def update_data():
global img, animation_grid, updateTime, fps, data1, ptr1, temp_index_y, temp_index_x
data1 = np.roll(data1, -1)
data = a.propagate(ecg=True)
data = ani_convert(data,
shape=a.shape,
rp=a.rp,
animation_grid=animation_grid)
voltage = e.voltage(data, (temp_index_y, temp_index_x))
data1[-1] = voltage
ptr1 += 1
curve.setData(data1)
curve.setPos(ptr1, 0)
time.sleep(1 / 120.) # gives larger more stable fps.
img.setImage(data.T, levels=(0, 50))
QtCore.QTimer.singleShot(1, update_data)
now = ptime.time()
fps2 = 1.0 / (now - updateTime)
updateTime = now
fps = fps * 0.9 + fps2 * 0.1
constrainedx = [20, 179]
ecg_num = 0 #Num of ECGs
process_list = [] # process list
final_rotor_position = None # Final rotor position tuple
ECG_start_flag = False # Setting measurment flag to False (ecg measurments start when flag is triggered).
ECG_located_flag = False # To keep the system propogating until the rotor is found or limit is reached.
state = 0 # State for pipe work.
while not ECG_located_flag:
# Propogates model and converts into ani_data (used in ecg data processing)
data = a.propagate(ecg=True)
data = ani_convert(data,
shape=a.shape,
rp=a.rp,
animation_grid=animation_grid)
# If flag triggered, then start taking measurments.
if ECG_start_flag:
process_list.append(copy.deepcopy(data))
ptr1 += 1
if ptr1 >= stability_time:
if not ECG_start_flag:
ECG_start_flag = True
if ptr1 % process_length == 0 and ptr1 != stability_time:
sample = rt_ecg_gathering(process_list,
sign_para='record_sign')
def update_data():
global updateTime, fps, ptr1, process_list, ECG_start_flag, state, vsign_short_memory, constrainedy, constrainedx
global current_ecg_y_pos, current_ecg_x_pos, y_short_memory, x_short_memory, ecg_count, previousR, prev_y_vector
global prev_x_vector, hsign_short_memory
data = a.propagate(ecg=True)
data = ani_convert(data,
shape=a.shape,
rp=a.rp,
animation_grid=animation_grid)
# Initial Crosshair drawing
if ptr1 == 0:
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
# If flag triggered, then start taking measurments.
if ECG_start_flag:
process_list.append(copy.deepcopy(data))
ptr1 += 1
if ptr1 >= stability_time:
if not ECG_start_flag:
ECG_start_flag = True
if ptr1 % process_length == 0 and ptr1 != stability_time:
if state == 0:
sample, bsigns = rt_ecg_gathering(
process_list, sign_para="record_sign_plus"
) # ECG Recording and feature gathering
ecg_count += 1
sample = sample.reshape(
1, -1) # Get deprication warning if this is not done.
vsign = sample[0, :][-3]
# hsign = sample[0, :][-2]
# print hsign
# constrainedx = autojump_constrainx(h_sign = hsign, constrained_ = constrainedx, x_pos = current_ecg_x_pos, constrainedy = constrainedy)
# con_midpoint = int(np.mean(constrainedx))
# current_ecg_x_pos = con_midpoint
# first potential x constraint
# sample_ = sample[0, :][0:-3].reshape(1, -1) # Get sample without sign information.
sample_ = sample[0, :][0:].reshape(1, -1)
y_class_value = y_class.predict(sample_)[0]
y_probarg = reg_predictor(
y_classifier_full.predict_proba(sample_)[0, :])
if y_class_value == 1:
state = 1 # Change to state 1 for y axis regression/classification.
