TrafficPerformanceScore.py

#!/usr/bin/env python
# coding: utf-8

import streamlit as st
import pandas as pd
import numpy as np
import time
import pypyodbc as pyodbc
import datetime
import pydeck as pdk
import plotly.graph_objects as go
from plotly.subplots import make_subplots
from PIL import Image
from requests_html import HTMLSession
import locale
import folium
import base64
import copy
import math

from sys import platform 
if platform == "linux" or platform == "linux2":
    # linux
	SQL_DRIVER = 'ODBC Driver 17 for SQL Server'

elif platform == "darwin":
    # OS X
	SQL_DRIVER = 'ODBC Driver 17 for SQL Server'

elif platform == "win32":
    # Windows...
	SQL_DRIVER = 'ODBC Driver 17 for SQL Server'
	SQL_DRIVER = 'SQL Server'
	
import warnings
warnings.filterwarnings("ignore")


from Visualization import GenerateGeo, GenerateGeoAnimation
#####################################################
# SQL query functions
#####################################################
#@st.cache(allow_output_mutation=True)
def getDatabaseConnection():
	return pyodbc.connect(f'DRIVER={SQL_DRIVER};SERVER=128.95.29.74;DATABASE=RealTimeLoopData;UID=starlab;PWD=star*lab1')

@st.cache
def getLoopDetectorLocation():
	conn = getDatabaseConnection()
	SQL_Query = pd.read_sql_query(
	'''	SELECT Distinct [CabName]
		      ,cab.[Lat]
		      ,cab.[Lon]
		  FROM [RealTimeLoopData].[dbo].[cabinets] as cab join [RealTimeLoopData].[dbo].[MinuteDataDefnNW] as def on cab.UnitName = def.id
		  WHERE SUBSTRING(def.id, 12, 2) = '_M' AND (SUBSTRING(def.id, 15, 3) = '___' OR SUBSTRING(def.id, 15, 3) = 'H__')
		  AND def.[covered_dist] > 0 AND def.[covered_dist] <= 2
		  AND [CabName] IS NOT NULL AND cab.[Lat] IS NOT NULL''', conn)

	return pd.DataFrame(SQL_Query)


def getTrafficIndex(date):
	conn = getDatabaseConnection()

	SQL_Query = pd.read_sql_query(
	'''SELECT DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0) as [time]
	      , AVG([AVG_Spd_GP]) AS [AVG_Spd_GP]
	      , AVG([AVG_Vol_GP]) AS [AVG_Vol_GP]
	      , AVG([TrafficIndex_GP]) AS [TrafficIndex_GP]
	      , AVG([AVG_Spd_HOV]) AS [AVG_Spd_HOV]
	      , AVG([AVG_Vol_HOV]) AS [AVG_Vol_HOV]
	      , AVG([TrafficIndex_HOV]) AS [TrafficIndex_HOV]
	  FROM [RealTimeLoopData].[dbo].[TrafficIndex]
	  WHERE CAST([time] AS DATE) = ?
	  GROUP BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0) 
	  ORDER BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0) 
	  ''', conn, params = [date])

	return pd.DataFrame(SQL_Query)


def getTrafficIndexMultiDays(sdate, edate):
    conn = getDatabaseConnection()

    SQL_Query = pd.read_sql_query(
    '''SELECT DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0) as [time]
	      , AVG([AVG_Spd_GP]) AS [AVG_Spd_GP]
	      , AVG([AVG_Vol_GP]) AS [AVG_Vol_GP]
	      , AVG([TrafficIndex_GP]) AS [TrafficIndex_GP]
	      , AVG([AVG_Spd_HOV]) AS [AVG_Spd_HOV]
	      , AVG([AVG_Vol_HOV]) AS [AVG_Vol_HOV]
	      , AVG([TrafficIndex_HOV]) AS [TrafficIndex_HOV]
	  FROM [RealTimeLoopData].[dbo].[TrafficIndex]
	  WHERE time between ? and ?
	  GROUP BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0) 
	  ORDER BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0)
      ''', conn, params = [sdate,edate])

    return pd.DataFrame(SQL_Query)


def getDailyIndex(sdate, edate):
    conn = getDatabaseConnection()
    SQL_Query = pd.read_sql_query(
    '''	SELECT convert(varchar, CAST([time] AS DATE), 107) as Date, AVG([TrafficIndex_GP]) as daily_index_GP, AVG([TrafficIndex_HOV]) as daily_index_HOV
		FROM [RealTimeLoopData].[dbo].[TrafficIndex]
		WHERE CAST([time] AS DATE) BETWEEN ? AND ?
			AND ( (DATEPART(HOUR, [time]) >= 6 AND DATEPART(HOUR, [time]) <= 9) OR (DATEPART(HOUR, [time]) >= 15 and DATEPART(HOUR, [time]) <= 18) )
		GROUP BY CAST([time] AS DATE)
		ORDER BY CAST([time] AS DATE)
      	''', conn, params = [sdate,edate])
    return pd.DataFrame(SQL_Query)


def getSegments():
	conn = getDatabaseConnection()
	SQL_Query = pd.read_sql_query(
	'''	SELECT *
		  FROM [RealTimeLoopData].[dbo].[Segments]''', conn)
	return pd.DataFrame(SQL_Query)

def getSegmentTPS_Day(sdate, edate, segmentID):
	conn = getDatabaseConnection()
	SQL_Query = pd.read_sql_query(
	'''	SELECT CONVERT(varchar, CAST([time] AS DATE), 107) as [time]
		      ,AVG([AVG_Spd_GP]) AS [AVG_Spd_GP]
		      ,AVG([AVG_Spd_HOV]) AS [AVG_Spd_HOV]
		      ,AVG([AVG_Vol_GP]) AS [AVG_Vol_GP]
		      ,AVG([AVG_Vol_HOV]) AS [AVG_Vol_HOV]
		      ,AVG([TrafficIndex_GP]) AS [TrafficIndex_GP]
		      ,AVG([TrafficIndex_HOV]) AS [TrafficIndex_HOV]
		FROM [RealTimeLoopData].[dbo].[SegmentTrafficIndex]
		WHERE [time] BETWEEN ? and ?
		AND [segmentID] = ?
		GROUP BY CAST([time] AS DATE)
		ORDER BY CAST([time] AS DATE)
		''', conn, params = [sdate, edate, segmentID])
	return pd.DataFrame(SQL_Query)

# Zhiyong to confirm whether this is a no-use code clip
# def getSegmentTPS_5Min(sdate, edate, segmentID):
# 	conn = getDatabaseConnection()
# 	SQL_Query = pd.read_sql_query(
# 	'''	SELECT DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0) as [time]
# 		      ,AVG([AVG_Spd_GP]) AS [AVG_Spd_GP]
# 		      ,AVG([AVG_Spd_HOV]) AS [AVG_Spd_HOV]
# 		      ,AVG([AVG_Vol_GP]) AS [AVG_Vol_GP]
# 		      ,AVG([AVG_Vol_HOV]) AS [AVG_Vol_HOV]
# 		      ,AVG([TrafficIndex_GP]) AS [TrafficIndex_GP]
# 		      ,AVG([TrafficIndex_HOV]) AS [TrafficIndex_HOV]
# 		FROM [RealTimeLoopData].[dbo].[SegmentTrafficIndex]
# 		WHERE [time] BETWEEN ? and ?
# 		AND [segmentID] = ?
# 		GROUP BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0)
# 		ORDER BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0)
# 		''', conn, params = [sdate, edate, segmentID])
# 	return pd.DataFrame(SQL_Query)

def getSegmentTPS_5Min(sdate, edate):
	conn = getDatabaseConnection()
	SQL_Query = pd.read_sql_query(
	'''	SELECT DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0) as [time]
				,[segmentID]
		      	,AVG([AVG_Spd_GP]) AS [AVG_Spd_GP]
		      	,AVG([AVG_Spd_HOV]) AS [AVG_Spd_HOV]
		      	,AVG([AVG_Vol_GP]) AS [AVG_Vol_GP]
		      	,AVG([AVG_Vol_HOV]) AS [AVG_Vol_HOV]
		      	,AVG([TrafficIndex_GP]) AS [TrafficIndex_GP]
		      	,AVG([TrafficIndex_HOV]) AS [TrafficIndex_HOV]
		FROM [RealTimeLoopData].[dbo].[SegmentTrafficIndex]
		WHERE [time] BETWEEN ? and ?
		GROUP BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0),[segmentID]
		ORDER BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/5*5, 0),[segmentID]
		''', conn, params = [sdate, edate])
	return pd.DataFrame(SQL_Query)


def getSegmentTPS_1Hour(sdate, edate):
	conn = getDatabaseConnection()
	SQL_Query = pd.read_sql_query(
	'''	SELECT DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/60*60, 0) as [time]
				,[segmentID]
		      	,AVG([AVG_Spd_GP]) AS [AVG_Spd_GP]
		      	,AVG([AVG_Spd_HOV]) AS [AVG_Spd_HOV]
		      	,AVG([AVG_Vol_GP]) AS [AVG_Vol_GP]
		      	,AVG([AVG_Vol_HOV]) AS [AVG_Vol_HOV]
		      	,AVG([TrafficIndex_GP]) AS [TrafficIndex_GP]
		      	,AVG([TrafficIndex_HOV]) AS [TrafficIndex_HOV]
		FROM [RealTimeLoopData].[dbo].[SegmentTrafficIndex]
		WHERE [time] BETWEEN ? and ?
		GROUP BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/60*60, 0),[segmentID]
		ORDER BY DATEADD(MINUTE, DATEDIFF(MINUTE, 0, [time])/60*60, 0),[segmentID]
		''', conn, params = [sdate, edate])
	return pd.DataFrame(SQL_Query)


def getMorningPeakVolume(sdate, edate):
    conn = getDatabaseConnection()

