def get_coords(end_date):
#get_prediction_for_daterange in timeser.py -> creates foreasts
data = get_incoords(timeser.get_prediction_for_daterange(end_date))
#ret is the dataframe for the visualization graph
ret = [{'data':data, 'name':'forecast', 'color':'red'}]
#retrieves the input data, transforms to ret dataframe
actual = get_incoords(timeser.get_actual())
#appends forecat data to input data
ret.append({'data':actual,'name':'actual','color':'blue'})
return ret
def predict():
#Redundant: start_data=first proceeding period of end of dataset
#start_date = request.args.get('start_date', '')
end_date = request.args.get('end_date', '2012-05-05 23:00:00+00:00')
if end_date=='null':
end_date = '2012-05-05 23:00:00+00:00'
#Depending on output type perform the appropriate prediction action
#timeser.get_prediction_for_daterange builds forecast
output_type = request.args.get('type', 'json')
if output_type=='coords':
#For visualization below methods are called
return dumps(get_coords(end_date))
elif output_type=='json':
predicted_demand = timeser.get_prediction_for_daterange(end_date, asutc=True)
return jsonify(predicted_demand.to_dict())
else:
predicted_demand = timeser.get_prediction_for_daterange(end_date, asutc=True)
odir = './generated_csvs/'
tfile = 'temp'+str(datetime.now())+'.csv'
predicted_demand.to_csv(odir + tfile)
return send_from_directory(odir, tfile)
