app.py

import gradio as gr
import pandas as pd
import pickle


def load_model(model_name):
    filename=model_name
    loaded_model = pickle.load(open(filename, 'rb'))
    
    return loaded_model

rf_clf=load_model('Random_forest_perm_clf.sav')



from zipfile import ZipFile
import csv
import gradio as gr


def Perm_pred(file_obj):
    dataframe=pd.read_csv(file_obj.name, delimiter=',')
    new_df=dataframe[['pdp_avg','RETURN_RATE','conversion','Cat_1','Cat_2','Cat_3']]
    sku=dataframe['SKU']
    pred_score = rf_clf.predict(new_df)
    #pred=pd.DataFrame(pred_score,columns=['Prediction'])
    pred_list=pred_score.tolist()
    y=[]
    for i in range(len(pred_list)):
    
        if pred_score[i]==2:
           y.append('Never_Perm')
        elif pred_score[i]==1:
           y.append('Middle_Perm')
        else:
           y.append('Perm_A')
    
    pred=pd.DataFrame(y,columns=['Prediction'])
    pred['SKU']=sku
    df=pred[['SKU','Prediction']]


    #with open('Prediction.csv','w', encoding='UTF8') as f:
    #    writer=csv.writer(f)

    #    for i in range(len(y)):
    #        writer.writerow(y[i])

    #return 'Prerdiction.csv'
    return df

    
   

    
    
            
    


demo = gr.Interface(Perm_pred, inputs=[gr.inputs.File(label='Enter CSV File')], outputs= [gr.outputs.Dataframe(label='Predicted Label')])

if __name__ == "__main__":
    demo.launch(share=True)