    SQL_Query = pd.read_sql_query(
    '''	SELECT convert(varchar, CAST([time] AS DATE), 107) as Date, AVG(AVG_Vol_GP) as AVG_Vol_GP, AVG(AVG_Vol_HOV) as AVG_Vol_HOV
		FROM [RealTimeLoopData].[dbo].[TrafficIndex]
		WHERE CAST([time] AS DATE) between ? and ?
			AND DATEPART(HOUR, [time]) >= 6 AND DATEPART(HOUR, [time]) <= 9
		GROUP BY CAST([time] AS DATE)
		ORDER BY CAST([time] AS DATE)
      	''', conn, params = [sdate,edate])

    return pd.DataFrame(SQL_Query)


def getEveningPeakVolume(sdate, edate):
    conn = getDatabaseConnection()

    SQL_Query = pd.read_sql_query(
    '''	SELECT convert(varchar, CAST([time] AS DATE), 107) as Date, AVG(AVG_Vol_GP) as AVG_Vol_GP, AVG(AVG_Vol_HOV) as AVG_Vol_HOV
		FROM [RealTimeLoopData].[dbo].[TrafficIndex]
		WHERE CAST([time] AS DATE) between ? and ?
			AND DATEPART(HOUR, [time]) >= 15 and DATEPART(HOUR, [time]) <= 18
		GROUP BY CAST([time] AS DATE)
		ORDER BY CAST([time] AS DATE)
      	''', conn, params = [sdate,edate])

    return pd.DataFrame(SQL_Query)


def getVMT(sdate, edate):
    conn = getDatabaseConnection()

    SQL_Query = pd.read_sql_query(
    '''	SELECT convert(varchar, CAST([time] AS DATE), 107) as Date, SUM(VMT_GP) + SUM(VMT_HOV) as VMT
		FROM [RealTimeLoopData].[dbo].[TrafficIndex]
		WHERE CAST([time] AS DATE) between ? and ?
		GROUP BY CAST([time] AS DATE)
		ORDER BY CAST([time] AS DATE)
      	''', conn, params = [sdate,edate])

    return pd.DataFrame(SQL_Query)

def getCOVID19Info():
	return pd.read_csv('Washington_COVID_Cases.csv') 

def showCOVID19Figure():
	# get COVID info and update csv
	url = 'https://en.wikipedia.org/wiki/Template:2019%E2%80%9320_coronavirus_pandemic_data/United_States/Washington_State_medical_cases_chart'
	df_COVID19 = update_and_get_covid19_info(url)
	df_COVID19['date'] = df_COVID19['date'].astype('datetime64[ns]')
	# st.write(df_COVID19)
	sdate = datetime.datetime(2020, 2, 28)
	edate = df_COVID19.loc[len(df_COVID19)-1, 'date']
	# sdate = st.date_input('Select a start date', value=datetime.datetime(2020, 2, 28))
	# edate = st.date_input('Select an end date', value=df_COVID19.loc[len(df_COVID19)-1, 'date'])
	# daily index
	df_DailyIndex = getDailyIndex(sdate, edate)

	# # remove outliers from HOV traffic index
	# df_DailyIndex.loc[df_DailyIndex['avg_vol_hov'] == 0, 'trafficindex_hov'] = 1.0

	df_DailyIndex['date'] = df_DailyIndex['date'].astype('datetime64[ns]')
	df_DailyIndex = df_DailyIndex[['date', 'daily_index_gp', 'daily_index_hov']]

	df_DailyIndex['daily_index_gp'] = df_DailyIndex['daily_index_gp'] * 100
	# df_DailyIndex['daily_index_gp'] = df_DailyIndex['daily_index_gp'].astype('int64')
	df_DailyIndex['daily_index_hov'] = df_DailyIndex['daily_index_hov'] * 100
	# df_DailyIndex['daily_index_hov'] = df_DailyIndex['daily_index_hov'].astype('int64')

	# # peak volume
	# df_mpv = getMorningPeakVolume(sdate, edate)
	# df_mpv['date'] = df_mpv['date'].astype('datetime64[ns]')
	# df_mpv.rename(columns = {'avg_vol_gp':'Morning_GP', 'avg_vol_hov':'Morning_HOV'}, inplace = True) 
	# df_epv = getEveningPeakVolume(sdate, edate)
	# df_epv['date'] = df_epv['date'].astype('datetime64[ns]')
	# df_epv.rename(columns = {'avg_vol_gp':'Evening_GP', 'avg_vol_hov':'Evening_HOV'}, inplace = True) 
	# df_pv = pd.merge(df_mpv, df_epv, on='date')
	
	data = pd.merge(df_DailyIndex, df_COVID19, on='date', how='left')

	# st.write(data['confirmed case'].max())
	confirmed_case_axis_max = data['confirmed case'].max() + 500
	lw = 2  # line width

	# Create figure with secondary y-axis
	fig = make_subplots(specs=[[{"secondary_y": True}]])

	# Add traces for axis-2
	fig.add_trace(go.Scatter(x=data['date'], y=data['daily_index_gp'],
							 mode='lines', line=dict(dash='dot', width=lw, color='#1f77b4'),
							 name='Network-wide TPS - GP',
							 legendgroup='group2'),
					secondary_y=False)
	fig.add_trace(go.Scatter(x=data['date'], y=data['daily_index_hov'],
							 mode='lines', line=dict(dash='dot', width=lw, color='#2ca02c'),
							 name='Network-wide TPS - HOV',
							 legendgroup='group2'),
					secondary_y=False)

	# Add traces for axis-1
	fig.add_trace(go.Scatter(x=data['date'], y=data['confirmed case'],
							 mode='lines+markers', line=dict(dash='solid', width=lw, color='red'),
							 name='Confirmed Cases',
							 legendgroup='group1'),
					secondary_y=True)
	fig.add_trace(go.Scatter(x=data['date'], y=data['new case'],
							 mode='lines+markers', line=dict(dash='solid', width=lw, color='orange'),
							 name='New Cases',
							 legendgroup='group1'),
					secondary_y=True)
	fig.add_trace(go.Scatter(x=data['date'], y=data['death case'],
	 						 mode='lines+markers', line=dict(dash='solid', width=lw, color='black'),
	 						 name='Total Death',
	 						 legendgroup='group1'),
	 				secondary_y=True)
	

	fig.update_traces(textposition='top center')
	# Set x-axis title
	fig.update_xaxes(title_text="Date")
	# Set y-axes titles
	fig.update_yaxes(title_text="Daily Traffic Performance Score (%)", 
						range=[70, 100], 
						showline=True, 
						linecolor='rgb(204, 204, 204)', 
						linewidth=2, 
						showticklabels=True, 
						ticks='outside', 
						secondary_y=False)
	fig.update_yaxes(title_text="COVID-19 Case Amount",
						range=[0, confirmed_case_axis_max], 
						showline=True, 
						linecolor='rgb(204, 204, 204)', 
						linewidth=2, 
						showticklabels=True, 
						ticks='outside', 
						secondary_y=True)
	fig.update_layout(xaxis=dict(
						    showline=True,
						    showgrid=False,
						    showticklabels=True,
						    linecolor='rgb(204, 204, 204)',
						    linewidth=2,
						    ticks='outside',
						    tickfont=dict(
						        family='Arial',
						        size=12,
						        color='rgb(82, 82, 82)',
						    ),
						),
						legend=dict(x= 0.4, y=1.3, orientation="h"),
					  	margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=4), 
					  	width = 700, 
					  	height = 450,
					  	plot_bgcolor='white')
	st.plotly_chart(fig)

def checkDateRange(date):
	out_of_range = False
	if date > datetime.datetime.now().date():
		date = datetime.datetime.now().date()
		out_of_range = True
	elif date < datetime.datetime(2019, 11, 1).date():
		date = datetime.datetime(2019, 11, 1).date()
		out_of_range = True
	return date, out_of_range

def checkDatesRange(sdate, edate):
	out_of_range = False
	dates_reversed = False
	dates_equal = False

	if sdate > edate:
		temp = sdate
		sdate = edate
		edate = temp
		dates_reversed = True

	if sdate == edate:
		sdate = sdate - datetime.timedelta(days=1)
		edate = edate + datetime.timedelta(days=1)
		dates_equal= True

	if sdate >= datetime.datetime.now().date():
		sdate = datetime.datetime.now().date() - datetime.timedelta(days=1)
		out_of_range = True
	elif sdate < datetime.datetime(2019, 11, 1).date():
		sdate = datetime.datetime(2019, 11, 1).date()
		out_of_range = True

	if edate > datetime.datetime.now().date():
		edate = datetime.datetime.now().date()
		out_of_range = True
	elif edate < datetime.datetime(2019, 11, 1).date():
		edate = datetime.datetime(2019, 11, 1).date() + datetime.timedelta(days=1)
		out_of_range = True

	return sdate, edate, out_of_range, dates_reversed, dates_equal

def showDatesWarnings(out_of_range, dates_reversed, dates_equal):
	if out_of_range and (dates_reversed or dates_equal):
		st.write('(Note: Date available from', datetime.datetime(2019, 11, 1).date(), 'to', datetime.datetime.now().date(), '. End Date should be greater than Start Date)')
	elif out_of_range:
		st.write('(Note: Date available from', datetime.datetime(2019, 11, 1).date(), 'to', datetime.datetime.now().date(), ')')
	elif dates_reversed or dates_equal:
		st.write('(Note: End Date should be greater than Start Date)')
	

