# -*- coding: utf-8 -*-

"""

Created on Thu Feb 15 16:30:00 2018

@author: monakhov

"""

import numpy as np

import matplotlib.pyplot as plt

from savitzky_golay import savitzky_golay

arr_test_subjects = ['Lasse', 'Ramin', 'Jani', 'Antti', 'xingyang ni','Nahid', 'Vida', 'Ling yu',

'Jari', 'Pouria','Sujeet','emre', 'Adriana', 'kashyap', 'You yu', 'Peter',

'Kimmo', 'Yuri', 'Kalle', 'Toni']

# =============================================================================

# clip='armor'

# cuts = [8, 26.026, 38.672, 50.684]

# clip_length_sec = 66

# =============================================================================

# =============================================================================

# clip='martial'

# cuts = [12.56, 21.76, 44.76, 54.76]

# clip_length_sec = 62

# arr_test_subjects = ['Lasse', 'Ramin', 'Jani', 'Antti', 'xingyang ni','Nahid', 'Vida', 'Ling yu',

# 'Jari', 'Sujeet', 'Adriana', 'kashyap', 'You yu', 'Peter',

# 'Kimmo', 'Yuri', 'Kalle', 'Toni']

# =============================================================================

clip='lion'

cuts = [11.011, 32.866, 55.122, 67.034]

clip_length_sec = 76

cuts_small = np.array([int(x*1000/50) for x in cuts])

res = {}

#%%

seg_length = 250

w = seg_length/1000

folder_name = 'yaw_diff_max/'+str(seg_length)+'/'

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

seconds_arr = np.arange(0,clip_length_sec,w)*1000

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

seconds_cuts = []

for cut in cuts:

seconds_cuts.append(np.argmin(np.abs(seconds_arr-cut*1000)))#array of indeces for cuts

#print(second)

tmp_indexes = []

jjj=0

for (ind,ttt) in enumerate(data['Vida'][clip]['time']):

if ttt/1000>seconds_cuts[jjj]:

tmp_indexes.append(ind)

jjj += 1

if jjj >= len(seconds_cuts):

break

#%%

#seconds_cuts = [19,39,59,79,99,119,139]

res = np.zeros((len(arr_test_subjects),len(seconds_arr)))

for cur_iter,user in enumerate(arr_test_subjects):

tmp_clip = []

tmp_arr = []

current_sec_barrier = 0

i = 0

#current_trajectory = savitzky_golay(np.array(data[user][clip]['yaw']), window_size=11, order=3,deriv=0)

current_trajectory = np.array(data[user][clip]['yaw'])

while (i < len(data[user][clip]['time']) and current_sec_barrier < len(seconds_arr) ):

if (data[user][clip]['time'][i] >= seconds_arr[current_sec_barrier] and tmp_clip != []):

# abs_diff = np.abs([s-tmp_clip[0] for s in tmp_clip])

# shortest_abs_diff = np.minimum(abs_diff,360-abs_diff)

# shortest_abs_diff = abs_diff

shortest_abs_diff = np.max(tmp_clip) - np.min(tmp_clip)

#shortest_abs_diff = tmp_clip[-1] - tmp_clip[0]

tmp_arr.append(np.max(shortest_abs_diff))

current_sec_barrier += 1

tmp_clip = []

# if seconds_arr[current_sec_barrier] >= 159*1000:

# break

tmp_clip.append(current_trajectory[i])

i += 1

#if (len(tmp_clip) != 0):

# abs_diff = np.abs([s-tmp_clip[0] for s in tmp_clip])

# shortest_abs_diff = np.minimum(abs_diff,360-abs_diff)

# tmp_arr.append(np.max(shortest_abs_diff))

# tmp_clip = []

#tmp_arr = tmp_arr/np.max(tmp_arr)

res[cur_iter,:] = tmp_arr

#%%

#folder_name='range/martial/'

#plt.ioff()

for cur_iter,user in enumerate(arr_test_subjects):

plt.figure()

plt.bar(seconds_arr/1000,res[cur_iter,:],width=(-1)*w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,res[cur_iter,seconds_cuts],width=(-1)*w,align='edge')

plt.title('exploration range: '+user)

plt.xlabel('time, sec')

plt.ylabel('exploration range, degrees')

#plt.ylim( (0, 100) )

