import streamlit as st
import pandas as pd
import requests
import time


# Streamlit UI for Customer Churn Prediction
st.title("Telecom Customer Churn Prediction App")
st.write("This tool predicts customer churn risk based on their details. Enter the required information below.")

# Collect user input based on dataset columns
CustomerID = st.number_input("Customer ID", min_value=10000000, max_value=99999999)
SeniorCitizen = st.selectbox("Senior citizen", ["Yes", "No"])
Partner = st.selectbox("Does the customer have a partner?", ["Yes", "No"])
Dependents = st.selectbox("Does the customer have dependents?", ["Yes", "No"])
PhoneService = st.selectbox("Does the customer have phone service?", ["Yes", "No"])
InternetService = st.selectbox("Type of Internet Service", ["DSL", "Fiber optic", "No"])
Contract = st.selectbox("Type of Contract", ["Month-to-month", "One year", "Two year"])
PaymentMethod = st.selectbox("Payment Method", ["Electronic check", "Mailed check", "Bank transfer", "Credit card"])
tenure = st.number_input("Tenure (Months with the company)", min_value=0, value=12)
MonthlyCharges = st.number_input("Monthly Charges", min_value=0.0, value=50.0)
TotalCharges = st.number_input("Total Charges", min_value=0.0, value=600.0)

# Convert user input into a DataFrame
customer_data = pd.DataFrame([{
    'SeniorCitizen': 1 if SeniorCitizen == "Yes" else 0,
    'Partner':Partner,
    'Dependents': Dependents,
    'tenure': tenure,
    'PhoneService': PhoneService,
    'InternetService': InternetService,
    'Contract': Contract,
    'PaymentMethod': PaymentMethod,
    'MonthlyCharges': MonthlyCharges,
    'TotalCharges': TotalCharges
}])


# Single prediction section
if st.button("Predict", type='primary'):
    response = requests.post("https://MainiSandeep1987-BackEndFlaskAPITelecomChurnPrediction.hf.space/v1/customer", json=customer_data)    # enter user name and space name before running the cell
    if response.status_code == 200:
        result = response.json()
        churn_prediction = result["Prediction"]  # Extract only the value
        st.write(f"Based on the information provided, the customer with ID {CustomerID} is likely to {churn_prediction}.")
    else:
        st.error("Error in API request")


# Single prediction section
# if st.button("Predict"):
#     try:
#         print(customer_data)
#         response = requests.post(
#             "https://MainiSandeep1987-BackEndFlaskAPITelecomChurnPrediction.hf.space/v1/customer",
#             json=customer_data.to_dict(orient='records')[0]
#         )  # Send data to Flask API
#         response.raise_for_status()  # Raise an error if the request fails
#         result = response.json()
#         churn_prediction = result["Prediction"]  # Extract only the value
#         st.write(f"Based on the information provided, the customer with ID {CustomerID} is likely to {churn_prediction}.")
#     except requests.exceptions.RequestException as e:
#         st.error(f"Error making prediction: {e}")
        


# # Function for handling retries with exponential backoff
# def make_request_with_backoff(url, payload, max_retries=5):
#     retry_delay = 2  # Initial retry delay (in seconds)

#     for attempt in range(max_retries):
#         try:
#             response = requests.post(url, json=payload)
#             response.raise_for_status()  # Ensure successful response

#             return response.json()  # Return data if request succeeds

#         except requests.exceptions.RequestException as e:
#             if response.status_code == 429:  # Too many requests
#                 st.warning(f"Rate limit exceeded. Retrying in {retry_delay} seconds...")
#                 time.sleep(retry_delay)  # Wait before retrying
#                 retry_delay *= 2  # Exponential backoff (2s, 4s, 8s, etc.)
#             else:
#                 st.error(f"Request failed: {e}")
#                 return None

#     st.error("Max retries reached. Try again later.")
#     return None

# # Single prediction section with retry mechanism
# if st.button("Predict"):
#     try:
#         print(customer_data)  # Display input data for debugging
        
#         result = make_request_with_backoff(
#             "https://MainiSandeep1987-BackEndFlaskAPITelecomChurnPrediction.hf.space/v1/customer",
#             customer_data.to_dict(orient='records')[0]
#         )

#         if result:
#             churn_prediction = result.get("Prediction", "Unknown")  # Safe key lookup
#             customer_id = customer_data.get("CustomerID", "Unknown")  # Prevent undefined errors
#             st.write(f"Based on the information provided, the customer with ID {customer_id} is likely to {churn_prediction}.")
        
#     except KeyError as e:
#         st.error(f"Unexpected response format: Missing key {e}")
#     except Exception as e:
#         st.error(f"Unexpected error: {e}")



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

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("https://MainiSandeep1987-BackEndFlaskAPITelecomChurnPrediction.hf.space/v1/customerbatch", files={"file": file})    # enter user name and space name before running the cell
        if response.status_code == 200:
            result = response.json()
            st.header("Batch Prediction Results")
            st.write(result)
        else:
            st.error("Error in API request")