app.py

import streamlit as st
import pandas as pd
import numpy as np
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.ensemble import RandomForestClassifier
import joblib

import warnings
warnings.filterwarnings('ignore')


st.set_page_config(page_title='Product Type Predictor')
st.title('Detect Product Type')
st.subheader('Upload your CSV file')
uploaded_file = st.file_uploader('Choose a CSV file', type='csv')
if uploaded_file is not None:
    st.markdown('---')
    # Loading the data
    @st.cache_data
    def load_excel(file1):
            df = pd.read_csv(file1)
            return df
        
    data = load_excel(uploaded_file)
    st.subheader('Data Preview')
    st.dataframe(data.head(20))
    # Feature selection
    features = ['a_ApplicableMarkets', 'Manufacturing Plant','Number of Unique Finished Packs in BOM',
       'Total Number of Finished Packs in BOM', 'GMN', 'Product_Description',
       'EA_GTIN', 'CV_GTIN', 'Product_Hierarchy_Code',
       'Product_Hierarchy_Units_Per_Pack_L8', 'myPSR_Pack_Variant',
       'Stibo_Pack_variant']
    df = data[features]

    df['Manufacturing Plant'] = df['Manufacturing Plant'].replace({'Commerical Plant':'Commercial Plant'})
    df['Stibo_Pack_variant'] = df['Stibo_Pack_variant'].replace({'Migration Open Stock':'Migration OpenStock'})
    df = df.replace(np.nan, 0, regex=True)
    df['EA_GTIN'] = df['EA_GTIN'].astype(str)
    df['CV_GTIN'] = df['CV_GTIN'].astype(str)

    def GTIN_validity(x):
        gtin=str(x)
        if x=="0.0":
            return False
        if x:
            gtin=gtin[:-2]
            original_digits = [int(x) for x in gtin]
            digits_without_check_digit = original_digits[:-1]

            digits_without_check_digit.reverse()
            multiplied_digits = [x*3 if not i%2 else x
                                for i,x
                                in enumerate(digits_without_check_digit)]
            digits_sum = sum(multiplied_digits)

            if (digits_sum % 10):
                uprounded_sum = digits_sum + (10 - digits_sum % 10)
            else:   
                uprounded_sum = digits_sum
            expected_check_digit = uprounded_sum - digits_sum
            return (original_digits[-1] == expected_check_digit)

    df['EA_GTIN_valid']=df.apply(lambda x: GTIN_validity(x['EA_GTIN']),axis=1)
    df['CV_GTIN_valid']=df.apply(lambda x: GTIN_validity(x['CV_GTIN']),axis=1)
    text_cols = ['a_ApplicableMarkets', 'Manufacturing Plant', 'Product_Hierarchy_Units_Per_Pack_L8', 'myPSR_Pack_Variant', 'Stibo_Pack_variant']
    df = pd.get_dummies(data=df, columns=text_cols)

    v = CountVectorizer()
    text_vectors = v.fit_transform(df['Product_Description'])
    text_vectors_df = pd.DataFrame(text_vectors.toarray(), columns=v.get_feature_names_out())
    df_ext = pd.concat([df, text_vectors_df],axis=1)
    df = df_ext.drop(['GMN','Product_Description','EA_GTIN','CV_GTIN'],axis=1)

    loaded_model = joblib.load(open('rfc_model_grid.sav', 'rb'))
    result = loaded_model.predict(df)
    data['Product_Type_Predicted']=result

    out=data.to_csv().encode('utf-8')
    st.download_button(label='DOWNLOAD RESULT',data=out, file_name='Product_Type_Output.csv',mime='csv')