = '[EE] error_message: %s' % e
content_type = ('application/json'
if output_format == 'json' else 'text/plain')
return HttpResponse(result, content_type=content_type)
def get(self, request):
self.request = request
output_format = 'json'
try:
epidate_s = self._get(
'epidate', error_message='epidate sent is empty.'
)
epidate = datetime.strptime(epidate_s, '%Y-%m-%d')
epi_year_week = episem(epidate, sep='')
if output_format == 'json':
result = json.dumps(dict(
epi_year_week=epi_year_week,
epi_year=epi_year_week[:4],
epi_week=epi_year_week[4:],
))
else:
result = '' % epi_year_week
except Exception as e:
if output_format == 'json':
result = '{"error_message": "%s"}' % e
else:
result = '[EE] error_message: %s' % e
content_type = (
'application/json' if output_format == 'json' else 'text/plain'
)
return HttpResponse(result, content_type=content_type)
def _format_data(cls, df: pd.DataFrame) -> pd.DataFrame:
"""
Format the data for plotting.
Parameters
----------
df : pd.DataFrame
Returns
-------
pd.DataFrame
"""
# make a copy
df = df.copy()
df.fillna(0, inplace=True)
if not df.empty:
dfs = []
# merge with a range date dataframe to keep empty week on the data
ts_date = pd.date_range(
df['init_date_week'].min(),
df['init_date_week'].max(),
freq='7D',
)
df_date = pd.DataFrame({'init_date_week': ts_date})
for geocode in df.geocode.unique():
df_ = df[df.geocode == geocode].sort_values('init_date_week')
df_date_ = df_date.set_index(
df_.init_date_week.map(
lambda x: int(episem(str(x)[:10], sep=''))),
drop=True,
)
df_.index.name = None
df_date_.index.name = None
df_['init_date_week'] = pd.to_datetime(df_['init_date_week'],
errors='coerce')
dfs.append(
pd.merge(
df_,
df_date_,
how='outer',
on='init_date_week',
))
df = pd.concat(dfs)
df.sort_index(ascending=True, inplace=True)
for d in DISEASES_SHORT:
k = 'p_rt1_{}'.format(d)
df[k] = (df[k] * 100).fillna(0)
k = 'casos_est_{}'.format(d)
df[k] = df[k].fillna(0).round(0)
k = 'p_inc100k_{}'.format(d)
df[k] = df[k].fillna(0).round(0)
df.rename(
columns={
'p_inc100k_{}'.format(d): 'incidência {}'.format(d),
'casos_{}'.format(d): 'casos notif. {}'.format(d),
'casos_est_{}'.format(d): 'casos est. {}'.format(d),
'p_rt1_{}'.format(d): 'pr(incid. subir) {}'.format(d),
},
inplace=True,
)
df.n_tweets = df.n_tweets.fillna(0).round(0)
return df.rename(columns={
'umid_max': 'umid.max',
'temp_min': 'temp.min',
'n_tweets': 'tweets',
})
def read_disease_data(
cls,
geocode: int,
disease_code: str,
station_id: str,
year_week: int,
var_climate: str,
has_tweets: bool = False,
) -> pd.DataFrame:
"""
:param geocode:
:param disease_code:
:param station_id:
:param year_week:
:param var_climate:
:param has_tweets:
:return:
"""
k = [
var_climate,
'casos',
'p_inc100k',
'casos_est',
'p_rt1',
'nivel',
'level_code',
]
tweet_join = ''
tweet_field = ''
if has_tweets:
k.insert(1, 'n_tweets')
tweet_field = 'tweets.n_tweets,'
tweet_join = '''
LEFT JOIN (
SELECT
epi_week(data_dia) AS "SE",
SUM(numero) as n_tweets,
"Municipio_geocodigo"
FROM "Municipio"."Tweet"
WHERE "Municipio_geocodigo" = %(geocode)s
AND "CID10_codigo" = '%(disease_code)s'
AND epi_week(data_dia) <= %(year_week)s
GROUP BY "SE", "Municipio_geocodigo"
ORDER BY "SE" DESC
) AS tweets
ON (
"Municipio_geocodigo"="municipio_geocodigo"