# plt.savefig(folder_name+user+'.png',dpi=200)

# plt.close()

#plt.ion()

#%%

#plt.ioff()

plt.figure()

meanRange = np.mean(res,axis=0)

plt.bar(seconds_arr/1000,meanRange,width=(-1)*w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,meanRange[seconds_cuts],width=(-1)*w,align='edge', label='scene cut')

#plt.title('Average exploration range over users:')

plt.xlabel('Time (s)', fontsize=16)

plt.ylabel('Exploration range (degrees)', fontsize=16)

plt.legend()

plt.ylim( (0, 20) )

#plt.savefig(folder_name+'1average.png',dpi=200)

#plt.close()

#plt.ion()

#plt.savefig('SRM_Lions.jpg',dpi=600)

#%%

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

#Velocity values

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

#clip='helsinki_na'

#cuts = []

#clip_length_sec = data['Vida'][clip]['time'][-1]/1000

seg_length = 250

w = seg_length/1000

folder_name = 'velocity/'+str(seg_length)+'/'

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

seconds_arr = np.arange(0,clip_length_sec,w)*1000

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

seconds_cuts = []

for cut in cuts:

seconds_cuts.append(np.argmin(np.abs(seconds_arr-cut*1000)))

#print(second)

#%%

res2 = np.zeros((len(arr_test_subjects),len(seconds_arr)))

tmp_velocities = []

for cur_iter,user in enumerate(arr_test_subjects):

tmp_clip = []

tmp_arr = []

current_sec_barrier = 0

i = 0

current_trajectory = savitzky_golay(np.array(data[user][clip]['yaw']), window_size=11, order=3,deriv=0)

velocity = np.gradient(current_trajectory, data[user][clip]['time'])*1000

tmp_velocities.append(velocity)

while (i < len(data[user][clip]['time']) and current_sec_barrier < len(seconds_arr) ):

if (data[user][clip]['time'][i] >= seconds_arr[current_sec_barrier] and tmp_clip != []):

tmp_arr.append(np.mean(np.abs(tmp_clip)))

current_sec_barrier += 1

tmp_clip = []

# if current_sec_barrier >= 159:

# break

tmp_clip.append(velocity[i])

if current_sec_barrier == 0:

tmp_clip[-1] = 0

i += 1

#if (len(tmp_clip) != 0):

# abs_diff = np.abs([s-tmp_clip[0] for s in tmp_clip])

# shortest_abs_diff = np.minimum(abs_diff,360-abs_diff)

# tmp_arr.append(np.max(shortest_abs_diff))

# tmp_clip = []

res2[cur_iter,:] = tmp_arr

#%%

folder_name='velocity/martial/'

plt.ioff()

for cur_iter,user in enumerate(arr_test_subjects):

plt.figure()

plt.bar(seconds_arr/1000,res2[cur_iter,:],width=-w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,res2[cur_iter,seconds_cuts],width=-w,align='edge')

plt.title('Angular speed: '+user)

plt.xlabel('time, sec')

plt.ylabel('angular speed, degrees/sec')

plt.ylim( (0, 250) )

plt.savefig(folder_name+user+'.png',dpi=200)

plt.close()

plt.ion()

#%%

#plt.ioff()

meanRange = np.mean(res2,axis=0)

plt.bar(seconds_arr/1000,meanRange,width=-w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,meanRange[seconds_cuts],width=-w,align='edge', label='scene cut')

#plt.title('Angular speed average: Lions')

#plt.xlabel('time, sec')

#plt.ylabel('angular speed, degrees/sec')

plt.ylim( (0, 70) )

plt.xlabel('Time (s)', fontsize=16)

plt.legend()

plt.ylabel('Angular speed (degrees per sec)', fontsize=16)

#plt.savefig(folder_name+'1average'+'.png',dpi=200)

#plt.close()

#plt.ion()

#%%

tmp_velocities = np.abs(np.array(tmp_velocities))

plt.plot(np.arange(tmp_velocities.shape[1])*50/1000,np.mean(tmp_velocities,axis=0))

plt.vlines(cuts_small*50/1000,ymin=0,ymax=60,color='orange', label='scene cut')

plt.xlabel('time, sec',fontsize=14)

plt.ylabel('angular speed, degrees/sec',fontsize=14)

plt.legend(fontsize=14)

