import pandas 
import pickle
import gradio
import sklearn

    
 
file = open('RFC_model.pk', 'rb')
rf_clf=pickle.load(file)
file.close()


def Perm_pred(file_obj):
    dataframe=pandas.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=pandas.DataFrame(y,columns=['Prediction'])
    pred['SKU']=sku
    df=pred[['SKU','Prediction']]


    
    return df

    
   

    
    
            
    


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