def _get_master(self):
""" Get the global dataset that will be used for all calculations inside
this instance of the class
:return: None
"""
start_time = time.time()
start_date, end_date = self._get_date(None,
self.start_date), self._get_date(
None, self.end_date)
query = Utilities.open_sql(self.dir_path, 'user_sequence.sql')
# Temp fix for testing because my credentials are not working
cache = False
if cache:
df = pd.read_csv(
os.path.join(self.dir_path, 'data/manually_loaded_data',
'United Kingdom-Customer Service.csv'))
print("Converting object types...")
df = df.astype({'time_event': 'datetime64[ns]'})
df['time_event'] = df.time_event.apply(
lambda x: pytz.utc.localize(x))
df['date'] = df.time_event.dt.date
self.master = df
else:
self.master = self.query_db(query, self._formatted_flow_name(),
start_date.strftime('%Y-%m-%d'),
end_date.strftime('%Y-%m-%d'))
print(
f"Master Dataset Gathered in {round(time.time() - start_time, 0)} seconds"
)
def get_available_flows(self):
query = Utilities.open_sql(self.dir_path, 'flownames.sql')
df = self.client.query(query).to_dataframe()
return df['FlowName'].to_list()