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

#Acceleration values

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

#%%

seg_length = 250

w = seg_length/1000

folder_name = 'acceleration/'+str(seg_length)+'_mean/'

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

seconds_arr = np.arange(0,clip_length_sec,w)*1000

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

seconds_cuts = []

for cut in cuts:

seconds_cuts.append(np.argmin(np.abs(seconds_arr-cut*1000)))

#print(second)

#%%

res3 = np.zeros((len(arr_test_subjects),len(seconds_arr)))

for cur_iter,user in enumerate(arr_test_subjects):

tmp_clip = []

tmp_arr = []

current_sec_barrier = 0

i = 0

current_trajectory = savitzky_golay(np.array(data[user][clip]['yaw']), window_size=11, order=3,deriv=0)

velocity = np.gradient(current_trajectory, data[user][clip]['time'])*1000

acceleration = np.gradient(velocity, data[user][clip]['time'])*1000

while (i < len(data[user][clip]['time']) and current_sec_barrier < len(seconds_arr) ):

if (data[user][clip]['time'][i] >= seconds_arr[current_sec_barrier] and tmp_clip != []):

tmp_arr.append(np.median(np.abs(tmp_clip)))

# upper_quartile = np.percentile(np.abs(tmp_clip), 95,interpolation='nearest')

# lower_quartile = np.percentile(np.abs(tmp_clip), 5,interpolation='nearest')

# tmp_arr.append(upper_quartile)

# tmp_arr.append(lower_quartile)

current_sec_barrier += 1

tmp_clip = []

# if current_sec_barrier >= 159:

# break

tmp_clip.append(acceleration[i])

if current_sec_barrier == 0:

tmp_clip[-1] = 0

i += 1

res3[cur_iter,:] = tmp_arr#/np.max(tmp_arr)

#%%

folder_name='acceleration/martial/'

plt.ioff()

for cur_iter,user in enumerate(arr_test_subjects):

plt.bar(seconds_arr/1000,res3[cur_iter,:],width=-w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,res3[cur_iter,seconds_cuts],width=-w,align='edge')

plt.title('Angular acceleration: '+user)

plt.xlabel('time, sec')

plt.ylabel('angular acceleration, degrees/sec^2')

plt.ylim( (0, np.max(res3) ))

plt.savefig(folder_name+user+'.png',dpi=200)

plt.close()

plt.ion()

#%%

#plt.ioff()

meanRange = np.mean(res3,axis=0)

plt.bar(seconds_arr/1000,meanRange,width=-w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,meanRange[seconds_cuts],width=-w,align='edge', label='scene cut')

#plt.title('Angular acceleration mean: Martial training')

#plt.xlabel('time, sec')

#plt.ylabel('angular acceleration, degrees/sec^2')

plt.ylim( (0, 280) )

plt.xlabel('Time (s)', fontsize=16)

plt.ylabel('Angular acceleration (degrees per $sec^2$)', fontsize=16)

plt.legend()

#plt.savefig(folder_name+'1average'+'.png',dpi=200)

#plt.close()

#plt.ion()

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

#computing number of head turns

#%%

seg_length = 2000

folder_name = 'num_of_turns/1000/'

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

seconds_arr = np.arange(2,160,2)*1000

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

seconds_cuts = []

for cur_iter,second in enumerate(seconds_arr):

if second%20000 == 0:

seconds_cuts.append(cur_iter)

#%%

#seconds_cuts = [19,39,59,79,99,119,139]

res4 = np.zeros((len(arr_test_subjects),len(seconds_arr)))

for cur_iter,user in enumerate(arr_test_subjects):

tmp_clip = []

tmp_arr = []

current_sec_barrier = 0

i = 0

turns_buffer = [0,0,0,0,0,0,0,0,0,0]

prev_value = 0

num_of_turns = 0

while (i < len(data[user][clip]['time']) and current_sec_barrier < len(seconds_arr) ):

if (data[user][clip]['time'][i] >= seconds_arr[current_sec_barrier]):

tmp_arr.append(num_of_turns)

current_sec_barrier += 1

num_of_turns = 0

# if seconds_arr[current_sec_barrier] >= 159*1000:

# break

if data[user][clip]['yaw'][i] > prev_value:

turns_buffer.pop(0)