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
# del constrainedy
# constrainedy = [None, None]
x_class_value = x_class.predict(sample_)[0]
if x_class_value == 1:
previousR = "******"
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
# del constrainedy
# constrainedy = [None, None]
current_ecg_y_pos = randint(
20, 179) # reseting the process.
current_ecg_x_pos = randint(20, 179)
state = 0
ecg_count = 0
constrainedy = [None, None]
constrainedx = [20, 179]
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
if y_class_value == 0:
y_short_memory.append(current_ecg_y_pos)
vsign_short_memory.append(vsign)
constrainedy, vsign_short_memory = constrained_finder(
prev_y_vector, vsign_short_memory, current_ecg_y_pos,
constrainedy)
if condistance(constrainedy) == 1:
state = 1
x_class_value = x_class.predict(sample_)[0]
if x_class_value == 1:
previousR = "******"
# reseting the process.
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
# del constrainedy
# constrainedy = [None, None]
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
state = 0
ecg_count = 0
constrainedy = [None, None]
constrainedx = [20, 179]
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
else:
likelyp = prediction(y_probarg,
vector_constraint=vecdistance(
current_ecg_y_pos,
constrainedy),
axis='x')
y_vector = y_classifier_full.classes_[likelyp]
prev_y_vector = y_vector
# if np.abs(y_vector) > 3:
# pad_place = np.where(np.array(constrainedy) == current_ecg_y_pos)[0][0]
# if pad_place == 1:
# cons
current_ecg_y_pos -= y_vector
# print y_vector
if current_ecg_y_pos > 199 or current_ecg_y_pos < 0:
current_ecg_y_pos %= 200
# if current_ecg_y_pos in y_short_memory:
# print "Loop Y condition"
# loop_place = np.where(np.array(constrainedy) == current_ecg_y_pos)[0][0]
# if loop_place == 0:
# current_ecg_y_pos += 1
# if current_ecg_y_pos > 199:
# current_ecg_y_pos %= 200
# constrainedy[0] += 1
# if constrainedy[0] > 199:
# constrainedy[0] %= 200
# if loop_place == 1:
# current_ecg_y_pos -= 1
# if current_ecg_y_pos < 0:
# current_ecg_y_pos %= 200
# constrainedy[1] -= 1
# if constrainedy[1] < 0:
# constrainedy[1] %= 200
if state == 1:
sample, bsigns = rt_ecg_gathering(
process_list,
sign_para='record_sign_plus') # ECG feature Recording
ecg_count += 1
sample = sample.reshape(
1, -1) # Get deprication warning if this is not done.
hsign = sample[0, :][-2] # Gets the h sign
# sample_ = sample[0, :][0:-3].reshape(1, -1) # Takes a sample without sign
sample_ = sample[0, :][:].reshape(
1, -1) # Takes a sample without sign information
x_probarg = reg_predictor(
x_classifier_full.predict_proba(sample_)[0, :]) # Prob map
x_class_value = x_class.predict(sample_)[0]
if x_class_value == 1:
previousR = "******"
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
state = 0
ecg_count = 0
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
del constrainedx
constrainedy = [None, None]
constrainedx = [20, 179]
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
if x_class_value == 0:
x_short_memory.append(current_ecg_x_pos)
hsign_short_memory.append(hsign)
constrainedx, hsign_short_memory = constrained_finder(
prev_x_vector, hsign_short_memory, current_ecg_x_pos,
constrainedx)
if condistance(constrainedx) == 1:
previousR = "******"
state = 1
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
del constrainedx
constrainedx = [20, 179]
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
state = 0
ecg_count = 0
constrainedy = [None, None]
constrainedx = [20, 179]
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
else:
likelyp = prediction(x_probarg,
vector_constraint=vecdistance(
current_ecg_x_pos,
constrainedx),
axis='y')
x_vector = x_classifier_full.classes_[likelyp]
prev_x_vector = x_vector
current_ecg_x_pos -= x_vector
if current_ecg_x_pos > 199 or current_ecg_x_pos < 0:
current_ecg_x_pos %= 200
if current_ecg_x_pos in x_short_memory:
print "X LOOP condition."
