# -*- coding: utf-8 -*-
"""
Created on Tue Jun 25 11:20:27 2019

@author: mor96805

Door events analysis

Show number of events per day

plot all events for time of day

UPDATED Thur Nov 14 2019
"""

designated_csv = 'C:\\Users\\PATH TO YOUR CSV HERE\\door_tracker.csv'

import pandas as pd
pd.set_option('display.max_columns', 100)
pd.set_option('display.max_rows', 500)
import matplotlib.pyplot as plt
plt.rcParams.update({'font.size': 8})
import numpy as np
import datetime

df = pd.read_csv(designated_csv)

# Totals
total_days = df.Date.drop_duplicates().size
print("\nTotal days recorded: " + str(total_days))
total_events = df[['Door(YN)', 'Buzzer(YN)']].sum().sum()
print("Total events: " + str(total_events))
print("Average per day: " + str(round(total_events/total_days, 2)))
stamp = datetime.datetime.now()
date_stamp = str(stamp.strftime("%d/%m/%Y"))
if date_stamp in df.Date.tolist():
    print("No. events today: " + str(df.loc[df.Date == date_stamp][['Buzzer(YN)', 'Door(YN)']].sum().sum()))
else:
    print("No Rory events today.")

# Get Rory % but then proceed for all values as either Rory/non Rory
total_rory_events = df.loc[df['Non-RoryEvent(YN)']==0][['Buzzer(YN)', 'Door(YN)']].sum().sum()
print("Total Rory Events : " + str(total_rory_events))
print("Rory % of total: " + str(round(100*total_rory_events/total_events,2)) + " %")
deliveries = df.loc[df['Delivery(YN)']==1][['Door(YN)', 'Buzzer(YN)']].sum().sum()
print("% of total events from deliveries: " + str(round(100*deliveries/total_events,2))+" %")
glasgow_staff = df.loc[df['BVGlasgowEmployee(YN)']==1][['Door(YN)', 'Buzzer(YN)']].sum().sum()
print("% of total events from Glasgow office staff: " + str(round(100*glasgow_staff/total_events,2))+" %")


# 'Date', 'Time', 'weekday', 'Buzzer(YN)', 'Door(YN)', 'Delivery(YN)',
#       'BVGlasgowEmployee(YN)', 'Non-RoryEvent(YN)'

date_table = pd.pivot_table(df, index='Date', values=['Door(YN)', 'Buzzer(YN)'], aggfunc=np.sum)
date_table['Sum'] = date_table['Buzzer(YN)'] + date_table['Door(YN)']
date_table = date_table.sort_values(by=['Sum'])

print("-"*50)
print("\nMinimum events in a day: ")
print(date_table.head(10))
print("\nNumber of no-event days: " + str(sum(date_table.Sum == 0)))
print("\nMaximum events in a day: ")
print(date_table.tail(10))

# =============================================================================
# Graphs
# =============================================================================

#plt.figure(figsize=(12,8))
#plt.rcParams.update({'font.size': 16})
#plt.grid(linestyle="--")
#plt.hist(date_table.Sum, edgecolor='black', linewidth=1.0, bins=max(date_table.Sum))
#plt.title("Total number of events histogram")
#plt.ylabel("Days")
#plt.xlabel("Events")
#plt.savefig('Hist.jpg')

times = df.Time


fig, ax1 = plt.subplots(figsize=(8,3), constrained_layout=True)
ax1.grid(linestyle="--")
ax1.hist(date_table.Sum, edgecolor='black', linewidth=1.0, bins=max(date_table.Sum))
ax1.set_xlabel("Events")
ax1.set_ylabel("Days")
ax2 = ax1.twinx()
ax2.set_ylabel("Prob", color='red')
ax2.tick_params(axis='y', labelcolor='red')
ax2.hist(date_table.Sum, bins=max(date_table.Sum), density=True, histtype='step', cumulative=1, color='red')
plt.title("Total number of events histogram")
#ax1.plot([0.5, 6.5, 12.5], [12, 8, 0])
plt.show()

# =============================================================================
# For the hist of hourly events
# =============================================================================

df['interval'] = ""
for i in range(0, len(df)):
    x = str(df.loc[i, 'Time'])
    y = x[:3]
    if x[3] < "3":
        y += str("00")
    else:
        y+= str("30")
    df.loc[i, 'interval'] = y
    
df['interval'] = ""
for i in range(0, len(df)):
    x = str(df.loc[i, 'Time'])
    y = x[:3]
    if x[3] < "3":
        y += str("00")
    else:
        y+= str("30")
    df.loc[i, 'interval'] = y


time = []
for i in range(7,10):
    time.append("0"+str(i)+":00:00")
    time.append("0"+str(i)+":30:00")
for i in range(10,19):
    time.append(str(i)+":00:00")
    time.append(str(i)+":30:00")

time_table = pd.DataFrame({'interval':time, 'count':[0]*len(time)})

for i in range(0,len(time_table)-1):
    time_table.loc[i, 'count'] = len(df[(df.Time > time_table.loc[i, 'interval']) & (df.Time < time_table.loc[i+1, 'interval'])])
    
# make up some data
x = time_table['interval']
y = time_table['count']

# plot
from matplotlib.pyplot import figure
figure(num=1, figsize=(8, 3), dpi=80, facecolor='w', edgecolor='black')
plt.bar(x, y, edgecolor='black')
plt.gcf().autofmt_xdate()
#plt.yticks(range(0, 1+max(time_table['count'])))
plt.title('Time of day Hist')
plt.xlabel('Time of Day')
plt.ylabel('Total Events')
plt.grid(linestyle="--")
plt.show()

# =============================================================================
# Day of the week count
# =============================================================================


weekday_table = pd.pivot_table(df, index='weekday', values=['Door(YN)', 'Buzzer(YN)'], aggfunc=np.sum)

weekday_count_table = pd.pivot_table(df, index='Date', values=['weekday'])
weekdays = list(range(0,5))
weekdays_count = [np.sum([int(j == i) for j in weekday_count_table['weekday'].tolist()])for i in weekdays]
weekday_av = [(weekday_table['Buzzer(YN)'] + weekday_table['Door(YN)']).tolist()[i]/weekdays_count[i] for i in range(0,5)]


figure(num=2, figsize=(8, 3), dpi=80, facecolor='w', edgecolor='black')
plt.bar(weekdays, weekday_av, edgecolor='black')
#plt.gcf().autofmt_xdate()
#plt.yticks(range(0, 1+max(time_table['count'])))
plt.title('Day of Week Hist')
plt.xlabel('Weekday')
plt.ylabel('Total Events')
plt.grid(linestyle="--")
plt.show()