AND tweets."SE"=hist."SE"
)
'''
disease_table_suffix = ''
if disease_code == CID10['chikungunya']:
disease_table_suffix = '_chik'
elif disease_code == CID10['zika']:
disease_table_suffix = '_zika'
sql = (('''
SELECT
%(var_climate)s,
hist."SE",
''' + tweet_field + '''
hist.casos,
hist.p_rt1,
hist.casos_est,
hist.p_inc100k,
hist.nivel AS level_code,
(CASE
WHEN hist.nivel=1 THEN 'verde'
WHEN hist.nivel=2 THEN 'amarelo'
WHEN hist.nivel=3 THEN 'laranja'
WHEN hist.nivel=4 THEN 'vermelho'
ELSE '-'
END) AS nivel
FROM
"Municipio"."Historico_alerta%(disease_table_suffix)s" AS hist
LEFT JOIN (
SELECT
epi_week(data_dia) AS epiweek,
AVG(%(var_climate)s) AS %(var_climate)s
FROM "Municipio"."Clima_wu"
WHERE "Estacao_wu_estacao_id" = '%(station_id)s'
GROUP BY epiweek
ORDER BY epiweek
) AS climate_wu
ON (hist."SE"=climate_wu.epiweek)
''' + tweet_join + '''
WHERE
hist."SE" <= %(year_week)s
AND municipio_geocodigo=%(geocode)s
ORDER BY "SE" DESC
''') % {
'year_week': year_week,
'geocode': geocode,
'disease_code': disease_code,
'station_id': station_id,
'var_climate': var_climate,
'disease_table_suffix': disease_table_suffix,
})
df = pd.read_sql(sql, index_col='SE', con=db_engine)[k]
df.p_rt1 = (df.p_rt1 * 100).round(0).fillna(0)
df.casos_est = df.casos_est.round(0).fillna(0)
df.p_inc100k = df.p_inc100k.round(0).fillna(0)
if df.empty:
df['init_date_week'] = None
else:
df['init_date_week'] = df.index.map(episem2date)
# merge with a range date dataframe to keep empty week on the data
ts_date = pd.date_range(
df['init_date_week'].min(),
df['init_date_week'].max(),
freq='7D',
)
df_date = pd.DataFrame({'init_date_week': ts_date})
df_date.set_index(
df.init_date_week.map(
lambda x: int(episem(str(x)[:10], sep=''))),
drop=True,
inplace=True,
)
df.index.name = None
df_date.index.name = None
# TODO: these parameters is not valid in the future releases.
# issue #385
df = pd.merge(
df,
df_date,
how='outer',
on='init_date_week',
left_index=True,
right_index=True,
)
df.index.name = 'SE'
df.sort_index(ascending=True, inplace=True)
if has_tweets:
df.n_tweets = df.n_tweets.fillna(0).round(0)
return df.rename(
columns={
'umid_max': 'umid.max',
'temp_min': 'temp.min',
'p_inc100k': 'incidência',
'casos': 'casos notif.',
'n_tweets': 'tweets',
'p_rt1': 'pr(incid. subir)',
})
def _episem(dt):
return episem(dt, sep='')
def _read_disease_data(cls, cities: dict, station_id: str, year_week: int,
var_climate: str) -> pd.DataFrame:
"""
:param cities:
:param station_id:
:param year_week:
:param var_climate:
:return:
"""
k = ['casos', 'p_inc100k', 'casos_est', 'p_rt1', 'nivel', 'level_code']
k = ['{}_{}'.format(v, d) for v in k for d in cls.diseases]
k.append(var_climate)
k.append('n_tweets')
k.append('geocode_dengue AS geocode')
k.append('"SE_dengue" AS "SE"')
k.append('epiweek2date("SE_dengue") AS init_date_week')
general_param = {
'year_week_start': year_week - 100,
'year_week_end': year_week,
'geocodes': ','.join(map(lambda v: "'{}'".format(v), cities)),
'var_climate': var_climate,
'station_id': station_id,
}
sql = ''
previous_disease = ''
for disease in cls.diseases:
_param = dict(general_param)
_param['disease'] = disease
table_suffix = ''
if disease != 'dengue':
table_suffix = '_{}'.format(disease)
_param['table_suffix'] = table_suffix