loop_place = np.where(
np.array(constrainedx) ==
current_ecg_x_pos)[0][0]
if loop_place == 0:
current_ecg_x_pos += 1
if current_ecg_x_pos > 199:
current_ecg_x_pos %= 200
constrainedx[0] += 1
if constrainedx[0] > 199:
constrainedx[0] %= 200
if loop_place == 1:
current_ecg_x_pos -= 1
if current_ecg_x_pos < 0:
current_ecg_x_pos %= 200
constrainedx[1] -= 1
if constrainedx[1] < 0:
constrainedx[1] %= 200
ecg_processing.reset_singlegrid(
(current_ecg_y_pos, current_ecg_x_pos))
if constrainedy[0] is not None:
xLline.setPos(constrainedy[0])
if constrainedy[1] is not None:
xUline.setPos(constrainedy[1])
yLline.setPos(constrainedx[0])
yUline.setPos(constrainedx[1])
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
del process_list
process_list = []
update_label_text(cp_x_pos, cp_y_pos, current_ecg_x_pos,
current_ecg_y_pos, ecg_count, previousR,
constrainedy, constrainedx)
time.sleep(1 / 120.) # gives more stable fps.
img.setImage(data.T) # puts animation grid on image.
# Stuff to do with time and fps.
QtCore.QTimer.singleShot(1, update_data)
now = ptime.time()
fps2 = 1.0 / (now - updateTime)
updateTime = now
fps = fps * 0.9 + fps2 * 0.1
def update_data():
global updateTime, fps, ptr1, process_list, ECG_start_flag, state, vsign_short_memory, constrainedy, constrainedx
global current_ecg_y_pos, current_ecg_x_pos, y_short_memory, x_short_memory, ecg_count, previousR, prev_y_vector
global prev_x_vector, hsign_short_memory, num_ypos_class, num_xpos_class, num_Yloops, num_Xloops, num_yconstraint
global num_xconstraint, num_xzjump, num_yzjump
data = a.propagate(ecg=True)
data = ani_convert(data,
shape=a.shape,
rp=a.rp,
animation_grid=animation_grid)
# Initial Crosshair drawing
if ptr1 == 0:
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
# If flag triggered, then start taking measurments.
if ECG_start_flag:
process_list.append(copy.deepcopy(data))
ptr1 += 1
if ptr1 >= stability_time:
if not ECG_start_flag:
ECG_start_flag = True
if ptr1 % process_length == 0 and ptr1 != stability_time:
sample = rt_ecg_gathering(
process_list,
sign_para=args[5]) # ECG Recording and feature gathering
ecg_count += 1
if state == 0:
sample = sample.reshape(
1, -1) # Get deprication warning if this is not done.
vsign = sample[0, :][-3]
# first potential x constraint
# sample_ = sample[0, :][0:-3].reshape(1, -1) # Get sample without sign information.
y_class_value = y_class.predict(sample)[0]
y_probarg = movingaverage(
y_classifier_full.predict_proba(sample)[0, :], 10)
if y_class_value == 1:
state = 1 # Change to state 1 for y axis regression/classification.
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
x_class_value = x_class.predict(sample)[0]
num_ypos_class += 1
if x_class_value == 1:
previousR = "******" % (current_ecg_x_pos,
current_ecg_y_pos)
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
del constrainedy
constrainedy = [None, None]
current_ecg_y_pos = randint(
20, 179) # reseting the process.
current_ecg_x_pos = randint(20, 179)
state = 0
ecg_count = 0
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
num_xpos_class += 1
if y_class_value == 0:
y_short_memory.append(current_ecg_y_pos)
vsign_short_memory.append(vsign)
constrainedy, vsign_short_memory = constrained_finder(
prev_y_vector,
vsign_short_memory,
current_ecg_y_pos,
constrainedy,
axis='x')
if condistance(constrainedy) == 1:
state = 1
x_class_value = x_class.predict(sample)[0]
num_yconstraint += 1
if x_class_value == 1:
previousR = "******" % (
current_ecg_x_pos, current_ecg_y_pos)
# reseting the process.
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
del constrainedy
constrainedy = [None, None]
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
state = 0
ecg_count = 0
num_xpos_class += 1
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
else:
likelyp = prediction(y_probarg,
vector_constraint=vecdistance(
current_ecg_y_pos,
constrainedy),
axis='x')
y_vector = y_classifier_full.classes_[likelyp]
if y_vector == 0:
state = 1 # Change to state 1 for y axis regression/classification.
