# -*- coding: utf-8 -*-
"""
Created on Thu Jun  8 03:39:02 2023

@author: mritchey
"""
# streamlit run "C:\Users\mritchey\.spyder-py3\Python Scripts\streamlit projects\hail\hail all.py"

import pandas as pd
import numpy as np
import streamlit as st
from geopy.extra.rate_limiter import RateLimiter
from geopy.geocoders import Nominatim
import folium
from streamlit_folium import st_folium
from vincenty import vincenty
import duckdb

st.set_page_config(layout="wide")


@st.cache_data
def convert_df(df):
    return df.to_csv(index=0).encode('utf-8')


def duck_sql(sql_code):
    con = duckdb.connect()
    con.execute("PRAGMA threads=2")
    con.execute("PRAGMA enable_object_cache")
    return con.execute(sql_code).df()


def get_data(lat, lon, date_str):
    code = f"""
        select "#ZTIME" as "Date_utc", LON, LAT, MAXSIZE 
          from 
          'data/*.parquet'
                 where LAT<={lat}+1 and LAT>={lat}-1
                 and LON<={lon}+1 and LON>={lon}-1
                 and "#ZTIME"<={date_str}
    
    """
    return duck_sql(code)


def map_location(address, lat, lon):

    m = folium.Map(location=[lat, lon],

                   zoom_start=6,
                   height=400)
    folium.Marker(
        location=[lat, lon],
        tooltip=f'Address: {address}',
    ).add_to(m)

    return m


def distance(x):
    left_coords = (x[0], x[1])
    right_coords = (x[2], x[3])
    return vincenty(left_coords, right_coords, miles=True)


def geocode(address):
    try:
        address2 = address.replace(' ', '+').replace(',', '%2C')
        df = pd.read_json(
            f'https://geocoding.geo.census.gov/geocoder/locations/onelineaddress?address={address2}&benchmark=2020&format=json')
        results = df.iloc[:1, 0][0][0]['coordinates']
        lat, lon = results['y'], results['x']
    except:
        geolocator = Nominatim(user_agent="GTA Lookup")
        geocode = RateLimiter(geolocator.geocode, min_delay_seconds=1)
        location = geolocator.geocode(address)
        lat, lon = location.latitude, location.longitude
    return lat, lon


#Side Bar
address = st.sidebar.text_input(
    "Address", "Dallas, TX")
date = st.sidebar.date_input(
    "Loss Date",  pd.Timestamp(2023, 7, 14), key='date')
date_str = date.strftime("%Y%m%d")

#Geocode Addreses
lat, lon = geocode(address)

#Filter Data
df_hail_cut = get_data(lat, lon, date_str)


df_hail_cut["Lat_address"] = lat
df_hail_cut["Lon_address"] = lon
df_hail_cut['Miles to Hail'] = [
    distance(i) for i in df_hail_cut[['LAT', 'LON', 'Lat_address', 'Lon_address']].values]
df_hail_cut['MAXSIZE'] = df_hail_cut['MAXSIZE'].round(1)
df_hail_cut = df_hail_cut.query("`Miles to Hail`<10")
df_hail_cut['Category'] =  np.where(df_hail_cut['Miles to Hail'] < 1, "Within 1 Mile",
                                            np.where(df_hail_cut['Miles to Hail'] < 3, "Within 3 Miles",
                                             np.where( df_hail_cut['Miles to Hail'] < 5, "Within 5 Miles",
                                                     np.where(df_hail_cut['Miles to Hail'] < 10, "Within 10 Miles", 'Other'))))

df_hail_cut_group = pd.pivot_table(df_hail_cut, index='Date_utc',
                                   columns='Category',
                                   values='MAXSIZE',
                                   aggfunc='max')

cols = df_hail_cut_group.columns
cols_focus = [ "Within 1 Mile","Within 5 Miles",
              "Within 3 Miles", "Within 10 Miles"]

missing_cols = set(cols_focus)-set(cols)
for c in missing_cols:
    df_hail_cut_group[c] = np.nan

df_hail_cut_group2 = df_hail_cut_group[cols_focus].query(
    "`Within 3 Miles`==`Within 3 Miles`")

for i in range(len(cols)):
    df_hail_cut_group2[cols_focus[i+1]] = np.where(df_hail_cut_group2[cols_focus[i+1]].fillna(0) <
                                                   df_hail_cut_group2[cols_focus[i]].fillna(0),
                                                   df_hail_cut_group2[cols_focus[i]],
                                                   df_hail_cut_group2[cols_focus[i+1]])



df_hail_cut_group2 = df_hail_cut_group2.sort_index(ascending=False)

df_hail_cut_group2.index = pd.to_datetime(
    df_hail_cut_group2.index, format='%Y%m%d').strftime("%Y-%m-%d")


#Map Data
m = map_location(address, lat, lon)

#Display
col1, col2 = st.columns((3, 2))

with col1:
    st.header('Estimated Maximum Hail Size')
    st.write('Data from 2010 to 2023-09-24')
    df_hail_cut_group2
    csv2 = convert_df(df_hail_cut_group2.reset_index())
    st.download_button(
        label="Download data as CSV",
        data=csv2,
        file_name=f'{address}_{date_str}.csv',
        mime='text/csv')
with col2:
    st.header('Map')
    st_folium(m, height=400)