#####################################################
# display functions
#####################################################

def IntroduceTrafficIndex():
	###########
	# Sidebar #
	###########
	# st.sidebar.markdown("## Components")
	# st.sidebar.checkbox("Data Course")
	# st.sidebar.checkbox("Traffic Performance Score Caculation")
	###########
	# Content #
	###########
	st.markdown("# Traffic Performance Score in the Greater Seattle Area")
	# st.markdown("## Introduction to Traffic Performance Score")
	# st.markdown("In this website, Traffic Performance Score (TPS) indicating the overall performance "
	# 			"of the freeway networks in the Greater Seattle area is calculated and visualized. "
				
	# 			# "With this website, you can view "
	# 			# "\n * Temporal dynamic of network-wide TPS of different types of lanes with various time resolutions, ranging from 5 minutes to one day."
	# 			# "\n * Varying Spatial distribution of segment-based TPS on interactive maps. "
	# 			# "\n * Traffic changes in response to COVID-19 reflected by the TPS. "
	# 			# "\n * Other traffic performance metrics. " 
	# 			)
	# st.markdown("The **TPS** is a value ranges from 0% to 100%. "
	# 			"The closer to 100% the **TPS** is, the better the overall network-wide traffic condition is. "
	# 			"The TPS calculation and the data source are described in the ***About*** page. ")
	st.markdown("Urban traffic status affects our everyday lives. "
				"This website shows timely measurements of the traffic performance in the Greater Seattle area. "
				"To view more information, please select on the left ***navigation*** panel. "
				"Your feedback is greatly welcome.")
	

	#################################################################
	# st.markdown("## Traffic Changes in Response to COVID-19")
	
	# showCOVID19Figure()

	#################################################################
	# st.markdown("## Segment-based Traffic Performance Score")

	date = st.date_input('Select a date:', value = datetime.datetime.now().date())
	date, out_of_range = checkDateRange(date)
	if out_of_range:
		st.write('Selected date:', date, '(Data available from', datetime.datetime(2019, 11, 1).date(), 'to', datetime.datetime.now().date(), ')')
	datatime1 = datetime.datetime.combine(date, datetime.time(00, 00))
	datatime2 = datetime.datetime.combine(date, datetime.time(23, 59))
	df_SegTPS = getSegmentTPS_1Hour(datatime1, datatime2)

	#################################################################
	st.markdown("### Segment-based TPS on Animated Map")			
	GenerateGeoAnimation(copy.copy(df_SegTPS))

	#################################################################
	st.markdown("### Segment-based TPS Chart")

	segments = getSegments()
	segments['route_dir'] = 'Route ' + segments['route'].astype(int).astype(str) + '\t, ' + segments['direction'] + 'B'
	segments['milepost_pair'] = 'Milepost ( ' + segments['milepost_small'].astype(str) + '\t, ' + segments['milepost_large'].astype(str) +' )'
	segments_route_dir = segments['route_dir'].drop_duplicates()
	route_dir = st.selectbox("Select a route:", segments_route_dir.values.tolist())

	segments_milepost_pair = segments[segments['route_dir'] == route_dir]['milepost_pair']
	milepost_pair = st.selectbox("Select a segment:", segments_milepost_pair.values.tolist())

	segmentID = segments[(segments['route_dir'] == route_dir) & (segments['milepost_pair'] == milepost_pair)]['segmentid']

	df_TI = df_SegTPS[df_SegTPS['segmentid'] == segmentID.values[0]]

	# remove outliers from HOV traffic index
	df_TI.loc[df_TI['avg_vol_hov'] == 0, 'trafficindex_hov'] = 1.0

	df_TI['trafficindex_gp'] = df_TI['trafficindex_gp'] * 100
	# df_TI['trafficindex_gp'] = df_TI['trafficindex_gp'].astype('int64')
	df_TI['trafficindex_hov'] = df_TI['trafficindex_hov'] * 100
	# df_TI['trafficindex_hov'] = df_TI['trafficindex_hov'].astype('int64')

	sampling_interval = 1
	data = df_TI.loc[::sampling_interval, ['time', 'trafficindex_gp', 'trafficindex_hov']]
	# minimum_score = 0
	minimum_score = min(data['trafficindex_gp'].min(), data['trafficindex_hov'].min())
	# st.write(minimum_score)
	if not math.isnan(minimum_score):
		minimum_score = round(minimum_score//5 *5)
	else:
		minimum_score = 0
	lw = 1  # line width
	# Create traces
	fig = go.Figure()
	fig.add_trace(go.Scatter(x=data['time'], y=data['trafficindex_gp'],
							 mode='lines', line=dict(dash='solid', width=lw),
							 name='Main lane'))

	fig.add_trace(go.Scatter(x=data['time'], y=data['trafficindex_hov'],
							 mode='lines', line=dict(dash='solid', width=lw),
							 name='HOV lane'))

	fig.update_layout(xaxis=dict(title_text='Time', showticklabels=True),
					  yaxis=dict(title_text='Traffic Performance Score (%)', range = [minimum_score, 100], showticklabels=True),
					  legend = dict(x=.01, y=0),
					  margin = go.layout.Margin(l=50, r=0, b=50, t=10, pad=20), width = 700, height = 450)

	#fig.update_yaxes(range=[0, 1.1])

	# st.write('Traffic Performance Score of (', date, '):')

	st.plotly_chart(fig)

	# dataFields = st.multiselect('Show Data',  list(df_TI.columns.values), default = ['time', 'trafficindex_gp', 'trafficindex_hov'] )
	# st.write(df_TI[dataFields])
	st.write("Main lane: general purpose (GP) lane. ")
	st.write("HOV lane: high-occupancy vehicle lane, also known as carpool lane. ")

	st.write('<a href="https://clustrmaps.com/site/1b7ap" title="Visit tracker"><img src="//clustrmaps.com/map_v2.png?cl=ffffff&w=70&t=n&d=jn07mPkuDBD9jMBfRsCUgcfZN5e7Z2SydqZ3ItFsfv4&co=ffffff&ct=ffffff" style="display:none"/></a>', unsafe_allow_html=True)


	
def showTrafficIndex():
	###########
	# Sidebar #
	###########
	# st.sidebar.markdown("## Components")

	# index = st.sidebar.radio( "Display:", ("Daily Index", "Traffic Performance Score per Minute", "Tabular Data"))
	# daily_Index = st.sidebar.checkbox("Daily Index", value = True)
	# minute_Index = st.sidebar.checkbox("Traffic Performance Score per Minute")
	# tablular_Data = st.sidebar.checkbox("Tabular Data")
	daily_Index, five_minute_Index, tablular_Data = True, True, True
	########################
	# main content
	########################
	st.markdown("# Network-based TPS")
	st.markdown("* In this section, the network-wide TPS in the Greater Seattle area is provided based on the selected start and end dates. "
				"\n * TPS in both one day and 5-minute intervals is visualized. "
				"\n * Downloadable TPS tabular data in 5-minute intervals is also provided at the bottom of the page.")
	# "You can check or uncheck the checkbox in the left panel to adjuect the displayed information.")
	# sdate = st.date_input('Pick a start date', value = (datetime.datetime.now() - datetime.timedelta(days=30)))
	# edate = st.date_input('Pick an end date', value = datetime.datetime.now().date())
	# st.write('From ',sdate, ' to ', edate,':')
	

	########################
	# Daily Traffic Performance Score #
	########################

	if daily_Index:
		st.markdown("## Daily Traffic Performance Score")

		sdate_DI = st.date_input('Select a start date:', value = (datetime.datetime.now() - datetime.timedelta(days=90)))
		edate_DI = st.date_input('Select an end date:' , value = datetime.datetime.now().date())

