import streamlit as st
import pandas as pd
import numpy as np
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder

# Streamlit UI
st.set_page_config(page_title="BigMart Sales Predictor", page_icon="🛒", layout="centered")
st.title("🛒 BigMart Sales Prediction using Real World Dataset")
st.markdown("Fill in the product details to get a sales prediction.")

# Load and preprocess dataset
@st.cache_data
def load_data():
    data = pd.read_csv("Train.csv")  # 👈 Make sure Train.csv is in the same directory
    # Handle missing values
    data.fillna(data.mean(numeric_only=True), inplace=True)
    data.fillna("Unknown", inplace=True)
    
    # Encode categorical columns
    label_enc = LabelEncoder()
    for col in ['Item_Fat_Content', 'Item_Type', 'Outlet_Identifier', 'Outlet_Size', 'Outlet_Location_Type', 'Outlet_Type']:
        data[col] = label_enc.fit_transform(data[col])
    return data

df = load_data()

# Select features and target
features = ['Item_Weight', 'Item_Visibility', 'Item_MRP']
target = 'Item_Outlet_Sales'

X = df[features]
y = df[target]

# Train model
model = LinearRegression()
model.fit(X, y)

# Input UI
product_name = st.text_input("📦 Product Name")
item_weight = st.number_input("⚖️ Item Weight (kg)", min_value=0.0, step=0.1)
item_visibility = st.slider("👀 Item Visibility", 0.0, 1.0, 0.05)
item_mrp = st.number_input("💰 Item MRP", min_value=0.0, step=1.0)

# Prediction
if st.button("Predict Sales"):
    if not product_name:
        st.warning("Please enter a product name.")
    else:
        user_input = np.array([[item_weight, item_visibility, item_mrp]])
        predicted_sales = model.predict(user_input)[0]
        st.success(f"📈 Predicted Sales for '{product_name}': ₹{predicted_sales:,.2f}")
        
        # Optional: Download Prediction
        result_df = pd.DataFrame({
            "Product Name": [product_name],
            "Item Weight": [item_weight],
            "Item Visibility": [item_visibility],
            "Item MRP": [item_mrp],
            "Predicted Sales": [predicted_sales]
        })

        st.download_button("📥 Download Result as CSV", result_df.to_csv(index=False), file_name="prediction.csv", mime="text/csv")

# Sidebar Info
st.sidebar.title("📌 About")
st.sidebar.markdown("""
This app uses a **real BigMart dataset** from Kaggle and a **Linear Regression model** to predict sales.
You can customize features or switch to advanced ML models later!
""")