app.py

from datetime import datetime

import pandas as pd
import requests
import streamlit as st

USER_ID = "rakesh1715"
REPO_ID = "gl-backend"

st.title("SuperKart Sales Forecasting")

# Batch Prediction
st.subheader("Online Prediction")

# Input fields for product data
Product_ID = st.text_input("Product ID", value="FD6114")
Product_Type_Category = st.selectbox("Product Type", ["Perishable", "Non Perishable"])
Product_Weight = st.number_input("Product Weight", min_value=0, value=650)
Product_Sugar_Content = st.selectbox("Sugar Content in Product", ["Regular", "No Sugar", "Low Sugar"])
Product_Allocated_Area = st.number_input("Allocated Area for Product", min_value=0, value=30)
Product_MRP = st.number_input("Product Price", min_value=0, value=12)
Store_Id = st.selectbox("Store ID", ["OUT001", "OUT002", "OUT003", "OUT004"])
Store_Establishment_Year = st.number_input("Store Establishment Year", min_value=1970, max_value=datetime.now().year,
                                           step=1, format="%d")
Store_Location_City_Type = st.selectbox("Store City Type", ["Tier 1", "Tier 2", "Tier 3"])
Store_Type = st.selectbox("Store Type", ["Departmental Store", "Food Mart", "Supermarket Type1", "Supermarket Type2"])
Store_Size = st.selectbox("Store Size", ["High", "Medium", "Small"])

product_data = {
        'Product_Weight': Product_Weight,
        'Product_Sugar_Content': Product_Sugar_Content,
        'Product_Allocated_Area': Product_Allocated_Area,
        'Product_MRP': Product_MRP,
        'Store_Size': Store_Size,
        'Store_Id': Store_Id,
        'Store_Location_City_Type': Store_Location_City_Type,
        'Store_Type': Store_Type,
        'Store_Age': datetime.now().year - Store_Establishment_Year,
        'Product_Type_Categories': Product_Type_Category,
}

if st.button("Predict", type='primary'):
    response = requests.post(f"https://{USER_ID}-{REPO_ID}.hf.space/v1/predict", json=product_data)
    if response.status_code == 200:
        result = response.json()
        sales_prediction = result["Sales"]  # Extract only the value
        st.write(
                f"Based on the information provided, the product with ID {Product_ID} is likely to have sales of {float(sales_prediction):,.2f}")
    else:
        st.error("Error in API request")

# Batch Prediction
st.subheader("Batch Prediction")

# Create a sample dataframe (customize columns to match your model’s inputs)
sample_data = {
        "Product_ID": ["FD6114", "FD6115"],
        "Product_Type_Categories": ["Perishable", "Non Perishable"],
        "Product_Weight": [0.7, 10],
        "Product_Sugar_Content": ["Regular", "No Sugar"],
        "Product_Allocated_Area": [0.19, 20],
        "Product_MRP": [100.23, 19.001],
        "Store_Id": ["OUT003", "OUT002"],
        "Store_Establishment_Year": [2010, 2020],
        "Store_Location_City_Type": ["Tier 1", "Tier 2"],
        "Store_Type": ["Departmental Store", "Food Mart"],
        "Store_Size": ["High", "Medium"],
}
sample_df = pd.DataFrame(sample_data)

# Convert to CSV and create a download button
st.download_button(
        label="📥 Download Sample CSV",
        data=sample_df.to_csv(index=False).encode("utf-8"),
        file_name="sample_input.csv",
        mime="text/csv"
)

file = st.file_uploader("Upload CSV file", type=["csv"])
if file is not None:
    if st.button("Predict for Batch", type='primary'):
        response = requests.post(f"https://{USER_ID}-{REPO_ID}.hf.space/v1/bulk/predict", files={"file": file})
        if response.status_code == 200:
            result = response.json()
            if result:
                df = pd.DataFrame(list(result.items()), columns=["Product Id", "Predicted Sales"])
                df["Predicted Sales"] = df["Predicted Sales"].round(2)
                st.header("Batch Prediction Results")
                st.dataframe(df, use_container_width=True)
        else:
            st.error("Error in API request")