		# Check dates ranges and show warnings
		sdate_DI, edate_DI, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate_DI, edate_DI)
		showDatesWarnings(out_of_range, dates_reversed, dates_equal)

		st.write('**Selected Dates** from ',sdate_DI, ' to ', edate_DI,':')

		df_DailyIndex = getDailyIndex(sdate_DI, edate_DI)

		df_DailyIndex['date'] = df_DailyIndex['date'].astype('datetime64[ns]')
		df_DailyIndex['date'] = df_DailyIndex['date'].dt.date

		df_DailyIndex['daily_index_gp'] = df_DailyIndex['daily_index_gp'] * 100
		# df_DailyIndex['daily_index_gp'] = df_DailyIndex['daily_index_gp'].astype('int64')
		df_DailyIndex['daily_index_hov'] = df_DailyIndex['daily_index_hov'] * 100
		# df_DailyIndex['daily_index_hov'] = df_DailyIndex['daily_index_hov'].astype('int64')

		data = df_DailyIndex[['date', 'daily_index_gp', 'daily_index_hov']]
		minimum_score = min(data['daily_index_gp'].min(), data['daily_index_hov'].min())
		minimum_score = round(minimum_score // 5 * 5)
		lw = 1  # line width
		# Create traces
		fig = go.Figure()
		fig.add_trace(go.Scatter(x=data['date'], y=data['daily_index_gp'],
								 mode='lines', line=dict(dash='solid', width=lw),
								 name='Main lane'))
		fig.add_trace(go.Scatter(x=data['date'], y=data['daily_index_hov'],
								 mode='lines', line=dict(dash='solid', width=lw),
								 name='HOV lane'))
		fig.update_layout(xaxis_title='Date',
						  yaxis=dict(title_text='Traffic Performance Score (%)', range=[minimum_score, 100],
									 showticklabels=True),
						  legend=dict(x=.01, y=0),
						  margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=20), width=700, height=450)
		st.plotly_chart(fig)

	########################
	# Minute Traffic Performance Score 
	########################
	if five_minute_Index:

		st.markdown("## Traffic Performance Score per 5-Minute")

		sdate_MI = st.date_input('Pick a start date:', value = (datetime.datetime.now() - datetime.timedelta(days=30)))
		edate_MI = st.date_input('Pick an end date:', value = datetime.datetime.now().date())

		# Check dates ranges and show warnings
		sdate_MI, edate_MI, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate_MI, edate_MI)
		showDatesWarnings(out_of_range, dates_reversed, dates_equal)

		st.write('**Selected Dates** from ',sdate_MI, ' to ', edate_MI,':')

		df_TI_range = getTrafficIndexMultiDays(sdate_MI, edate_MI)
		
		# remove outliers from HOV traffic index
		df_TI_range.loc[df_TI_range['avg_vol_hov'] == 0, 'trafficindex_hov'] = 1.0

		df_TI_range['trafficindex_gp'] = df_TI_range['trafficindex_gp'] * 100
		# df_TI_range['trafficindex_gp'] = df_TI_range['trafficindex_gp'].astype('int64')
		df_TI_range['trafficindex_hov'] = df_TI_range['trafficindex_hov'] * 100
		# df_TI_range['trafficindex_hov'] = df_TI_range['trafficindex_hov'].astype('int64')

		sampling_interval = 1
		data = df_TI_range.loc[::sampling_interval, ['time', 'trafficindex_gp', 'trafficindex_hov']]

		# st.write(df_TI_range['time'].dtypes)

		lw = 1  # line width
		minimum_score = min(data['trafficindex_gp'].min(), data['trafficindex_hov'].min())
		minimum_score = round(minimum_score // 5 * 5)
		# Create traces
		fig = go.Figure()
		fig.add_trace(go.Scatter(x=data['time'], y=data['trafficindex_gp'],
								 mode='lines', line=dict(dash='solid', width=lw),
								 name='Main lane'))
		fig.add_trace(go.Scatter(x=data['time'], y=data['trafficindex_hov'],
								 mode='lines', line=dict(dash='solid', width=lw),
								 name='HOV lane'))
		fig.update_layout(xaxis_title='Time',
						  yaxis=dict(title_text='Traffic Performance Score (%)', range=[minimum_score, 100],
									 showticklabels=True),
						  legend=dict(x=.01, y=0),
						  margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=20), width=700, height=450)
		st.plotly_chart(fig)

		
	########################
	# Minute Traffic Performance Score 
	########################
	if tablular_Data:
		st.markdown("## Traffic Performance Score Tablular Data")

		sdate_TD = st.date_input('Pick a start date', value = (datetime.datetime.now() - datetime.timedelta(days=1)))
		edate_TD = st.date_input('Pick an end date', value = datetime.datetime.now().date())
		
		# Check dates ranges and show warnings
		sdate_TD, edate_TD, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate_TD, edate_TD)
		showDatesWarnings(out_of_range, dates_reversed, dates_equal)

		st.write('**Selected Dates** from ',sdate_TD, ' to ', edate_TD,':')

		df_TI_range = getTrafficIndexMultiDays(sdate_TD, edate_TD)
		
		# remove outliers from HOV traffic index
		df_TI_range.loc[df_TI_range['avg_vol_hov'] == 0, 'trafficindex_hov'] = 1.0
		df_TI_range['trafficindex_gp'] = df_TI_range['trafficindex_gp'] * 100
		df_TI_range['trafficindex_hov'] = df_TI_range['trafficindex_hov'] * 100
		

		# rename column headers
		df_TI_range.columns = ['Time', 'AVG_Spd_GP', 'AVG_Vol_GP', 'TPS_GP', 'AVG_Spd_HOV', 'AVG_Vol_HOV', 'TPS_HOV']
		
		# set precision of each column
		df_TI_range[['AVG_Spd_GP', 'TPS_GP', 'AVG_Spd_HOV', 'TPS_HOV']] \
			= df_TI_range[['AVG_Spd_GP', 'TPS_GP', 'AVG_Spd_HOV', 'TPS_HOV']].applymap("{0:.1f}".format)
		df_TI_range[['AVG_Vol_GP', 'AVG_Vol_HOV']] \
			= df_TI_range[['AVG_Vol_GP', 'AVG_Vol_HOV']].applymap("{0:.0f}".format)

		dataFields = st.multiselect('Show Data Type',  list(df_TI_range.columns.values)
						, default = ['Time', 'AVG_Spd_GP', 'AVG_Vol_GP', 'TPS_GP', 'AVG_Spd_HOV', 'AVG_Vol_HOV', 'TPS_HOV']
		 )
		st.write(df_TI_range[dataFields])

		st.markdown("Download the tabular data as a CSV file:")
		st.markdown(get_table_download_link(df_TI_range[dataFields]), unsafe_allow_html=True)

def get_table_download_link(df, filename = 'data'):
    """Generates a link allowing the data in a given panda dataframe to be downloaded
    in:  dataframe
    out: href string
    """
    csv = df.to_csv(index=False)
    b64 = base64.b64encode(csv.encode()).decode()  # some strings <-> bytes conversions necessary here
    href = f'<a href="data:file/csv;base64,{b64}" download="{filename}.csv">Download csv file</a>'
    return href

def get_data_from_sel(url, sel):
    session = HTMLSession()
    r = session.get(url)
    mylist = []
    try:
        results = r.html.find(sel)
        for result in results:
            mytext = result.text
            mylist.append(mytext)
        return mylist
    except:
        return None


def update_and_get_covid19_info(url):

    sel_date = '#mw-content-text > div > div.barbox.tright > div > table > tbody > tr > td:nth-child(1)'
    sel_cases = '#mw-content-text > div > div.barbox.tright > div > table > tbody > tr > td:nth-child(3) > span.cbs-ibr'
    sel_death = '#mw-content-text > div > div.barbox.tright > div > table > tbody > tr > td:nth-child(4) > span.cbs-ibr'
    
    try:

        df_csv = getCOVID19Info()
        date_list = get_data_from_sel(url, sel_date)
        if df_csv.loc[len(df_csv)-1,'date'] != date_list[len(date_list)-2]:

            # the first and last items are not data
            del date_list[len(date_list) - 1]
            del date_list[0]
            cases_list_0 = get_data_from_sel(url, sel_cases)
            death_list_0 = get_data_from_sel(url, sel_death)
            # remove the thousand seprators in cases_list_0 and death_list_0
            locale.setlocale(locale.LC_ALL, 'en_US.UTF-8')
            cases_list, death_list = [], []
            for n in cases_list_0:
                cases_list.append(locale.atoi(n))
            for n in death_list_0:
                death_list.append(locale.atoi(n))

            df_web = pd.DataFrame({'date': date_list, 'confirmed case': cases_list, 'death case': death_list})
            # calculate new case based on confirmed case
            df_web['new case'] = df_web['confirmed case'] - df_web['confirmed case'].shift(1)
            df_web.drop(df_web.index[0], inplace=True)
            df_web['date'] = df_web['date'].astype('datetime64[ns]')
            