turns_buffer.append(1)

else:

turns_buffer.pop(0)

turns_buffer.append(-1)

prev_value = data[user][clip]['yaw'][i]

i += 1

if (sum(turns_buffer[:5])== 5 and sum(turns_buffer[5:]) == -5) or (sum(turns_buffer[:5])== -5 and sum(turns_buffer[5:]) == 5):

num_of_turns+=1

res4[cur_iter,:] = tmp_arr

#%%

for cur_iter,user in enumerate(arr_test_subjects):

plt.figure()

plt.bar(seconds_arr/1000,res4[cur_iter,:],width=-1,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,res4[cur_iter,seconds_cuts],width=-1,align='edge')

plt.title('number of turns: '+user)

# plt.ylim( (0, 600) )

#%%

meanRange = np.median(res4,axis=0)

plt.bar(seconds_arr/1000,meanRange,width=-1,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,meanRange[seconds_cuts],width=-1,align='edge')

plt.title('mean number of turns')

#%%

#%%

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

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

#correlation

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

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

seg_length = 500

w = seg_length/1000

folder_name = 'correlation/'+str(seg_length)+'/'

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

seconds_arr = np.arange(0,160,w)*1000

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

seconds_cuts = []

for cur_iter,second in enumerate(seconds_arr):

if second%20000 == 0:

seconds_cuts.append(cur_iter)

#%%

#seconds_cuts = [19,39,59,79,99,119,139]

res = np.zeros((len(arr_test_subjects),len(seconds_arr)))

for cur_iter,user in enumerate(arr_test_subjects):

tmp_clip = []

tmp_arr = []

current_sec_barrier = 0

i = 0

current_trajectory = savitzky_golay(np.array(data[user][clip]['yaw']), window_size=11, order=3,deriv=0)

while (i < len(data[user][clip]['time']) and current_sec_barrier < len(seconds_arr) ):

if (data[user][clip]['time'][i] >= seconds_arr[current_sec_barrier] and tmp_clip != []):

# abs_diff = np.abs([s-tmp_clip[0] for s in tmp_clip])

# shortest_abs_diff = np.minimum(abs_diff,360-abs_diff)

# shortest_abs_diff = abs_diff

shortest_abs_diff = np.max(tmp_clip) - np.min(tmp_clip)

tmp_arr.append(np.max(shortest_abs_diff))

current_sec_barrier += 1

tmp_clip = []

# if seconds_arr[current_sec_barrier] >= 159*1000:

# break

tmp_clip.append(current_trajectory[i])

i += 1

#if (len(tmp_clip) != 0):

# abs_diff = np.abs([s-tmp_clip[0] for s in tmp_clip])

# shortest_abs_diff = np.minimum(abs_diff,360-abs_diff)

# tmp_arr.append(np.max(shortest_abs_diff))

# tmp_clip = []

res[cur_iter,:] = tmp_arr

#%%

plt.ioff()

for cur_iter,user in enumerate(arr_test_subjects):

plt.figure()

tmp = res[cur_iter,:]

tmp = tmp - np.mean(tmp)

norm = np.sum(tmp ** 2)

cur_corr = (np.correlate(tmp, tmp, mode='full')/norm)[len(tmp)-1:]

plt.bar(seconds_arr/1000,cur_corr,width=(-1)*w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,cur_corr[seconds_cuts],width=(-1)*w,align='edge')

plt.title('correlation: '+user)

plt.xlabel('shift, sec')

plt.ylabel('correlation value')

plt.savefig(folder_name+user+'.png',dpi=200)

plt.close()

plt.ion()

#%%

plt.ioff()

#plt.figure()

meanRange = np.mean(res,axis=0)

tmp = meanRange - np.mean(meanRange)

norm = np.sum(tmp ** 2)

cur_corr = (np.correlate(tmp, tmp, mode='full')/norm)[len(tmp)-1:]

plt.bar(seconds_arr/1000,cur_corr,width=(-1)*w,align='edge')

plt.bar(seconds_arr[seconds_cuts]/1000,cur_corr[seconds_cuts],width=(-1)*w,align='edge')

plt.title('average correlation')

plt.xlabel('shift, sec')

plt.ylabel('correlation value')

plt.savefig(folder_name+'1average.png',dpi=200)

plt.close()

plt.ion()

#%%