import streamlit as st
import requests
import pandas as pd

st.set_page_config(page_title="ExtraaLearn Predictor", layout="centered")
st.title("ExtraaLearn Lead Conversion Predictor")

# Input fields based on ExtraaLearn dataset features
age = st.number_input("Age", min_value=18, max_value=100, value=45)
current_occupation = st.selectbox("Current Occupation", ["Professional", "Unemployed", "Student"])
first_interaction = st.selectbox("First Interaction", ["Website", "Mobile App"])
profile_completed = st.selectbox("Profile Completed", ["Low", "Medium", "High"], index=2) # Default High
website_visits = st.number_input("Website Visits", min_value=0, max_value=50, value=3)
time_spent_on_website = st.number_input("Time Spent on Website (seconds)", min_value=0, value=500)
page_views_per_visit = st.number_input("Page Views Per Visit", min_value=0.0, value=3.0)
last_activity = st.selectbox("Last Activity", ["Email Activity", "Phone Activity", "Website Activity"])

# Marketing Channels
print_media_type1 = st.selectbox("Saw Ad in Newspaper?", ["Yes", "No"], index=1)
print_media_type2 = st.selectbox("Saw Ad in Magazine?", ["Yes", "No"], index=1)
digital_media = st.selectbox("Saw Ad on Digital Platforms?", ["Yes", "No"], index=1)
educational_channels = st.selectbox("Heard via Educational Channels?", ["Yes", "No"], index=1)
referral = st.selectbox("Heard via Referral?", ["Yes", "No"], index=1)

lead_data = {
    "age": age,
    "current_occupation": current_occupation,
    "first_interaction": first_interaction,
    "profile_completed": profile_completed,
    "website_visits": website_visits,
    "time_spent_on_website": time_spent_on_website,
    "page_views_per_visit": page_views_per_visit,
    "last_activity": last_activity,
    "print_media_type1": print_media_type1,
    "print_media_type2": print_media_type2,
    "digital_media": digital_media,
    "educational_channels": educational_channels,
    "referral": referral
}

if st.button("Predict Propensity", type='primary'):
    api_url = "https://sirisha335-extraalearn-api.hf.space/v1/predict"

    with st.spinner("Connecting to API (this may take 20s if the backend is waking up)..."):
        try:
            # Increased timeout to 20 seconds to allow for cold starts
            response = requests.post(api_url, json=lead_data, timeout=20)
            if response.status_code == 200:
                result = response.json()
                prediction = result.get("Conversion_Probability", 0)
                st.metric("Conversion Propensity", f"{prediction:.4f}")
                if prediction > 0.5:
                    st.success("This lead is highly likely to convert!")
                else:    
                    st.warning("This lead has a low probability of conversion.")
            else:
                st.error(f"API Error: {response.status_code} - {response.text}")
        except requests.exceptions.Timeout:
            st.error("The request timed out. The backend Space is likely still starting up or under heavy load. Please try again in a minute.")
        except Exception as e:
            st.error(f"Could not connect to the backend. Ensure the Backend Space is 'Running'. Error: {e}")