sql_ = ('''
(SELECT
hist."SE" AS "SE_%(disease)s",
hist.casos AS casos_%(disease)s,
hist.p_rt1 AS p_rt1_%(disease)s,
hist.casos_est AS casos_est_%(disease)s,
hist.p_inc100k AS p_inc100k_%(disease)s,
hist.nivel AS level_code_%(disease)s,
(CASE
WHEN hist.nivel=1 THEN 'verde'
WHEN hist.nivel=2 THEN 'amarelo'
WHEN hist.nivel=3 THEN 'laranja'
WHEN hist.nivel=4 THEN 'vermelho'
ELSE '-'
END) AS nivel_%(disease)s,
hist.municipio_geocodigo AS geocode_%(disease)s
FROM
"Municipio"."Historico_alerta%(table_suffix)s" AS hist
WHERE
hist."SE" BETWEEN %(year_week_start)s AND %(year_week_end)s
AND hist.municipio_geocodigo IN (%(geocodes)s)
ORDER BY "SE_%(disease)s" DESC
) AS %(disease)s
''' % _param)
if not sql:
sql = sql_
else:
sql += '''
LEFT JOIN {0}
ON (
{1}."SE_{1}" = {2}."SE_{2}"
AND {1}.geocode_{1} = {2}.geocode_{2}
)
'''.format(sql_, previous_disease, disease)
previous_disease = disease
tweet_join = ('''
LEFT JOIN (
SELECT
epi_week(data_dia) AS "SE_twitter",
SUM(numero) as n_tweets,
"Municipio_geocodigo"
FROM "Municipio"."Tweet"
WHERE
"Municipio_geocodigo" IN (%(geocodes)s)
AND epi_week(data_dia)
BETWEEN %(year_week_start)s AND %(year_week_end)s
GROUP BY "SE_twitter", "Municipio_geocodigo"
ORDER BY "SE_twitter" DESC
) AS tweets
ON (
"Municipio_geocodigo"=dengue."geocode_dengue"
AND tweets."SE_twitter"=dengue."SE_dengue"
)
''' % general_param)
climate_join = ('''
LEFT JOIN (
SELECT
epi_week(data_dia) AS epiweek_climate,
AVG(%(var_climate)s) AS %(var_climate)s
FROM "Municipio"."Clima_wu"
WHERE "Estacao_wu_estacao_id" = '%(station_id)s'
GROUP BY epiweek_climate
ORDER BY epiweek_climate
) AS climate_wu
ON (dengue."SE_dengue"=climate_wu.epiweek_climate)
''' % general_param)
sql += climate_join + tweet_join
sql = ' SELECT {} FROM ({}) AS data'.format(','.join(k), sql)
df = pd.read_sql(sql, index_col='SE', con=db_engine)
if not df.empty:
dfs = []
# merge with a range date dataframe to keep empty week on the data
ts_date = pd.date_range(
df['init_date_week'].min(),
df['init_date_week'].max(),
freq='7D',
)
df_date = pd.DataFrame({'init_date_week': ts_date})
for geocode in df.geocode.unique():
df_ = df[df.geocode == geocode].sort_values('init_date_week')
df_date_ = df_date.set_index(
df_.init_date_week.map(
lambda x: int(episem(str(x)[:10], sep=''))),
drop=True,
)
df_.index.name = "SE"
df_date_.index.name = None
df_['init_date_week'] = pd.to_datetime(df_['init_date_week'],
errors='coerce')
# TODO: these parameters is not valid in the future releases.
dfs.append(
pd.merge(
df_,
df_date_,
how='outer',
left_index=True,
right_index=True,
))
df = pd.concat(dfs)
df.sort_index(ascending=True, inplace=True)
for d in cls.diseases:
k = 'p_rt1_{}'.format(d)
df[k] = (df[k] * 100).fillna(0)
k = 'casos_est_{}'.format(d)
df[k] = df[k].fillna(0).round(0)
k = 'p_inc100k_{}'.format(d)
df[k] = df[k].fillna(0).round(0)
df.rename(
columns={
'p_inc100k_{}'.format(d): 'incidência {}'.format(d),
'casos_{}'.format(d): 'casos notif. {}'.format(d),
'casos_est_{}'.format(d): 'casos est. {}'.format(d),
'p_rt1_{}'.format(d): 'pr(incid. subir) {}'.format(d),
},
inplace=True,
)
df.n_tweets = df.n_tweets.fillna(0).round(0)
return df.rename(columns={
'umid_max': 'umid.max',
'temp_min': 'temp.min',
'n_tweets': 'tweets',
})