            # merge df_web and df_csv
            df_csv['date'] = df_csv['date'].astype('datetime64[ns]')
            df_new = df_csv.append(df_web, ignore_index=True)
            df_new = df_new.drop_duplicates(subset=['date'], keep='last')
            df_new.to_csv("Washington_COVID_Cases.csv", mode='w', header=True, index=False)

    finally:
        return getCOVID19Info()


def showSgementTPS():
	st.markdown("# Segment-based TPS")

	st.markdown("* In this section, TPS of freeway segments is provided and visualized on an interactive map. "
				"\n * Segment-based TPS is also visualized separately at the bottom of the page.")
	date = st.date_input('Select a date:', value = datetime.datetime.now().date())
	date, out_of_range = checkDateRange(date)
	if out_of_range:
		st.write('Selected date:', date, '(Data available from', datetime.datetime(2019, 11, 1).date(), 'to', datetime.datetime.now().date(), ')')
	
	datatime1 = datetime.datetime.combine(date, datetime.time(00, 00))
	datatime2 = datetime.datetime.combine(date, datetime.time(23, 59))
	

	
	# df_SegTPS_5Min = getSegmentTPS_5Min(datatime1, datatime2)
	# df_SegTPS_5Min.columns = ['time', 'segmentID', 'AVG_Spd_GP', 'AVG_Spd_HOV', 'AVG_Vol_GP', 'AVG_Vol_HOV', 'TrafficIndex_GP', 'TrafficIndex_HOV']
	

	# time = st.time_input('Pick an end date', value = datetime.datetime.now().time())
	# time = time.replace(second=0, microsecond=0)
	# dt = datetime.datetime.combine(date, time)
	# st.write(dt)
	# st.write(df_SegTPS_5Min)
	# st.write(df_SegTPS_5Min[df_SegTPS_5Min['time'] == dt])
	

	df_SegTPS = getSegmentTPS_1Hour(datatime1, datatime2)
	df_SegTPS.columns = ['time', 'segmentID', 'AVG_Spd_GP', 'AVG_Spd_HOV', 'AVG_Vol_GP', 'AVG_Vol_HOV', 'TrafficIndex_GP', 'TrafficIndex_HOV']
		
	annimation = True

	# annimation = st.radio( "Display Map", ('Dynamic Map', 'Animated Map')

	# if annimation == 'Dynamic Map':
	# GenerateGeoAnimation(df_SegTPS)
	# else:
	dt = st.selectbox('Select a time:', df_SegTPS['time'].astype(str).unique().tolist())
	TPS = df_SegTPS[df_SegTPS['time'] == dt]
	GenerateGeo(TPS)

	
	# map.save('index.html')
	# # st.write(m._repr_html_(), unsafe_allow_html=True)
	# st.write(map._repr_html_(), unsafe_allow_html=True)

	#####
	st.markdown("# Route-based Traffic Preformance Score")

	sdate = st.date_input('Select a start date:', value = (datetime.datetime.now() - datetime.timedelta(days=7)))
	edate = st.date_input('Select an end date:' , value = datetime.datetime.now().date())

	# Check dates ranges and show warnings
	sdate, edate, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate, edate)
	showDatesWarnings(out_of_range, dates_reversed, dates_equal)

	st.write('**Selected Dates** from ',sdate, ' to ', edate,':')

	# segments = getSegments()
	# segments['label'] = 'Route (' + segments['route'].astype(int).astype(str) + '),\t Direction (' + segments['direction'] + 'B),\t Milepost (' \
	# 					+ segments['milepost_small'].astype(int).astype(str) + ', ' + segments['milepost_large'].astype(int).astype(str) + ')'
	
	# segmentLabel = st.selectbox("", segments['label'].iloc[0:10].values.tolist())
	# segmentID = segments[segments['label'] == segmentLabel]['segmentid'].values.tolist()[0]

	segments = getSegments()
	segments['route_dir'] = 'Route ' + segments['route'].astype(int).astype(str) + '\t, ' + segments['direction'] + 'B'
	segments['milepost_pair'] = 'Milepost ( ' + segments['milepost_small'].astype(str) + '\t, ' + segments['milepost_large'].astype(str) +' )'
	segments_route_dir = segments['route_dir'].drop_duplicates()
	route_dir = st.selectbox("Select a route:", segments_route_dir.values.tolist())

	segments_milepost_pair = segments[segments['route_dir'] == route_dir]['milepost_pair']
	milepost_pair = st.selectbox("Select a segment:", segments_milepost_pair.values.tolist())

	segmentID = segments[(segments['route_dir'] == route_dir) & (segments['milepost_pair'] == milepost_pair)]['segmentid']

	df_SegTPS = getSegmentTPS_1Hour(sdate, edate)

	df_SegTPS = df_SegTPS[df_SegTPS['segmentid'] == segmentID.values[0]]

	# st.write(df_SegTPS)
	# df_SegTPS_Day = getSegmentTPS_Day(sdate, edate, segmentID)

	# st.write(segments['label'])

	# remove outliers from HOV traffic index
	df_SegTPS.loc[df_SegTPS['avg_vol_hov'] == 0, 'trafficindex_hov'] = 1.0

	df_SegTPS['trafficindex_gp'] = df_SegTPS['trafficindex_gp'] * 100
	df_SegTPS['trafficindex_hov'] = df_SegTPS['trafficindex_hov'] * 100
	df_SegTPS['time'] = pd.to_datetime(df_SegTPS['time'])

	lw = 1  # line width
	minimum_score = min(df_SegTPS['trafficindex_gp'].min(), df_SegTPS['trafficindex_hov'].min())
	minimum_score = round(minimum_score // 5 * 5)
	fig = go.Figure()
	fig.add_trace(go.Scatter(x=df_SegTPS['time'], y=df_SegTPS['trafficindex_gp'],
							 mode='lines', line=dict(dash='solid', width=lw),
							 name='Main lane'))
	fig.add_trace(go.Scatter(x=df_SegTPS['time'], y=df_SegTPS['trafficindex_hov'],
							 mode='lines', line=dict(dash='solid', width=lw),
							 name='HOV lane'))

	fig.update_layout(xaxis_title='Time',
					  yaxis=dict(title_text='Traffic Performance Score (%)', range=[minimum_score, 100],
								 showticklabels=True),
					  legend=dict(x=.01, y=0),
					  margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=20), width=700, height=450)
	#fig.update_yaxes(range=[0, 1.1])
	st.plotly_chart(fig)


def showCOVID19():

	# st.markdown("# Impact of COVID-19 on Traffic Changes")
	st.markdown("# How Does COVID-19 Affects TPS")
	st.markdown("## COVID-19 in Washington State")
	st.markdown("Since the early March 2020, the coronavirus outbreak has taken hold in the United States. "
				"Besides affecting public health, COVID-19 also has greatly impacted the whole transportation network, "
				"especially when businesses and public agencies shut down. The timeline of responses are listed as below: [\[source\]](https://en.wikipedia.org/wiki/2020_coronavirus_pandemic_in_Washington_(state))")
	st.markdown("* **March 6**: Major tech companies ask Seattle employees to work from home. Amazon and Facebook shut down individual offices as well. [\[link\]](https://www.theverge.com/2020/3/5/21166686/coronavirus-amazon-google-facebook-microsoft-twitter-seattle-staff-remote-work)\n"
				"* **March 9**: UW suspends on-site classes and finals. [\[link\]](https://www.washington.edu/coronavirus/2020/03/06/beginning-march-9-classes-and-finals-will-not-be-held-in-person-message-to-students/) \n"
				"* **March 13**: Gov. Inslee announces statewide school closures, expansion of limits on large gatherings. [\[link\]](https://medium.com/wagovernor/inslee-announces-statewide-school-closures-expansion-of-limits-on-large-gatherings-63d442111438) \n"
				"* **March 16**: Gov. Inslee announces statewide shutdown of restaurants, bars and expanded social gathering limits. [\[link\]](https://www.governor.wa.gov/news-media/inslee-statement-statewide-shutdown-restaurants-bars-and-limits-size-gatherings-expanded) \n"
				"* **March 23**: Gov. Inslee announces \"Stay Home, Stay Healthy\" order. [\[link\]](https://www.governor.wa.gov/news-media/inslee-announces-stay-home-stay-healthy%C2%A0order)\n"
				"* **April 2**: Gov. Inslee extends \"Stay Home, Stay Healthy\" through May 4. [\[link\]](https://www.governor.wa.gov/news-media/inslee-extends-stay-home-stay-healthy-through-may-4)")

	#################################################################
	st.markdown("## COVID-19 Cases")
	st.markdown("The following dynamic plot displays the progression of the coronavirus cases in Washington State.")

	st.write("<iframe src='https://public.flourish.studio/visualisation/1696713/embed' frameborder='0' scrolling='no' style='width:100%;height:300px;'></iframe><div style='width:100%!;margin-top:4px!important;text-align:right!important;'><a class='flourish-credit' href='https://public.flourish.studio/visualisation/1696713/?utm_source=embed&utm_campaign=visualisation/1696713' target='_top' style='text-decoration:none!important'><img alt='Made with Flourish' src='https://public.flourish.studio/resources/made_with_flourish.svg' style='width:105px!important;height:16px!important;border:none!important;margin:0!important;'> </a></div>", unsafe_allow_html=True)

	#################################################################
	st.markdown("## Impact of COVID-19 on Urban Traffic")
	st.markdown("This section shows the impact of COVID-19 on urban traffic. "
				"In the following chart, the trends of daily traffic performance scores and the coronavirus cases in Washington State are displayed together. "
				"Note: The coronavirus cases are caluculated since Feb. 28.")


