diff options
| -rw-r--r-- | Classifier.py | 4 | ||||
| -rw-r--r-- | click.c | 36 | ||||
| -rwxr-xr-x | record.py | 102 | ||||
| -rwxr-xr-x | visualize.py | 8 |
4 files changed, 87 insertions, 63 deletions
diff --git a/Classifier.py b/Classifier.py index 67e84fb..bbc0403 100644 --- a/Classifier.py +++ b/Classifier.py @@ -42,8 +42,8 @@ def gen_model(train_data, train_ans, verbose=True): tuned_parameters = [ {'kernel': ['rbf'], 'gamma': [], 'C': []}, ] - now = 4.0 - while now > 0.00001: + now = 1.0 + while now > 0.001: tuned_parameters[0]['gamma'].append(now) if now >= 1: tuned_parameters[0]['C'].append(now) @@ -5,7 +5,6 @@ #include <unistd.h> #include <xdo.h> -/* int main () { xdo_t* xdo = xdo_new (NULL); if (xdo == NULL) { @@ -24,22 +23,21 @@ int main () { return 0; } -*/ -static xdo_t *xdo = NULL; - -int myoparasite_mouse_init (void) { - xdo = xdo_new (NULL); - if (xdo == NULL) { - return -1; - } - return 0; -} - -void myoparasite_mouse_click (int action) { - switch (action) { - case 1: case 2: case 3: case 4: case 5: - xdo_click (xdo, CURRENTWINDOW, button); - break; - } -} +// static xdo_t *xdo = NULL; +// +// int myoparasite_mouse_init (void) { +// xdo = xdo_new (NULL); +// if (xdo == NULL) { +// return -1; +// } +// return 0; +// } +// +// void myoparasite_mouse_click (int action) { +// switch (action) { +// case 1: case 2: case 3: case 4: case 5: +// xdo_click (xdo, CURRENTWINDOW, button); +// break; +// } +// } @@ -9,10 +9,12 @@ import subprocess import sys import time import threading +import struct import Classifier from sklearn.externals import joblib from sklearn.preprocessing import StandardScaler +from sklearn.neighbors import KNeighborsClassifier import code @@ -28,17 +30,20 @@ class readdataThread (threading.Thread) : global buf while True : - buf += self.FILE.readline (50).strip () + buf += self.FILE.read (50 * 4) +def extract_feature (data1, data2): + return np.concatenate(( + (np.absolute(np.fft.fft(data1))) , + (np.absolute(np.fft.fft(data2))) ) ).tolist() + def getdata (raw) : data1 = [] data2 = [] m = 1 - for j in range (0, len (raw), 4) : - n = int (raw[j:j+4], 16) - if n >= 32768 : - n -= 65536 + for j in range (0, len (raw), 2) : + n = struct.unpack ("<h", raw[j:j+2])[0] if m == 1 : data1.append (float(n)) m = 2 @@ -49,14 +54,15 @@ def getdata (raw) : def record () : - labels = range(Classifier.NUM_OF_LABELS-1, -1, -1) * 15 + # labels = range(Classifier.NUM_OF_LABELS-1, -1, -1) * 15 + labels = [0] * 15 + [1] * 15 # random.shuffle(labels) datas = [] rawdata1 = [] rawdata2 = [] record_time = len (labels) - shcmd = "arecord -c 2 -d %d -t raw -r 2000 -f S16_BE - 2>/dev/null | xxd -p | tr -d '\n' > tmppp" % (record_time + 1) + shcmd = "arecord -c 2 -d %d -t raw -r 2000 -f S16_LE - 2>/dev/null | xxd -p | tr -d '\n' > tmppp" % (record_time + 1) proc = subprocess.Popen (shcmd , stdout = subprocess.PIPE, shell = True) for i in range (-1, record_time) : if i > -1 : @@ -66,15 +72,19 @@ def record () : f = open ('tmppp', 'r') for i in range (-1, record_time) : if i == -1 : - f.readline (16000) + f.read (2000 * 4) continue - data1, data2 = getdata (f.readline (16000)) + data1, data2 = getdata (f.read (2000 * 4)) # plt.subplot(2,1,1) # plt.plot(data1) # plt.subplot(2,1,2) # plt.plot(data2) # plt.show() + # TODO + data2 = data1 + # TODO + rawdata1.append(data1) rawdata2.append(data2) @@ -98,11 +108,15 @@ def train(rawdata1, rawdata2, y): X1 = rawdata1 X2 = rawdata2 y_2 = [] - for yi, x1, x2 in zip(y, X1, X2): - for i in range(700, 1300, Classifier.WINDOW_SHIFT_TRAIN): - X.append( np.concatenate(( - np.absolute(np.fft.fft(x1[i: i+Classifier.WINDOW_SIZE])) , - np.absolute(np.fft.fft(x2[i: i+Classifier.WINDOW_SIZE])) ) ).tolist()) + for yi, x1, x2 in zip(y, X1[:-1], X2[:-1]): + for i in range(700, 1500, Classifier.WINDOW_SHIFT_TRAIN): + X.append( extract_feature( + x1[i: i+Classifier.WINDOW_SIZE], + x2[i: i+Classifier.WINDOW_SIZE]) ) + # X.append(x1[i: i+Classifier.WINDOW_SIZE]) + # X.append( np.concatenate(( + # np.absolute(np.fft.fft(x1[i: i+Classifier.WINDOW_SIZE])) , + # np.absolute(np.fft.fft(x2[i: i+Classifier.WINDOW_SIZE])) ) ).tolist()) y_2.append( yi ) y = y_2 scalers, classifiers, scores = Classifier.gen_model(X, y, verbose=False) @@ -113,41 +127,48 @@ def train(rawdata1, rawdata2, y): def predict (scalers, classifiers, scores) : global buf + tmpp = [] + t=0 + sys.stderr.write ("start predict\n") - shcmd = "arecord -t raw -c 2 -r 2000 -f S16_BE - 2>/dev/null | xxd -p" + shcmd = "arecord -t raw -c 2 -r 2000 -B 25 -f S16_LE - 2>/dev/null" proc = subprocess.Popen (shcmd, stdout = subprocess.PIPE, shell = True) read_thread = readdataThread (proc.stdout) read_thread.start () count = 0 - p = {} - while True : - if len (buf) >= Classifier.WINDOW_SIZE * 8 : - data1, data2 = getdata (buf[-Classifier.WINDOW_SIZE * 8:]) - buf = buf[50:] - - data1 = np.absolute(np.fft.fft(data1)) - data2 = np.absolute(np.fft.fft(data2)) - X = data1.tolist () + data2.tolist () + p = [0] * Classifier.NUM_OF_LABELS + while t<50000000 : + if len (buf) >= Classifier.WINDOW_SIZE * 4 : + data1, data2 = getdata (buf[-Classifier.WINDOW_SIZE * 4:]) + buf = buf[-(Classifier.WINDOW_SIZE - 50) * 4:] + +# TODO + data2 = data1 + + X = extract_feature(data1, data2) + tmpp.append(X) tp = Classifier.multi_classification(X, scalers, classifiers, scores)[0] - #tp = clf.predict(X)[0] - if tp in p: - p[tp] += 1 - else : - p[tp] = 1 + print tp + + p[tp] += 1 count += 1 - if count >= 20 : - maxi = 0 - maj = -1 - for m in p : - if p[m] > maxi : - maxi = p[m] - maj = m +# TODO + if count >= 10 : +# TODO + maj = p.index (max (p)) count = 0 - p = {} + p = [0] * Classifier.NUM_OF_LABELS print (maj) sys.stdout.flush () + t+=1 + for i in tmpp: + plt.plot(i) + plt.show() + print tmpp[0] + + read_thread.join () @@ -168,9 +189,10 @@ def main () : sys.stderr.write ("Wrong arguments\n") exit (1) scalers, classifiers, scores = train (rawx1, rawx2, y) - - predict (scalers, classifiers, scores) - + for i, j in zip(rawx1[:-1], rawx2[:-1]): + print len(i), len(j) + print Classifier.multi_classification(extract_feature(rawx1[500:1000], rawx2[500:1000]), scalers, classifiers, scores) + #predict (scalers, classifiers, scores) if __name__ == "__main__" : main () diff --git a/visualize.py b/visualize.py index b85e33a..318a4ed 100755 --- a/visualize.py +++ b/visualize.py @@ -15,9 +15,13 @@ f = open (argv[1], "r") labels, datas1, datas2 = zip(*json.loads (f.readline ())) for i in range (0, len (datas1)) : print (labels[i]) + tmppp = np.absolute(np.fft.fft(np.array(datas1[i]))) + tmppp2 = np.absolute(np.fft.fft(np.array(datas1[-i]))) plt.subplot (2, 1, 1) - plt.plot (datas1[i]) + #plt.yscale('log') + plt.plot (tmppp) plt.subplot (2, 1, 2) - plt.plot (datas2[i]) + #plt.yscale('log') + plt.plot (tmppp2) plt.show () f.close () |