Notes

main_pp.py

# author : makcin

# Importing Libraries
import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.ensemble import RandomForestRegressor
import copy
from sklearn.metrics import mean_absolute_error
import openpyxl
from datetime import date, timedelta
from sklearn.linear_model import HuberRegressor, LinearRegression, TheilSenRegressor, RANSACRegressor
from sklearn.linear_model import Ridge, Lasso
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.metrics import mean_squared_error, r2_score
from sklearn.linear_model import ElasticNetCV
from sklearn.linear_model import ElasticNet
from sklearn.utils.validation import check_consistent_length, check_array
pd.set_option('display.max_rows', 6000)
from scripts import _check_reg_targets,mean_absolute_percentage_error,best_model_pipeline,taking_average
import warnings
warnings.filterwarnings("ignore")
import re
from openpyxl import load_workbook
from datetime import datetime


# ------- PP Models --------------------
# 1- PP HS Import Model
# 2- PP LS Import Model
# 3- PP MS Import Model

# Data Çekme kodu ile değişecek

# 1- PP MS Import Model

df_org = pd.read_excel("PP_tags_08082022.xlsx",index_col = 0,engine='openpyxl')


df_org.rename(columns = {'1701.FI605':'Pp_HS_Import',
'1701.FI604':'Pp_MS_Import',
'1701.FI603':'Pp_LS_Import',
'5801.11MBL01CT001_XQ60':'Ambient_Temperature',
'2101.FT109A.DACA.PV':'Amonyak_capacity_kg/h',
'1701.IIH301A_L':'PP_Ekskruder1',
'1701.IIH301B_L':'PP_Ekskruder2',
'1701.FI201':'Propilen_capacity',
'1701.FI226':'PP_Fg'}, inplace = True)

df_pp = copy.deepcopy(df_org)

# Confidence 0 olan dataları veri setinden düşürmek.
df_pp_manipulated = copy.deepcopy(df_pp)

confidence_cols = list([col for col in df_pp_manipulated.columns if 'Confidence' in col])

for i in confidence_cols:
df_pp_manipulated.drop(
df_pp_manipulated.loc[(df_pp_manipulated[i] == 0)].index,
inplace=True)

df_pp_manipulated.drop(confidence_cols,axis=1,inplace=True)

# Sıfırdan küçük değerlerin sıfıra eşitlenmesi

main_features = list(df_pp_manipulated.columns)

main_features.remove('Ambient_Temperature')


for i in main_features:
df_pp_manipulated.loc[df_pp_manipulated[i] < 0, i] = 0

df_pp_manipulated.loc[df_pp_manipulated['Ambient_Temperature']<=-10,'Ambient_Temperature'] = -10

# Kapasitelere göre veride bazı düzenlemeler yapılması

df_pp_manipulated.loc[df_pp_manipulated['Propilen_capacity']>=18.5,'Propilen_capacity'] = 18.5

df_pp_manipulated.loc[df_pp_manipulated['Propilen_capacity']<=5,'Propilen_capacity'] = 0

df_pp_manipulated.loc[df_pp_manipulated['Propilen_capacity']<=5,'Pp_HS_Import'] = 0

target_pta_hs_import_model = 'PTA_HS_Import'

# Ekskruder durumunun veri temizliği

df_pp_manipulated.loc[df_pp_manipulated['PP_Ekskruder1']<=20,'PP_Ekskruder1'] = 0
df_pp_manipulated.loc[df_pp_manipulated['PP_Ekskruder2']<=20,'PP_Ekskruder2'] = 0

# Ekskruder on/off feature

df_pp_manipulated['Pp_Ekskruder1_on_off'] = 0
df_pp_manipulated['Pp_Ekskruder2_on_off'] = 0

df_pp_manipulated.loc[df_pp_manipulated['PP_Ekskruder1']>20,'Pp_Ekskruder1_on_off'] = 1
df_pp_manipulated.loc[df_pp_manipulated['PP_Ekskruder2']>20,'Pp_Ekskruder2_on_off'] = 1

df_pp_manipulated['Number_of_ekskruder_on'] = df_pp_manipulated['Pp_Ekskruder1_on_off'] + df_pp_manipulated['Pp_Ekskruder2_on_off']


# Model datasının hazırlanması (kullanılmayacak feature'lar çıkarılmalı)

df_pp_ms_import_model_df = df_pp_manipulated[['Number_of_ekskruder_on','Pp_Ekskruder1_on_off','Propilen_capacity','Pp_MS_Import']]

df_pp_ms_import_model_df = df_pp_ms_import_model_df.dropna()

target_pp_ms_import_model = 'Pp_MS_Import'

# Model kodları

# Model Pipeline with GridSearchCV

result_pp_ms_import_prediction = best_model_pipeline(n=1000,df=df_pp_ms_import_model_df,target=target_pp_ms_import_model,name_of_model='PP_MS_Import_Model')

