import streamlit as st
import requests
import pandas as pd

# Set up Streamlit page
st.set_page_config(page_title="🌍 Geospatial Querying System", page_icon="📍", layout="wide")

# Custom Styling
st.markdown("""
    <style>
        .stButton > button {
            background-color: #ff5a5f;
            color: white;
            font-size: 16px;
            border-radius: 10px;
        }
    </style>
""", unsafe_allow_html=True)

# Title & Instructions
st.title("📍 Geospatial Querying System")
st.write("🌍 **Enter any query like:**")
st.markdown("""
- `"Find hospitals in Delhi"`
- `"Show malls near Bangalore"`
- `"Where are the police stations in Mumbai?"`
""")

# User input
query = st.text_input("🔍 Enter your query (e.g., 'Find restaurants in Mumbai'):")

# Buttons
col1, col2 = st.columns(2)
with col1:
    find_location = st.button("📍 Find Location & Category")
with col2:
    find_nearby = st.button("🔎 Find Nearby Places")

# Backend API URLs
BASE_URL = "https://geospatial-api.onrender.com"

# Global variables to store extracted data
if "location" not in st.session_state:
    st.session_state["location"] = None
if "category" not in st.session_state:
    st.session_state["category"] = None

# Step 1: Extract location & category from query
if find_location:
    if query:
        with st.spinner("🔍 Extracting location & category..."):
            response = requests.post(f"{BASE_URL}/query", json={"text": query})
            data = response.json()
        
        if "error" in data:
            st.error("❌ Could not extract location. Try again!")
        else:
            # Store extracted data in session state
            st.session_state["location"] = data["place"]
            st.session_state["category"] = data["category"]

            # Display extracted information
            st.success(f"✅ **Extracted Location:** `{data['place']}`")
            st.success(f"📌 **Extracted Category:** `{data['category']}`")
            
            st.write(f"🌍 **Latitude:** `{data['latitude']}`")
            st.write(f"🌍 **Longitude:** `{data['longitude']}`")

            # Google Maps Link
            maps_link = f"https://www.google.com/maps/search/?api=1&query={data['latitude']},{data['longitude']}"
            st.markdown(f"[📍 View on Google Maps]({maps_link})", unsafe_allow_html=True)

# Step 2: Find Nearby Places
if find_nearby:
    if st.session_state["location"] and st.session_state["category"]:
        with st.spinner("🔎 Searching for nearby places..."):
            response = requests.post(f"{BASE_URL}/nearby", json={"place": st.session_state["location"], "category": st.session_state["category"]})
            data = response.json()

        if "nearby_places" in data and data["nearby_places"]:
            st.success(f"✅ **Nearby {st.session_state['category']} found in {st.session_state['location']}:**")
            
            # Show Places in a Table
            places_df = pd.DataFrame(data["nearby_places"])
            places_df = places_df.rename(columns={"name": "Place Name", "lat": "Latitude", "lon": "Longitude"})
            st.dataframe(places_df)

            # Show Google Maps Links
            for place in data["nearby_places"]:
                maps_link = f"https://www.google.com/maps/search/?api=1&query={place['lat']},{place['lon']}"
                st.markdown(f"📍 **{place['name']}** [🔗 View on Maps]({maps_link})", unsafe_allow_html=True)

            # Show on Interactive Map
            map_data = pd.DataFrame({
                "lat": [p["lat"] for p in data["nearby_places"]],
                "lon": [p["lon"] for p in data["nearby_places"]]
            })
            st.map(map_data)
        else:
            st.error("❌ No places found in the selected category.")
    else:
        st.error("❌ Please extract location & category first by clicking 'Find Location & Category'.")