Motor_tracking/tracking_helper_functions.py at master · lisamnash/Motor_tracking · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
import os
import sys
from settings import tracking_settings
import cPickle as pickle
import numpy as np
import matplotlib.pyplot as plt

def dump_pickled_data(output_dir, filename, data):
    if not os.path.exists(output_dir): os.mkdir(output_dir)
    of = open(output_dir + '/' + filename + '.pickle', 'wb')
    pickle.dump(data, of, pickle.HIGHEST_PROTOCOL)
    of.close()


def find_files_by_extension(root_dir, ext, tot=False):
    filenames = []
    for root, dirs, files in os.walk(root_dir):
        for file in files:
            if file.endswith(ext):
                if tot == False:
                    filenames.append(file)
                else:
                    filenames.append(root + '/' + file)
    return filenames


def set_output_directory():
    if 'output_dir' in tracking_settings:
        output_dir = tracking_settings['output_dir']
    else:
        output_dir = './'

    if not os.path.exists(output_dir):
        os.mkdir(output_dir)

    return output_dir

def shift(array):
    return array - np.mean(array)

def fft_on_data(dat, output_dir):
    '''
    Performs a fourier transform on motor tracked data. colors gyroscopes by speed in an image if maxima of frequencies are over 100

  Parameters
        -----------------
       dat: position vs time for each gyroscope in each frame
       output_dir: string specifying output directory

    '''

    t_dat = (dat.T)[0]

    print t_dat

    min_t = min(t_dat)

    num_gyros = len(t_dat[np.where(t_dat == min_t)])

    x_dat = (dat.T)[1]
    y_dat = (dat.T)[2]

    partitions = 1

    num_time_steps = len(x_dat) / (num_gyros)
    ind = np.array([k * num_gyros for k in xrange(num_time_steps)])
    t_f = np.fft.fftfreq(len(ind), 1)

    tot_power = np.zeros_like(t_f)
    m_f = []
    fft_list = []
    coords = []

    # get the data ready
    output_dir_1 = output_dir + 'fourier/'
    output_dir_old = output_dir
    if not os.path.exists(output_dir_1): os.mkdir(output_dir_1)
    copy_dir = output_dir

    for u in range(1):  # data can be partitioned, but currently I am just putting everything in one partition.

        num_in_partition = np.floor(num_time_steps / partitions)

        for j in range(num_gyros):
            pp = u * num_time_steps
            output_dir = os.path.join(output_dir_1, 'gy_%d/' % j)
            if not os.path.exists(output_dir): os.mkdir(output_dir)

            ind = np.array([k * num_gyros + j for k in range(num_time_steps)])

            x_gy_full = shift(x_dat[ind])
            y_gy_full = shift(y_dat[ind])
            t_gy_full = t_dat[ind]

            x_gy_full = x_gy_full[:num_in_partition * partitions]
            y_gy_full = y_gy_full[:num_in_partition * partitions]
            t_gy_full = t_gy_full[:num_in_partition * partitions]
            for u in range(partitions):

                x_gy = x_gy_full[u * num_in_partition:(u + 1) * num_in_partition]
                y_gy = y_gy_full[u * num_in_partition:(u + 1) * num_in_partition]
                t_gy = t_gy_full[u * num_in_partition:(u + 1) * num_in_partition]

                wind = np.hanning(len(t_gy))

                coords.append([np.mean(x_dat[ind]), np.mean(y_dat[ind])])

                fft_cylindrical_a = np.fft.fft(wind * np.array(x_gy + 1j * y_gy))
                fft_cylindrical_negative = np.fft.fft(wind * np.array(x_gy - 1j * y_gy))
                fft_cylindrical = abs(fft_cylindrical_a) ** 2
                fft_cylindrical_n = abs(fft_cylindrical_negative) ** 2
                fft_x = np.fft.fft(wind * np.array(x_gy))
                fft_y = np.fft.fft(wind * np.array(y_gy))
                fft_freq = np.fft.fftfreq(len(t_gy), t_gy[1] - t_gy[0])

                dump_pickled_data(output_dir, 'x_gy_%01d' % u, x_dat[ind])
                dump_pickled_data(output_dir, 'y_gy_%01d' % u, y_dat[ind])
                dump_pickled_data(output_dir, 'fft_x_%01d' % u, np.array([fft_freq, fft_x]).T)
                dump_pickled_data(output_dir, 'fft_y_%01d' % u, np.array([fft_freq, fft_y]).T)
                dump_pickled_data(output_dir, 'fft_complex_%01d_positive' % u, np.array([fft_freq, fft_cylindrical_a]).T)
                dump_pickled_data(output_dir, 'fft_complex_%01d_negative' % u,
                                  np.array([fft_freq, fft_cylindrical_negative]).T)
                dump_pickled_data(output_dir, 'coords_%01d.pickle' % u, coords[j])
                fft_list.append(fft_cylindrical_a)

                max_xf = max(fft_cylindrical)

                maximum_f = fft_freq[np.where(fft_cylindrical == max_xf)[0][0]]

                if True:  # j ==0:
                    fig = plt.figure()
                    ax = fig.add_subplot(1, 1, 1)
                    plt.plot(np.array(t_gy), x_gy)
                    plt.plot(np.array(t_gy), y_gy)

                    plt.savefig(output_dir + 'data_xy_%01d.png' % u)
                    print output_dir + 'data_xy_%01d.png' % u

                    lab = ['max of x + iy fft = %0.3f ' % maximum_f, '', '']
                    fig3 = plt.figure()
                    ax = fig3.add_subplot(1, 1, 1)
                    plt.plot(fft_freq[1:], abs(fft_cylindrical[1:]), label=lab[0])
                    plt.plot(fft_freq[1:], abs(fft_cylindrical_n[1:]), label=lab[0])
                    plt.xlabel('Freq (Hz)', fontsize=11)
                    plt.ylim(0, max_xf * 1.2)
                    handles, labels = ax.get_legend_handles_labels()
                    ax.legend(handles, labels, fancybox=True, fontsize=11, loc=1)
                    # plt.yscale('log')
                    plt.xlim(0, 3.5)
                    plt.savefig(output_dir + 'fft_cyl_%01d.png' % u)

                    plt.close()

                if max(fft_freq > 100):
                    if abs(maximum_f) < 10:
                        selected_f = fft_freq[fft_freq < -100]
                        selected = fft_cylindrical[fft_freq < -100]

                        max_xf = max(selected)
                        maximum_f = selected_f[np.where(selected == max_xf)[0][0]]

                    if abs(maximum_f) > 150 and abs(maximum_f) < 3200:
                        m_f = list(m_f)
                        m_f.append([j, maximum_f])

                plt.close()

                if j == 0:
                    tot_power = np.zeros_like(fft_cylindrical)
                else:

                    tot_power += fft_cylindrical

        if max(fft_freq) > 100:

            if len(m_f) > 1:
                dump_pickled_data(copy_dir, '/speed', np.array(m_f))
                m_f = np.array(m_f).T[1]
                fig55 = plt.figure()
                ax = fig55.add_subplot(111)
                n, bins, patches = ax.hist(abs(np.array(m_f)), 100, normed=False, facecolor='green', alpha=0.75)
                plt.savefig(copy_dir + '/speed.png')
                m_f = np.array(m_f)
                plt.close()

                print output_dir_old + 'com_data.pickle'
                color_by_speed(output_dir_old + 'com_data.pickle', output_dir_old)