# Model sonuçları Excel'e yazılıyor.

n = (len(result_pp_ms_import_prediction) - 2)/2

model_pred_pp_ms_import_model = round(result_pp_ms_import_prediction[0], 2)

model_intercept_pp_ms_import_model = round(result_pp_ms_import_prediction[1], 2)

values_used_in_pp_ms_import_model = [round(num, 2) for num in result_pp_ms_import_prediction[int(-n):]]

model_coefs_of_pp_ms_import_model = [round(num, 2) for num in result_pp_ms_import_prediction[2:int(2+n)]]


wb = load_workbook('Draft_Final_Output.xlsx')

main_summary_sheet = wb['Main_Summary_Sheet']

main_summary_sheet['D13'] = values_used_in_pp_ms_import_model[2]

pp_sheet = wb['PP']

pp_sheet['B3'] = values_used_in_pp_ms_import_model[0]

pp_sheet['B4'] = values_used_in_pp_ms_import_model[1]

pp_sheet['B5'] = values_used_in_pp_ms_import_model[2]


pp_sheet['D3'] = model_coefs_of_pp_ms_import_model[0]

pp_sheet['D4'] = model_coefs_of_pp_ms_import_model[1]

pp_sheet['D5'] = model_coefs_of_pp_ms_import_model[2]

pp_sheet['B6'] = model_intercept_pp_ms_import_model

pp_sheet['L2'] = model_pred_pp_ms_import_model

pp_sheet['R1'] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

# 2- PP LS Import Model

df_pp_ls_import_model_df = df_pp_manipulated[['Propilen_capacity', 'PP_Ekskruder2', 'Pp_LS_Import']]

target_pp_ls_import_model = "Pp_LS_Import"

# Model Pipeline with GridSearchCV

result_pp_ls_import_prediction = best_model_pipeline(n=1000,df=df_pp_ls_import_model_df,target=target_pp_ls_import_model,name_of_model='PP_LS_Import_Model')

# Model sonuçları Excel'e yazılıyor.

n = (len(result_pp_ls_import_prediction) - 2)/2

model_pred_pp_ls_import_model = round(result_pp_ls_import_prediction[0], 2)

model_intercept_pp_ls_import_model = round(result_pp_ls_import_prediction[1], 2)

values_used_in_pp_ls_import_model = [round(num, 2) for num in result_pp_ls_import_prediction[int(-n):]]

model_coefs_of_pp_ls_import_model = [round(num, 2) for num in result_pp_ls_import_prediction[2:int(2+n)]]

# Sonuclar Excel'e yazılıyor.

pp_sheet['B8'] = values_used_in_pp_ls_import_model[0]

pp_sheet['B9'] = values_used_in_pp_ls_import_model[1]

pp_sheet['D8'] = model_coefs_of_pp_ls_import_model[0]

pp_sheet['D9'] = model_coefs_of_pp_ls_import_model[1]

pp_sheet['B10'] = model_intercept_pp_ls_import_model

pp_sheet['L3'] = model_pred_pp_ls_import_model

pp_sheet['R1'] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

# 2- PP HS Import Model

df_pp_hs_import_model_df = df_pp_manipulated[['Propilen_capacity', 'Number_of_ekskruder_on', 'Pp_HS_Import']]

target_pp_hs_import_model = "Pp_HS_Import"

# Model Pipeline with GridSearchCV

result_pp_hs_import_prediction = best_model_pipeline(n=1000,df=df_pp_hs_import_model_df,target=target_pp_hs_import_model,name_of_model='PP_HS_Import_Model')

# Model sonuçları Excel'e yazılıyor.

n = (len(result_pp_hs_import_prediction) - 2)/2

model_pred_pp_hs_import_model = round(result_pp_hs_import_prediction[0], 2)

model_intercept_pp_hs_import_model = round(result_pp_hs_import_prediction[1], 2)

values_used_in_pp_hs_import_model = [round(num, 2) for num in result_pp_hs_import_prediction[int(-n):]]

model_coefs_of_pp_hs_import_model = [round(num, 2) for num in result_pp_hs_import_prediction[2:int(2+n)]]

# Arom FG Prediction Model Sonuçları Excel'e yazılıyor.

pp_sheet['D12'] = model_coefs_of_pp_hs_import_model[0]

pp_sheet['D13'] = model_coefs_of_pp_hs_import_model[1]

pp_sheet['B14'] = model_intercept_pp_ls_import_model

pp_sheet['L4'] = model_pred_pp_hs_import_model

pp_sheet['R1'] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

wb.save(r'C:Usersmert.akcinDesktopDaily NG DemandNG_Demand_NewNg_Demand_Prod_CodeDraft_Final_Output.xlsx')
wb.close()