	#################################################################
	# get COVID info and update csv
	# url = 'https://en.wikipedia.org/wiki/Template:2019%E2%80%9320_coronavirus_pandemic_data/United_States/Washington_State_medical_cases_chart'
	url = 'https://en.wikipedia.org/wiki/Template:COVID-19_pandemic_data/United_States/Washington_State_medical_cases_chart'
	df_COVID19 = update_and_get_covid19_info(url)
	df_COVID19['date'] = df_COVID19['date'].astype('datetime64[ns]')
	# st.write(df_COVID19)
	#sdate = datetime.datetime(2020, 2, 28)
	#edate = df_COVID19.loc[len(df_COVID19)-1, 'date']
	sdate = st.date_input('Select a start date', value=datetime.datetime(2020, 2, 28))
	edate = st.date_input('Select an end date', value=df_COVID19.loc[len(df_COVID19)-1, 'date'])

	# Check dates ranges and show warnings
	sdate, edate, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate, edate)
	showDatesWarnings(out_of_range, dates_reversed, dates_equal)

	st.write('**Selected Dates** from ',sdate, ' to ', edate,':')
	# daily index
	df_DailyIndex = getDailyIndex(sdate, edate)

	# # remove outliers from HOV traffic index
	# df_DailyIndex.loc[df_DailyIndex['avg_vol_hov'] == 0, 'trafficindex_hov'] = 1.0

	df_DailyIndex['date'] = df_DailyIndex['date'].astype('datetime64[ns]')
	df_DailyIndex = df_DailyIndex[['date', 'daily_index_gp', 'daily_index_hov']]

	df_DailyIndex['daily_index_gp'] = df_DailyIndex['daily_index_gp'] * 100
	# df_DailyIndex['daily_index_gp'] = df_DailyIndex['daily_index_gp'].astype('int64')
	df_DailyIndex['daily_index_hov'] = df_DailyIndex['daily_index_hov'] * 100
	# df_DailyIndex['daily_index_hov'] = df_DailyIndex['daily_index_hov'].astype('int64')

	# # peak volume
	# df_mpv = getMorningPeakVolume(sdate, edate)
	# df_mpv['date'] = df_mpv['date'].astype('datetime64[ns]')
	# df_mpv.rename(columns = {'avg_vol_gp':'Morning_GP', 'avg_vol_hov':'Morning_HOV'}, inplace = True) 
	# df_epv = getEveningPeakVolume(sdate, edate)
	# df_epv['date'] = df_epv['date'].astype('datetime64[ns]')
	# df_epv.rename(columns = {'avg_vol_gp':'Evening_GP', 'avg_vol_hov':'Evening_HOV'}, inplace = True) 
	# df_pv = pd.merge(df_mpv, df_epv, on='date')
	
	data = pd.merge(df_DailyIndex, df_COVID19, on='date', how='left')

	# st.write(data['confirmed case'].max())
	confirmed_case_axis_max = pd.to_numeric(data['confirmed case']).max() + 2000

	# st.write(confirmed_case_axis_max)

	lw = 2  # line width

	# Create figure with secondary y-axis
	fig = make_subplots(specs=[[{"secondary_y": True}]])

	# Add traces for axis-2
	fig.add_trace(go.Scatter(x=data['date'], y=data['daily_index_gp'],
							 mode='lines', line=dict(dash='dot', width=lw, color='#1f77b4'),
							 name='TPS - GP',
							 legendgroup='group2'),
					secondary_y=False)
	fig.add_trace(go.Scatter(x=data['date'], y=data['daily_index_hov'],
							 mode='lines', line=dict(dash='dot', width=lw, color='#2ca02c'),
							 name='TPS - HOV',
							 legendgroup='group2'),
					secondary_y=False)

	# Add traces for axis-1
	fig.add_trace(go.Scatter(x=data['date'], y=data['confirmed case'],
							 mode='lines+markers', line=dict(dash='solid', width=lw, color='red'),
							 name='Confirmed Cases',
							 legendgroup='group1'),
					secondary_y=True)
	fig.add_trace(go.Scatter(x=data['date'], y=data['new case'],
							 mode='lines+markers', line=dict(dash='solid', width=lw, color='orange'),
							 name='New Cases',
							 legendgroup='group1'),
					secondary_y=True)
	fig.add_trace(go.Scatter(x=data['date'], y=data['death case'],
	 						 mode='lines+markers', line=dict(dash='solid', width=lw, color='black'),
	 						 name='Total Death',
	 						 legendgroup='group1'),
	 				secondary_y=True)
	

	fig.update_traces(textposition='top center')
	# Set x-axis title
	fig.update_xaxes(title_text="Date")
	# Set y-axes titles
	fig.update_yaxes(title_text="Daily Traffic Performance Score (%)", 
						range=[70, 100], 
						showline=True, 
						linecolor='rgb(204, 204, 204)', 
						linewidth=2, 
						showticklabels=True, 
						ticks='outside', 
						secondary_y=False)
	fig.update_yaxes(title_text="COVID-19 Case Amount",
						range=[0, confirmed_case_axis_max], 
						showline=True, 
						linecolor='rgb(204, 204, 204)', 
						linewidth=2, 
						showticklabels=True, 
						ticks='outside', 
						secondary_y=True)
	fig.update_layout(xaxis=dict(
						    showline=True,
						    showgrid=False,
						    showticklabels=True,
						    linecolor='rgb(204, 204, 204)',
						    linewidth=2,
						    ticks='outside',
						    tickfont=dict(
						        family='Arial',
						        size=12,
						        color='rgb(82, 82, 82)',
						    ),
						),
						legend=dict(x= 0.5, y=1.3, orientation="h"),
					  	margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=4), 
					  	width = 700, 
					  	height = 450,
					  	plot_bgcolor='white')
	st.plotly_chart(fig)

	st.markdown("Download COVID-19 related data as a CSV file:")
	st.markdown(get_table_download_link(data), unsafe_allow_html=True)

def showVMT():
	
	st.markdown("# Vehicle Miles of Travel (VMT)")
	
	st.markdown('This page shows the variations of Vehicle Miles of Travel. '
		'Downloadable tablular data is shown at the bottom of this page. ')
	sdate_2 = st.date_input('Select a start date', value = (datetime.datetime.now() - datetime.timedelta(days=90)))
	edate_2 = st.date_input('Select an end date', value = datetime.datetime.now().date())
	
	# Check dates ranges and show warnings
	sdate_2, edate_2, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate_2, edate_2)
	showDatesWarnings(out_of_range, dates_reversed, dates_equal)

	st.write('**Selected Dates** from ',sdate_2, ' to ', edate_2,':')

	df_vmt = getVMT(sdate_2, edate_2)
	df_vmt['date'] = df_vmt['date'].astype('datetime64[ns]')
	df_vmt.rename(columns = {'vmt':'VMT'}, inplace = True) 
	df_vmt['VMT'] = df_vmt['VMT'].astype(int)

	df_vmt.set_index('date')
	df_vmt['dow'] = df_vmt['date'].dt.weekday
	vmt_baseline = pd.DataFrame(np.array([[0, 15395589], [1, 15633580], [2, 15831616], [3, 16128306], [4, 16857911], [5, 14527603], [6, 12425555]]),
                   columns=['dow', 'vmt_base'])
	df_vmt = pd.merge(df_vmt, vmt_baseline, on='dow')
	df_vmt['change'] = (df_vmt['VMT'] - df_vmt['vmt_base'])/df_vmt['vmt_base']
	df_vmt = df_vmt.sort_values(by='date')
	df_vmt.set_index('date')
	#print(list(df_pv.columns.values))
	# dataFields = st.multiselect('Data fields', ['VMT'] , default = ['VMT'] )

	# data = df_vmt[['date'] + dataFields]

	start_date = df_vmt.date.iloc[0].to_pydatetime() - pd.DateOffset(days=1)
	end_date = df_vmt.date.iloc[-1].to_pydatetime() + pd.DateOffset(days=1)

	# st.write(type(start_date))
	data = df_vmt[['date', 'VMT', 'change']]
	dataFields = ['VMT']
	lw = 1  # line width
	maximun_vol = 0

	fig = make_subplots(specs=[[{"secondary_y": True}]])

	# Add traces for axis-1
	fig.add_trace(go.Scatter(x=data['date'], y=data['VMT'],
							 mode='lines+markers', line=dict(dash='solid', width=lw, color='blue'), marker=dict(size=4),
							 name='VMT',
							 legendgroup='group1'),
					secondary_y=True)

	# Add traces for axis-2
	fig.add_trace(go.Scatter(x=data['date'], y=data['change'],
							 mode='lines+markers', line=dict(dash='solid', width=lw, color='green'), marker=dict(size=4),
							 name='VMT Changes w.r.t. Baseline (Jan & Feb 2020)',
							 legendgroup='group2'),
					secondary_y=False)