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
x_class_value = x_class.predict(sample)[0]
num_yzjump += 1
if x_class_value == 1:
previousR = "******" % (
current_ecg_x_pos, current_ecg_y_pos)
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
del constrainedy
constrainedy = [None, None]
current_ecg_y_pos = randint(
20, 179) # reseting the process.
current_ecg_x_pos = randint(20, 179)
state = 0
ecg_count = 0
num_xpos_class += 1
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(
300
) # If it predicts a jump of 0, then take the position it's on.
prev_y_vector = y_vector
current_ecg_y_pos -= y_vector
if current_ecg_y_pos > 199 or current_ecg_y_pos < 0:
current_ecg_y_pos %= 200
if state == 1:
sample = sample.reshape(
1, -1) # Get deprication warning if this is not done.
hsign = sample[0, :][-2] # Gets the h sign
# sample_ = sample[0, :][0:-3].reshape(1, -1) # Takes a sample without sign information
x_probarg = movingaverage(
x_classifier_full.predict_proba(sample)[0, :],
10) # Prob map
x_class_value = x_class.predict(sample)[0]
if x_class_value == 1:
previousR = "******" % (current_ecg_x_pos,
current_ecg_y_pos)
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
state = 0
ecg_count = 0
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
del constrainedx
constrainedx = [20, 179]
del constrainedy
constrainedy = [None, None]
num_xpos_class += 1
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
if x_class_value == 0:
x_short_memory.append(current_ecg_x_pos)
hsign_short_memory.append(hsign)
constrainedx, hsign_short_memory = constrained_finder(
prev_x_vector,
hsign_short_memory,
current_ecg_x_pos,
constrainedx,
axis='y')
if condistance(
constrainedx
) == 1: # Row is constrained to be have distance 1, take position.
previousR = "******" % (
current_ecg_x_pos, current_ecg_y_pos)
state = 1
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
del constrainedx
constrainedx = [20, 179]
del constrainedy
constrainedy = [None, None]
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
state = 0
ecg_count = 0
num_xconstraint += 1
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
else:
likelyp = prediction(x_probarg,
vector_constraint=vecdistance(
current_ecg_x_pos,
constrainedx),
axis='y')
x_vector = x_classifier_full.classes_[likelyp]
if x_vector == 0: # If the predicted X jump is 0
previousR = "******" % (
current_ecg_x_pos, current_ecg_y_pos)
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
state = 0
ecg_count = 0
num_xzjump += 1
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
del constrainedx
constrainedx = [20, 179]
del constrainedy
constrainedy = [None, None]
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
prev_x_vector = x_vector
current_ecg_x_pos -= x_vector
if current_ecg_x_pos > 199 or current_ecg_x_pos < 0:
current_ecg_x_pos %= 200
if current_ecg_x_pos in x_short_memory:
previousR = '******' % current_ecg_y_pos
state = 1
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
del constrainedx
constrainedx = [20, 179]
del constrainedy
constrainedy = [None, None]
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
state = 0
ecg_count = 0
num_Xloops += 1
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
ecg_processing.reset_singlegrid(
(current_ecg_y_pos, current_ecg_x_pos))
if constrainedy[0] is not None:
xLline.setPos(constrainedy[0])
if constrainedy[1] is not None:
xUline.setPos(constrainedy[1])
yLline.setPos(constrainedx[0])
yUline.setPos(constrainedx[1])
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
del process_list
process_list = []
update_label_text(cp_x_pos, cp_y_pos, current_ecg_x_pos,
current_ecg_y_pos, ecg_count, previousR,
constrainedy, constrainedx, num_ypos_class,
num_xpos_class, num_Yloops, num_Xloops,
num_yconstraint, num_xconstraint, num_yzjump,
num_xzjump)
time.sleep(1 / 120.) # gives more stable fps.
img.setImage(data.T) # puts animation grid on image.
