app.py

# -*- coding: utf-8 -*-
"""Gradio_final_version_upwork_.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1aoMHIwjlYzuHId1tfByJINN1Gazb7bzL
"""



from joblib import dump, load
import pickle
import gradio as gr
import pandas as pd
import json
from random import randint
filename="Pre_Trained_Model_Setting_v15.sav"
modlist_loaded = pickle.load(open(filename, 'rb'))

clf = modlist_loaded[0]
enc = modlist_loaded[1]
lis = modlist_loaded[2]
cs  = modlist_loaded[3]
df  = modlist_loaded[4]
df3 = modlist_loaded[5]
df5 = modlist_loaded[6]
vendor_small = modlist_loaded[7]
vendor_mid = modlist_loaded[8]
vendor_large = modlist_loaded[9]

def get_probability(Industry, Company_Size):
  var=(clf.predict_proba(enc.transform([[Industry,Company_Size]])))
  var=var.transpose()
  var=var.tolist()
  name=clf.classes_.tolist()
  dataset = pd.DataFrame({'Prob': var, 'Name':name},columns=['Prob','Name'])
  df=dataset.sort_values('Prob',ascending=False)
  return df

def start(Industry, Company_Size):

  
    if (Industry == "" or Industry== None) and (Company_Size !="" and  Company_Size!=None):
        return "Please Select Industry Drop Down", Company_Size
   
    if (Company_Size== "" or Company_Size== None)  and (Industry != "" and  Industry!=None):
            return Industry,"Please Select Company Size Drop Down"
   
    if (Industry == ""  or Industry== None) and (Company_Size =="" or Industry== None): 
            return "Please Select Industry Drop Down", "Please Select Company Size Drop Down"
    else:


        ac="ActiveCampaign"
        df_=get_probability(Industry, Company_Size)
        spec_titles = ["recommendedVendor", "recommendationScore", "vendorPros", "vendorCons","userSatisfaction"]
        cars = {}
        title="Recommendation for " + str(Industry) + " Industry having company size " + str(Company_Size)
        vehical_data = {title: [cars]}
        lst_prob=[]
        sum_first_prob=df_["Prob"].iloc[0][0]

        for x in range(5,len(df_)):
            sum_first_prob+=df_["Prob"].iloc[x][0]
        
        lst_prob.append(sum_first_prob)

        for x in range(1,5):
            prob=df_["Prob"].iloc[x][0]
            lst_prob.append(prob)
            
    
        if df_["Name"].iloc[2]==ac:
        
            df_["Name"].iloc[2],df_["Name"].iloc[0]=df_["Name"].iloc[0],df_["Name"].iloc[2]
            dig=float(lst_prob[0])
            dig=dig+0.01
            lst_prob[0]=float(dig)
        
            
        elif df_["Name"].iloc[3]==ac:
        
            df_["Name"].iloc[3],df_["Name"].iloc[0]=df_["Name"].iloc[0],df_["Name"].iloc[3]
            dig=float(lst_prob[0])
            dig=dig+0.01
            lst_prob[0]=float(dig)
            
        elif df_["Name"].iloc[4]==ac:
        
            df_["Name"].iloc[4],df_["Name"].iloc[0]=df_["Name"].iloc[0],df_["Name"].iloc[4]
            dig=float(lst_prob[0])
            dig=dig+0.01
            lst_prob[0]=float(dig)
            
        myflag=True
        for x in range(0,5):
            if df_["Name"].iloc[x] == ac:
                myflag=False
        if myflag==True:
            df_["Name"].iloc[2]=ac
            if int(lst_prob[1])==int(lst_prob[2]):
                
            
                dig=float(lst_prob[1])
                dig=dig+0.01
                lst_prob[1]=float(dig)
                df_["Name"].iloc[1],df_["Name"].iloc[2]=df_["Name"].iloc[2],df_["Name"].iloc[1]
            
            else:
                dig=float(lst_prob[2])
                dig=dig+0.01
                lst_prob[2]=float(dig)
            
        
        
        for n in range(0,5):
            name = "Recommendation Number " + str(n+1)
            vendor_name=df_["Name"].iloc[n]
            vendor_prob=lst_prob[n]
           # vendor_prob=round(float(vendor_prob)*100)
            if n==0:
                vendor_prob=round(float(vendor_prob)*100)+30
            else:
                 vendor_prob=round(float(vendor_prob)*100)+45
                 if vendor_prob>round(float(lst_prob[0])*100)+30:
                     vendor_prob=vendor_prob-10
                 
                 
            
            vendor_pros=(df5.loc[df5['Vendor Name'] == str(vendor_name)]["Pros"]).item()
            vendor_cons=(df5.loc[df5['Vendor Name'] == str(vendor_name)]["Cons"]).item()
            user_satisfaction=""



            

            if Company_Size=="Self Employed" or Company_Size=="2-10 Employees" or Company_Size=="11-50 Employees":
                user_satisfaction=str(vendor_small[str(vendor_name)]) +"% of " + str(vendor_name) + "'s small business users are satisfied with it"
                
            elif Company_Size=="51-200 Employees" or Company_Size=="201-500 Employees" or Company_Size=="501-1,000 Employees":
                user_satisfaction=str(vendor_mid[vendor_name]) +"% of " + vendor_name + "'s medium business users are satisfied with it"
    
            elif Company_Size=="1,001-5,000 Employees" or Company_Size=="5,001-10,000 Employees" or Company_Size=="10,000+ Employees":
                user_satisfaction=str(vendor_large[vendor_name]) +"% of " + vendor_name + "'s large business users are satisfied with it"
      









            spec_details = [str(vendor_name), str(vendor_prob)+"%", str(vendor_pros), str(vendor_cons),str(user_satisfaction)]
            car_data = dict(zip(spec_titles, spec_details))
            cars[name] = car_data
        js=json.dumps(vehical_data)
        return vehical_data

face = gr.Interface(fn=start, inputs=[gr.Dropdown(lis), gr.Dropdown(cs),], outputs=["json"],allow_flagging="never")
face.launch()