mymodel.py

import mmlibrary as mm
import sys
import pandas as pd
import pickle
import datetime
import json
import zipfile
from collections import OrderedDict

# mm.parseArguments(sys.argv) #point to address - remove this on mmlibrary/cloud

newdata_humidity = pd.DataFrame(mm.getArgument("Humidity"))
newdata_air_pressure = pd.DataFrame(mm.getArgument("Air_Pressure"))


lastdate_str = newdata_humidity.iloc[-1,0]
lastdate_date = datetime.datetime.strptime(lastdate_str,"%Y-%m-%d %H:%M:%S")

newdate = []

for i in range(12):
    nextdate = lastdate_date + datetime.timedelta(minutes=5)
    newdate.append(nextdate)
    lastdate_date = nextdate

# Separate the independent variables
humidity = newdata_humidity[['Humidity']]
air_pressure = newdata_air_pressure[['Air_Pressure']]

# Load the models
modelBinary = mm.getModel()
# Extract and read the binaries files
with zipfile.ZipFile(modelBinary) as zipper:
    with zipper.open('MM Model Binary Temperature.pckl', 'r') as LnRg_File:
        LnRg_Model = pickle.load(LnRg_File)
        LnRg_File.close()
    with zipper.open('MM Model Binary Humidity.pckl', 'r') as LnRg_File_Humidity:
        LnRg_Model_Humidity = pickle.load(LnRg_File_Humidity)
        LnRg_File_Humidity.close()
    with zipper.open('MM Model Binary Air Pressure.pckl', 'r') as LnRg_File_Air_Pressure:
        LnRg_Model_Air_Pressure = pickle.load(LnRg_File_Air_Pressure)
        LnRg_File_Air_Pressure.close()

# Create the predicted values
predicted_values_Humidity = LnRg_Model_Humidity.predict(humidity)
prediction_series_Humidity = list(pd.Series(predicted_values_Humidity))

predicted_values_Air_Pressure = LnRg_Model_Air_Pressure.predict(air_pressure)
prediction_series_Air_Pressure = list(pd.Series(predicted_values_Air_Pressure))

newpred = [('Date_Time', newdate), ('Humidity', prediction_series_Humidity), ('Air_Pressure', prediction_series_Air_Pressure)]
newpred = pd.DataFrame.from_dict(OrderedDict(newpred))

newpred_x = newpred[['Humidity','Air_Pressure']]

predicted_values = LnRg_Model.predict(newpred_x)
prediction_series = list(pd.Series(predicted_values))
predictions = pd.DataFrame({'Date_Time': newdate, 'Predicted_Temperature': prediction_series})

stingdate = predictions['Date_Time'].dt.strftime("%Y-%m-%d %H:%M:%S")
predictions = pd.DataFrame({'Date_Time': stingdate, 'Predicted_Temperature': prediction_series})

score = predictions.to_dict(orient='records')

mm.returnScore(score)