# Stuff to do with time and fps.
QtCore.QTimer.singleShot(1, update_data)
now = ptime.time()
fps2 = 1.0 / (now - updateTime)
updateTime = now
fps = fps * 0.9 + fps2 * 0.1
def update_data():
global updateTime, fps, ptr1, process_list, ECG_start_flag, state, y_regress_treshold
global current_ecg_y_pos, current_ecg_x_pos, y_short_memory, x_short_memory
data = a.propagate(ecg=True)
data = ani_convert(data,
shape=a.shape,
rp=a.rp,
animation_grid=animation_grid)
# Initial Crosshair drawing
if ptr1 == 0:
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
# If flag triggered, then start taking measurments.
if ECG_start_flag:
process_list.append(copy.deepcopy(data))
ptr1 += 1
if ptr1 >= stability_time:
if not ECG_start_flag:
print "Starting Measurment Process"
ECG_start_flag = True
if ptr1 % process_length == 0 and ptr1 != stability_time:
if state == 0:
# ECG Recording and feature gathering
sample = rt_ecg_gathering(process_list, "record sign")
hsign = sample[-2]
print hsign
# Get deprication warning if this is not done.
sample = sample.reshape(1, -1)
y_class_value = y_estimator.predict(sample)[0]
print "Y classification: %s" % y_class_value
y_vector = int(y_regress.predict(sample)[0])
print "Y Vector prediction: %s" % y_vector
if y_class_value == 1:
# Change to state 1 for y axis regression/classification.
state = 1
print "Found X-Axis"
# Temporary
del y_short_memory
y_short_memory = []
x_class_value = x_class.predict(sample)[0]
print "X classification: %s" % x_class_value
if x_class_value == 1:
print "Found the rotor!"
print "Predicted Rotor position: (%s, %s)" % (
current_ecg_x_pos, current_ecg_y_pos)
# reseting the process.
current_ecg_x_pos = randint(30, 169)
current_ecg_y_pos = randint(0, 199)
state = 0
if y_class_value == 0:
y_short_memory.append(current_ecg_y_pos)
current_ecg_y_pos -= y_vector
current_ecg_x_pos -= int(3 * hsign * np.absolute(hsign) *
np.absolute(y_vector))
if current_ecg_y_pos > 199 or current_ecg_y_pos < 0: #Is this an error?
current_ecg_y_pos %= 200
if current_ecg_x_pos > 199 or current_ecg_x_pos < 0:
current_ecg_x_pos += int(3 * hsign *
np.absolute(hsign) *
np.absolute(y_vector))
if current_ecg_y_pos in y_short_memory:
print "Entered Y Loop"
# print "Loop: %s" % y_short_memory
# loop_average = int((float(sum(y_short_memory))/len(y_short_memory)))
# print "Loop Average: %s" % loop_average
# del y_short_memory
# y_short_memory = []
# current_ecg_y_pos = loop_average
current_ecg_x_pos = randint(30, 169)
current_ecg_y_pos = randint(0, 199)
if state == 1:
# ECG Recording and feature gathering
sample = rt_ecg_gathering(process_list, "record sign")
hsign = sample[-2]
print hsign
# Get deprication warning if this is not done.
sample = sample.reshape(1, -1)
x_class_value = x_class.predict(sample)[0]
print "X classification: %s" % x_class_value
x_vector = int(x_regress.predict(sample)[0])
print "X Vector prediction: %s" % x_vector
if x_class_value == 1:
print "Found the rotor!"
print "Predicted Rotor position: (%s, %s)" % (
current_ecg_x_pos, current_ecg_y_pos)
# Temporary
del x_short_memory
x_short_memory = []
# reseting the process.
current_ecg_x_pos = randint(30, 169)
current_ecg_y_pos = randint(0, 199)
state = 0
if x_class_value == 0:
x_short_memory.append(current_ecg_x_pos)
current_ecg_x_pos -= x_vector + 13
if current_ecg_x_pos > 199 or current_ecg_x_pos < 0:
current_ecg_x_pos %= 200
if current_ecg_x_pos in x_short_memory:
print "Entered X Loop"
# print "Loop: %s" % x_short_memory
# loop_average = int((float(sum(x_short_memory))/len(x_short_memory)))
# print "Loop Average: %s" % loop_average
# del x_short_memory
# x_short_memory = []
# current_ecg_x_pos = loop_average
current_ecg_x_pos = randint(30, 169)
current_ecg_y_pos = randint(0, 199)
print "New ECG Probe position: (%s, %s)" % (current_ecg_x_pos,
current_ecg_y_pos)
print '\n'
ecg_processing.reset_singlegrid(
(current_ecg_y_pos, current_ecg_x_pos))
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
del process_list
process_list = []
updating_text = """ECG Position: (%s, %s)<br>\n
Rotor Position: (%s, %s)""" % (
current_ecg_x_pos, current_ecg_y_pos, cp_x_pos, cp_y_pos)
label.setText(updating_text)
