import streamlit as st
import numpy as np
import pandas as pd
import joblib
import matplotlib.pyplot as plt
import plotly.express as px

st.title("Customer Segmentation")

kmeans = joblib.load("kmeans.pkl")
scaler = joblib.load("scaler.pkl")
rfm = pd.read_csv("transformation.csv")

cluster_label = {0: 'Loyal Customers', 1: 'At Risk', 2: 'Champions', 3: 'New Customers'}

def customer_segmentation(num1,num2,num3):
  print("Customer Segmentation")
  data_recency = np.log1p(num1)
  data_frequency = np.log1p(num2)
  data_monetary = np.log1p(num3)
  data = pd.DataFrame({'Recency': [data_recency], 'Frequency': [data_frequency], 'Monetary': [data_monetary]})
  X_data = scaler.transform(data)
  pred = kmeans.predict(X_data)
  return cluster_label[pred[0]]

col1,col2,col3 = st.columns(3)
num1 = col1.number_input("Enter Recency",min_value=1,max_value=400,step=1)
num2 = col2.number_input("Enter Frequency",min_value=1,max_value=6000,step=1)
num3 = col3.number_input("Enter Monetary",min_value=1,step=10)

value = ""
if st.button(label="Predict"):
    value = customer_segmentation(num1,num2,num3)

st.markdown(f"<span style='font-size:20px; font-weight:bold; font-style:italic'>{value}</span>",unsafe_allow_html=True)


custom_colors = {
    'Loyal Customers': '#99ff99',
    'Champions': '#66b3ff',
    'At Risk': '#ff9999',
    'New Customers': '#ffcc99'
}

figx = px.scatter_3d(
    rfm,
    x='Recency',
    y='Frequency',
    z='Monetary',
    color='Cluster Labels',
    color_discrete_map=custom_colors,
    labels={'Recency': 'Recency', 'Frequency': 'Frequency', 'Monetary': 'Monetary'},
    title='Customer Segmentation Visualization'
)
st.plotly_chart(figx)



customers = rfm.shape[0]
labels = ['Loyal Customers','At Risk','Champions','New Customers']
sizes = (rfm["Cluster"].value_counts()/customers)*100
colors = ['#99ff99', '#ff9999', '#66b3ff', '#ffcc99']

fig,ax = plt.subplots(figsize=(8,6))

ax.pie(
    sizes, labels=labels, colors=colors, autopct='%1.1f%%',
    startangle=120, wedgeprops={'edgecolor': 'black'}
)

ax.set_title('Customer Segmentation', fontsize=14)
ax.legend([0,1,2,3],title='Clusters',loc='best',)
st.pyplot(fig)