class GetDataEcon(d6tflow.tasks.TaskPqPandas):
date_start = d6tflow.DateParameter() # define parameters
date_end = d6tflow.DateParameter() # define parameters
def run(self):
df_gdp = pddr.DataReader('CPGDPAI', 'fred', self.date_start, self.date_end)
self.save(df_gdp)
class TaskGetData(
d6tflow.tasks.TaskPqPandas
): # save dataframe as parquet, see https://d6tflow.readthedocs.io/en/latest/targets.html
dt_start = d6tflow.DateParameter(
default=cfg.dt_start
) # workflow parameters. See https://d6tflow.readthedocs.io/en/latest/advparam.html
dt_end = d6tflow.DateParameter(default=cfg.dt_end)
def run(self):
iris = sklearn.datasets.load_iris()
df_train = pd.DataFrame(
iris.data, columns=['feature{}'.format(i) for i in range(4)])
df_train['y'] = iris.target
# optional: df_train[df_train['date']>=self.dt_start]
self.save(df_train) # quickly save dataframe
class Task_getPlot(TaskPickle):
runDate = d6tflow.DateParameter()
plottype = d6tflow.Parameter()
def run(self):
tickers, Data = self.input()
tickers, Data = tickers.load().set_index('Ticker symbol'), Data.load()
status = Data['status']
if status == 'failure':
self.save({'status': status, 'plot': None})
return None
company_timeseries = pd.read_json(Data['dataframe'])
performance_data = company_timeseries.set_index('Date')
# status, company_timeseries = Data['status'], Data['company_timeseries']
# performance_data = pd.read_sql(
# company_timeseries.statement,
# company_timeseries.session.bind).set_index('Date')
# performance_data = self.input().load().set_index('Date')
myplot = getPlot(
plottype=self.plottype,
performance_data=performance_data,
status=status,
ticker=tickers.at[self.stockticker, 'Company'],
)
self.save(myplot)
class Task_getDates(TaskJson):
runDate = d6tflow.DateParameter()
type = d6tflow.Parameter()
def run(self):
input = self.input().load()
status, timeseries = input['status'], input['timeseries']
if status == 'failure':
self.save({'status': status, 'dates': None})
return None
timeseries = pd.read_json(timeseries)
options_date_map = sorted(timeseries['date'].unique(), reverse=True)
options_date_map = pd.DataFrame({'date':options_date_map})
options_date_map['date'] = options_date_map['date'].apply(lambda x: x.strftime('%Y-%m-%d'))
out = {
'timeseries': options_date_map.to_json(),
'status': 'success'
}
self.save(out)
class Task_getTickers(TaskCSVPandas):
runDate = d6tflow.DateParameter()
def run(self):
tickers = getTickers()
self.save(tickers)
class Task_getData(TaskJson):
runDate = d6tflow.DateParameter()
type = d6tflow.Parameter()
def run(self):
timeseries = getTimeseries_T(self.type)
self.save(timeseries)
class Task_getData(TaskJson):
runDate = d6tflow.DateParameter()
stockticker = d6tflow.Parameter()
def run(self):
# tickers = self.input().load()
#assert self.stockticker in tickers.keys(), 'Symbol should be in ticker list'
out = getData(stockticker=self.stockticker)
self.save(out)
class Task_getCountries(TaskJson):
runDate = d6tflow.DateParameter()
def run(self):
out = getCountries()
self.save(out)