# gives more stable fps.
time.sleep(1 / 120.)
# puts animation grid on image.
img.setImage(data.T)
# Stuff to do with time and fps.
QtCore.QTimer.singleShot(1, update_data)
now = ptime.time()
fps2 = 1.0 / (now - updateTime)
updateTime = now
fps = fps * 0.9 + fps2 * 0.1
def update_data():
global updateTime, fps, ptr1, process_list, ECG_start_flag, state, vsign_short_memory, constrainedy, constrainedx
global current_ecg_y_pos, current_ecg_x_pos, y_short_memory, x_short_memory, ecg_count, previousR, prev_y_vector
global prev_x_vector, hsign_short_memory, num_ypos_class, num_xpos_class, num_Yloops, num_Xloops, num_yconstraint
global num_xconstraint, num_xzjump, num_yzjump, rotors_found, perminant_constraints, N_rotors, total_sign_info, vconsistent, hconsistent, bconsistent
global x_history, y_history, special_state, jump_x, special_vsign, special_y
data = a.propagate(ecg=True)
data = ani_convert(data,
shape=a.shape,
rp=a.rp,
animation_grid=animation_grid)
# Initial Crosshair drawing
if ptr1 == 0:
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
xUline.setPos(300)
xLline.setPos(300)
pUline.setPos(-300)
pLline.setPos(-300)
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
# If flag triggered, then start taking measurments.
if ECG_start_flag:
process_list.append(copy.deepcopy(data))
ptr1 += 1
# CONDITION TO START THE LOCATING PROCESS
if ptr1 >= stability_time:
if not ECG_start_flag:
ECG_start_flag = True
# CONDITION TO TAKE A MEASURMENT
if ptr1 % process_length == 0 and ptr1 != stability_time:
sample, signs = rt_ecg_gathering(
process_list, sign_para='record_sign_plus'
) # ECG Recording and feature gathering
ecg_count += 1
total_sign_info.append(signs)
x_history.append(current_ecg_x_pos)
y_history.append(current_ecg_y_pos)
# LOOKING FOR THE ROTORS X AXIS
if state == 0:
sample = sample.reshape(
1, -1) # Get deprication warning if this is not done.
vsign = sample[0, :][-3]
# first potential x constraint
# sample_ = sample[0, :][0:-3].reshape(1, -1) # Get sample without sign information.
y_class_value = y_class.predict(sample)[0]
y_probarg = reg_predictor(
y_classifier_full.predict_proba(sample)[0, :])
# POSITIVE CLASSIFICATION FOR Y
if y_class_value == 1:
if special_state:
special_state = False
state = 1 # Change to state 1 for y axis regression/classification.