	# Add shapes
	fig.add_shape(
	        # Line Vertical
	        dict(
	            type="line",
	            x0=start_date,
	            y0=0,
	            x1=end_date,
	            y1=0,
	            line=dict(
	                color="gray",
	                width=1,

	            )
	))

	# Add shapes
	# fig.add_shape(
	#         # Line Vertical
	#         dict(
	#             type="line",
	#             x0=start_date,
	#             y0=-1,
	#             x1=start_date,
	#             y1=1,
	#             line=dict(
	#                 color="pink",
	#                 width=1,
	#                 dash="dashdot",
	#             )
	# ))

	fig.add_annotation(dict(font=dict(color="green",size=12),
                            #x=x_loc,
                            x=start_date + pd.DateOffset(days=5),
                            y=-0.1,
                            showarrow=False,
                            text='<b>baseline</b>',
                            textangle=0,
                           ))

	# fig.add_annotation(
	#             x=start_date + pd.DateOffset(days=5),
	#             y=-0.5,
	#             text="dict Text",
	#             arrowhead=1,
	#             ax=0,
	#             ay=40)

	# fig.add_annotation(dict(
	#             x=start_date + pd.DateOffset(days=5),
 #                y=-0.1,
	#             showarrow=True,
	#             arrowhead=1,
	#             ax=0,
	#             ay=40
	# ))

	fig.update_traces(textposition='top center')
	# Set x-axis title
	fig.update_xaxes(title_text="Date")
	# Set y-axes titles
	fig.update_yaxes(title_text="VMT Changes w.r.t. Baseline", 
					range=[-0.7, 0.7], 
					showline=True, 
					linecolor='rgb(204, 204, 204)', 
					linewidth=2, 
					showticklabels=True, 
					ticks='outside', 
					tickformat=',.0%',
					secondary_y=False)
	fig.update_yaxes(title_text="VMT",
						range=[0, 20000000], 
						showline=True, 
						linecolor='rgb(204, 204, 204)', 
						linewidth=2, 
						showticklabels=True, 
						ticks='outside', 
						secondary_y=True)
	fig.update_layout(xaxis=dict(
					    showline=True,
					    showgrid=False,
					    showticklabels=True,
					    linecolor='rgb(204, 204, 204)',
					    linewidth=2,
					    ticks='outside',
					    tickfont=dict(
					        family='Arial',
					        size=12,
					        color='rgb(82, 82, 82)',
					    ),
					),
					legend=dict(x= 0, y=1, orientation="h"),
				  	margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=4), 
				  	width = 700, 
				  	height = 450,
				  	plot_bgcolor='white')
	st.plotly_chart(fig)


	# Create traces
	# fig = go.Figure()
	# for item in dataFields:
	# 	maximun_vol = max(data[item].max(), maximun_vol)
	# 	fig.add_trace(go.Scatter(x=data['date'], y=data[item], mode='lines', line=dict(dash='solid', width=lw), name=item))
	# maximun_vol = round(maximun_vol // 5 * 5)+5
	# fig.update_layout(xaxis_title='Date',
	# 				  yaxis=dict(title_text='VMT', range=[0, maximun_vol],
	# 							 showticklabels=True),
	# 				  legend=dict(x=.01, y=0),
	# 				  margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=20), width=700, height=450)
	# st.plotly_chart(fig)

	st.markdown("#### VMT Tablular Data")

	df_vmt = df_vmt[['date', 'VMT', 'change']]
	# rename column headers
	df_vmt.columns = ['Date', 'VMT', 'VMT Change']

	dataFields = st.multiselect('Show Data Type', list(df_vmt.columns.values),
			default=['Date', 'VMT', 'VMT Change'])

	
	df_vmt.reset_index(drop=True, inplace=True) # Resets the index, makes factor a column
	# df_vmt.drop("Factor",axis=1,inplace=True)
	
	# st.table(df_vmt[dataFields])
	st.write(df_vmt[dataFields])
	st.markdown("Download the tabular data as a CSV file:")
	st.markdown(get_table_download_link(df_vmt[dataFields], filename = 'VMT'), unsafe_allow_html=True)


def showOtherMetrics():
	###########
	# Sidebar #
	###########
	# st.sidebar.markdown("## Components")
	# rushHourVolume = st.sidebar.checkbox("Volume at Rush Hours", value = True)
	# rushHourVolume = True
	
	########################
	# main content
	########################
	st.markdown("# Other Traffic Performance Metrics")

	st.markdown('This page shows the variations of traffic metrics, such as Volume per Lane at Rush Hour. '
				'Other traffic metrics might be added in the future. Enjoy! :sunglasses:'
				# ' \n * Vehicle Miles of Travel '
				# ' \n * Volume per Lane at Rush Hour '
				# 'Please customize the rush hours and lanes as you need. '
				)
	st.markdown('Downloadable tablular data is shown at the bottom of this page. ')
	
	#################################################################

	# st.markdown("## Vehicle Miles of Travel (VMT)")
	
	# sdate_2 = st.date_input('Select a start date', value = (datetime.datetime.now() - datetime.timedelta(days=90)))
	# edate_2 = st.date_input('Select an end date', value = datetime.datetime.now().date())
	
	# # Check dates ranges and show warnings
	# sdate_2, edate_2, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate_2, edate_2)
	# showDatesWarnings(out_of_range, dates_reversed, dates_equal)

	# st.write('**Selected Dates** from ',sdate_2, ' to ', edate_2,':')

	# df_vmt = getVMT(sdate_2, edate_2)
	# df_vmt['date'] = df_vmt['date'].astype('datetime64[ns]')
	# df_vmt.rename(columns = {'vmt':'VMT'}, inplace = True) 
	# df_vmt['VMT'] = df_vmt['VMT'].astype(int)

	# df_vmt.set_index('date')
	# #print(list(df_pv.columns.values))
	# # dataFields = st.multiselect('Data fields', ['VMT'] , default = ['VMT'] )

	# # data = df_vmt[['date'] + dataFields]
	# data = df_vmt
	# dataFields = ['VMT']
	# lw = 1  # line width
	# maximun_vol = 0

	# # Create traces
	# fig = go.Figure()
	# for item in dataFields:
	# 	maximun_vol = max(data[item].max(), maximun_vol)
	# 	fig.add_trace(go.Scatter(x=data['date'], y=data[item], mode='lines', line=dict(dash='solid', width=lw), name=item))
	# maximun_vol = round(maximun_vol // 5 * 5)+5
	# fig.update_layout(xaxis_title='Date',
	# 				  yaxis=dict(title_text='Traffic Volume per Lane', range=[0, maximun_vol],
	# 							 showticklabels=True),
	# 				  legend=dict(x=.01, y=0),
	# 				  margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=20), width=700, height=450)
	# st.plotly_chart(fig)

	# st.markdown("#### VMT Tablular Data")

	# # rename column headers
	# df_vmt.columns = ['Date', 'VMT']

	# dataFields = st.multiselect('Show Data Type', list(df_vmt.columns.values),
	# 		default=['Date', 'VMT'])


	# st.write(df_vmt[dataFields])
	# st.markdown("Download the tabular data as a CSV file:")
	# st.markdown(get_table_download_link(df_vmt[dataFields], filename = 'VMT'), unsafe_allow_html=True)


	#################################################################################################################

	st.markdown("## Volume per Lane at Rush Hour per Minute")

	st.markdown("Tips: Morning rush hours: 6:00AM-9:00AM; Evening rush hours: 3:00PM-6:00PM")


	sdate = st.date_input('Select a start date:', value = (datetime.datetime.now() - datetime.timedelta(days=90)))
	edate = st.date_input('Select an end date:', value = datetime.datetime.now().date())
	# Check dates ranges and show warnings
	sdate, edate, out_of_range, dates_reversed, dates_equal = checkDatesRange(sdate, edate)
	showDatesWarnings(out_of_range, dates_reversed, dates_equal)

	st.write('**Selected Dates** from ',sdate, ' to ', edate,':')



	df_mpv = getMorningPeakVolume(sdate, edate)
	df_mpv['date'] = df_mpv['date'].astype('datetime64[ns]')
	df_mpv.rename(columns = {'avg_vol_gp':'Morning_GP', 'avg_vol_hov':'Morning_HOV'}, inplace = True) 

	df_epv = getEveningPeakVolume(sdate, edate)
	df_epv['date'] = df_epv['date'].astype('datetime64[ns]')
	df_epv.rename(columns = {'avg_vol_gp':'Evening_GP', 'avg_vol_hov':'Evening_HOV'}, inplace = True) 

	df_pv = pd.merge(df_mpv, df_epv, on='date')
	df_pv.set_index('date')
	#print(list(df_pv.columns.values))
	dataFields = st.multiselect('Data fields', ['Morning_GP', 'Evening_GP', 'Morning_HOV', 'Evening_HOV'] , default = ['Morning_GP', 'Evening_GP'] )

	data = df_pv[['date'] + dataFields]
	lw = 1  # line width
	maximun_vol = 0

	# Create traces
	fig = go.Figure()
	for item in dataFields:
		maximun_vol = max(data[item].max(), maximun_vol)
		fig.add_trace(go.Scatter(x=data['date'], y=data[item], mode='lines', line=dict(dash='solid', width=lw), name=item))
	maximun_vol = round(maximun_vol // 5 * 5)+5
	fig.update_layout(xaxis_title='Date',
					  yaxis=dict(title_text='Traffic Volume per Lane per Minute', range=[0, maximun_vol],
								 showticklabels=True),
					  legend=dict(x=.01, y=0),
					  margin=go.layout.Margin(l=50, r=0, b=50, t=10, pad=20), width=700, height=450)
	st.plotly_chart(fig)

	st.markdown("#### Rush Hour Traffic Volume Tablular Data")

	
	# rename column headers
	df_pv.columns = ['Time', 'Morning_GP', 'Evening_GP', 'Morning_HOV', 'Evening_HOV']

	# set precision of each column
	df_pv[['Morning_GP', 'Evening_GP', 'Morning_HOV', 'Evening_HOV']] \
		= df_pv[['Morning_GP', 'Evening_GP', 'Morning_HOV', 'Evening_HOV']].applymap("{0:.0f}".format)



	dataFields = st.multiselect('Show Data Type', list(df_pv.columns.values),
			default=['Time', 'Morning_GP', 'Evening_GP', 'Morning_HOV', 'Evening_HOV'])


	st.write(df_pv[dataFields])
	st.markdown("Download the tabular data as a CSV file:")
	st.markdown(get_table_download_link(df_pv[dataFields], filename = 'Volume'), unsafe_allow_html=True)


	

def showAbout():
	#################################################################
	st.markdown("# Traffic Performance Score")
	st.markdown("Traffic Performance Score (TPS) can intuitively indicate the overall performance of urban traffic networks. "
				"In this website, the TPS is calculated and visualized to quantify the overall traffic condition in the Greater Seattle area. "
				"With this website, you can view "
				"\n * Temporal dynamic of network-wide TPS of different types of lanes with various time resolutions, ranging from 5 minutes to one day."
				"\n * Varying Spatial distribution of segment-based TPS on interactive maps. "
				"\n * Traffic changes in response to COVID-19 reflected by the TPS. "
				"\n * Other traffic performance metrics. " )
	st.markdown("## Traffic Performance Score Calculation")
	st.markdown("The raw data contains lane-wise speed, volume, and occupancy information collected by each loop detector. "
				"Each detector's meta data includes detector category, route, milepost, director, direction, address. "
				"Based on the consecutive detectors' location information, we separate the freeways into segments, "
				"each of which only contains one loop detector per lane. We consider a road segment's length is the corresponding detector's covered length. "
				"The time interval of the data is one-minute. ")
	st.markdown("The **Traffic Performance Score** (**TPS**) at time $t$ is calculated using the following equation:")
	st.latex(r'''
		\text{TPS}_t = \frac{\displaystyle\sum_{i=1}^n V_t^i \cdot Q_t^i \cdot L^i }{ V_f \cdot \displaystyle \sum_{i=1}^n Q_t^i \cdot L^i } * 100\%
		''')
	st.markdown("where $V_t^i$ and $Q_t^i$ represent the *speed* and *volume* "
				"of each road segment $i$ at time $t$. $L^i$ is the length of $i$-th detector's covered road segment."
				"$V_f$ is the free flow speed. "
				"The unit of speed is mile per hour (mph), and the unit of covered distance is mile.")
	st.markdown("In this way, the **TPS** is a value ranges from 0% to 100%. "
			#	"($TPS \in [0, 100]$). "
				"The closer to 100% the **TPS** is, the better the overall network-wide traffic condition is. ")

	#################################################################
	st.markdown("## Data Source")
	st.markdown("The TPS is calculated based on data collected from more than 44800 inductive loop detectors deployed on freeways in Seattle, WA area. "
				"Freeways include: I-5, I-90, I-99, I-167, I-405, and SR-520. The raw data comes from Washington State Department of Transportation (WSDOT). "
				"Representative detectors are shown in the following map. ")
	showLoopDetectorMap()

	#################################################################
	st.markdown("## Development Credit")
	st.markdown("This website is developed by the Artificial Intelligence group in the Smart Transportation Application and Research Lab ([STAR Lab](http://www.uwstarlab.org/)). ")
	st.markdown("Group member: Zhiyong Cui, Meixin Zhu, Pengfei Wang, Yang Zhou, Qianxia Cao, Shuo Wang, Cole Kopca, and John Ash.")

	# st.write('<script type="text/javascript" id="clstr_globe" src="//cdn.clustrmaps.com/globe.js?d=q_qd3mQ6FdC52ZJRUtern-mmVaK1RER3n2BPh-FTy-Y"></script>', unsafe_allow_html=True)
	# st.write('<a href="https://clustrmaps.com/site/1b7ap" title="Visit tracker"><img src="//clustrmaps.com/map_v2.png?cl=080808&w=a&t=t&d=jn07mPkuDBD9jMBfRsCUgcfZN5e7Z2SydqZ3ItFsfv4&co=ffffff&ct=808080" /></a>', unsafe_allow_html=True)


def showLoopDetectorMap():	
	df_loop_location = getLoopDetectorLocation()

	st.pydeck_chart(pdk.Deck(
		# map_style='mapbox://styles/mapbox/dark-v9',
		map_style='mapbox://styles/mapbox/light-v9',
			initial_view_state=pdk.ViewState(
			latitude=47.59,
			longitude=-122.24,
			zoom=9,
			pitch=50,
		),
		layers=[
			# pdk.Layer(
			#    'HexagonLayer',
			#    data=df_loop_location,
			#    get_position='[lon, lat]',
			#    radius=200,
			#    elevation_scale=4,
			#    elevation_range=[0, 1000],
			#    pickable=True,
			#    extruded=True,
			# ),
			pdk.Layer(
				'ScatterplotLayer',
				data=df_loop_location,
				get_position='[lon, lat]',
				get_color='[0, 160, 187, 160]',
				get_radius=300,
			),
		],
 	))






#####################################################
# Main
#####################################################
def main():
	# hide upper-right menu
	hide_menu_style = """
        <style>
        #MainMenu {visibility: hidden;}
        </style>
        """
	st.markdown(hide_menu_style, unsafe_allow_html=True)

	# st.title("Traffic Performance Score in Seattle Area")

	st.sidebar.title("Traffic Performance Score (TPS)")
	app_mode = st.sidebar.radio("Navigation",
	        ["Home", "About this Website & TPS", "Network-based TPS", "Segment-based TPS", \
	        "Impact of COVID-19", "Vehicle Miles of Travel", "Other Traffic Metrics"])
	# st.sidebar.markdown("[![this is an image link](./images/STARLab.png)](https://streamlit.io)")
	if  app_mode == "Home":
		IntroduceTrafficIndex()
	elif app_mode == "About this Website & TPS":
		showAbout()
	elif app_mode == "Network-based TPS":
		showTrafficIndex()
	elif app_mode == "Segment-based TPS":
		showSgementTPS()
	elif app_mode == "Impact of COVID-19":
		showCOVID19()
	elif app_mode == "Vehicle Miles of Travel":
		showVMT()
	elif app_mode == "Other Traffic Metrics":
		showOtherMetrics()
	

	st.write('<a href="https://clustrmaps.com/site/1b7ap" title="Visit tracker"><img src="//clustrmaps.com/map_v2.png?cl=ffffff&w=70&t=n&d=jn07mPkuDBD9jMBfRsCUgcfZN5e7Z2SydqZ3ItFsfv4&co=ffffff&ct=ffffff" style="display:none"/></a>', unsafe_allow_html=True)


	# st.sidebar.title("About")
	# st.sidebar.info(
 #        "This an open source project developed and maintained by the "
 #        "*Artificial Intelligence GROUP* in the [Smart Transportation Application and Research Lab (**STAR Lab**)](http://www.uwstarlab.org/) "
 #        "at the [University of Washington](https://www.washington.edu/). "
	        
	# )

	st.sidebar.title("Contact Us")
	st.sidebar.info(
		"This an open source project developed and maintained by the "
        "*Artificial Intelligence GROUP* in the [Smart Transportation Application and Research Lab (**STAR Lab**)](http://www.uwstarlab.org/) "
        "at the [University of Washington](https://www.washington.edu/). "
		"For more information, "
		"please contact "
        "[Prof. Yinhai Wang](https://www.ce.washington.edu/facultyfinder/yinhai-wang) "
        "([yinhai@uw.edu](mailto:yinhai@uw.edu))."
	)

	image_starlab = Image.open('images/STARLab.png')
	st.sidebar.image(image_starlab, width = 240, use_column_width=False)

	image_pactrans = Image.open('images/PacTrans.png')
	st.sidebar.image(image_pactrans, width = 80, use_column_width=False)

	
	
	# image_uw = Image.open('images/UW2.png')
	# st.sidebar.image(image_uw, width = 300, use_column_width=False)

if __name__ == '__main__':
	main()

	# code = read()
	# edit = st.checkbox("Edit")
	# ph = st.empty()
	# code = ph.text_area("Your code here 👇", code)
	# if not edit:
	#     ph.empty()
	#     save(code)
	#     st.code(code, "c++")