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
x_class_value = x_class.predict(sample)[0]
num_ypos_class += 1
# CHECKS IF ITS ON THE CORRECT Y AXIS
if x_class_value == 1:
previousR = "******" % (current_ecg_x_pos,
current_ecg_y_pos)
state = 0
ecg_count = 0
pred_rotor_y.append(current_ecg_y_pos)
pred_rotor_x.append(current_ecg_x_pos)
# ALL ROTORS ARE FOUND
if rotors_found == N_rotors:
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
xUline.setPos(300)
xLline.setPos(300)
pUline.setPos(-300)
pLline.setPos(-300)
rotors_found = 0
perminant_constraints = []
# ONE OF THE ROTORS IS FOUND
else:
rotors_found += 1
upper = current_ecg_y_pos - 5
upper %= 200
lower = current_ecg_y_pos + 5
lower %= 200
perminant_constraints.append([lower, upper])
xUline.setPos(300)
xLline.setPos(300)
pUline.setPos(upper)
pLline.setPos(lower)
xvec, yvec = relative_vectors(
x_history, y_history, current_ecg_x_pos,
current_ecg_y_pos)
xvec = np.array(xvec)
yvec = np.array(yvec)
for i in range(len(x_history)):
vconsistent.append(
check_signs(xvec[i],
yvec[i],
total_sign_info[i][0],
vsign_check,
thr=0.05))
hconsistent.append(
check_signs(xvec[i],
yvec[i],
total_sign_info[i][1],
hsign_check,
thr=0.02))
bconsistent.append(
check_bsign(total_sign_info[i][-2],
total_sign_info[i][-1]))
vconsistent = np.array(vconsistent)[
np.absolute(yvec) > 4]
hconsistent = np.array(hconsistent)[
np.absolute(yvec) > 4]
bconsistent = np.array(bconsistent)[
np.absolute(yvec) > 4]
total_sign_info = np.array(total_sign_info)
tsign = total_sign_info[np.absolute(yvec) > 4]
yvec_use = yvec[np.absolute(yvec) > 4]
xvec_use = xvec[np.absolute(yvec) > 4]
special_y = (yvec_use[vconsistent == False] +
current_ecg_y_pos) % 200
special_vsign = tsign[:, 0][vconsistent == False]
current_ecg_y_pos = (current_ecg_y_pos + 100) % 200
special_state = True
constrainedy = [None, None]
constrainedx = [20, 179]
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
num_xpos_class += 1
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
total_sign_info = []
vconsistent = []
hconsistent = []
bconsistent = []
x_history = []
y_history = []
# NEGATIVE CLASSIFIACTION FOR Y
if y_class_value == 0:
y_short_memory.append(current_ecg_y_pos)
vsign_short_memory.append(vsign)
if not special_state:
constrainedy, vsign_short_memory = constrained_finder(
prev_y_vector, vsign_short_memory,
current_ecg_y_pos, constrainedy, 'x',
perminant_constraints)
if special_state:
constrainedx, constrainedy, jump_x = special_constraint_finder(
current_ecg_x_pos, current_ecg_y_pos,
total_sign_info, constrainedy, constrainedx,
perminant_constraints, special_y, special_vsign)
# might need this to fix weird constraints
vsign_short_memory = []
# CONSTRAINED CONDITION FOR Y
if condistance(constrainedy) == 0:
state = 0
ecg_count = 0
num_yconstraint += 1
xUline.setPos(300)
xLline.setPos(300)
constrainedy = [None, None]
constrainedx = [20, 179]
if rotors_found > 0:
current_ecg_y_pos = (pred_rotor_y[-1] + 100) % 200
else:
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
total_sign_info = []
vconsistent = []
hconsistent = []
bconsistent = []
x_history = []
y_history = []
# MOVING THE PROBE IN THE Y AXIS
else:
likelyp = prediction(y_probarg,
vector_constraint=vecdistance(
current_ecg_y_pos,
constrainedy),
axis='x')
y_vector = y_classifier_full.classes_[likelyp]
# SPECIAL CONDITION
if np.abs(y_vector) > 45 and special_state:
print y_vector
y_vector = int(45 * copysign(1, y_vector))
special_state = False
# JUMPS IN THE X DIRECTION AS WELL UNDER SPECIAL CONDITION
if jump_x:
h = total_sign_info[-1][1]
if np.abs(h) > 1.0:
print 'constraint jump'
if h > 0:
current_ecg_x_pos = int(
(current_ecg_x_pos + 20) / 2.)
else:
current_ecg_x_pos = int(
(current_ecg_x_pos + 179) / 2.)
else:
if 179 - current_ecg_x_pos >= 80:
current_ecg_x_pos += 80
else:
current_ecg_x_pos -= 80
jump_x = False
# IF THE PREDICTED Y JUMP IS ZERO
if y_vector == 0:
previousR = "******" % (
current_ecg_x_pos, current_ecg_y_pos)
state = 0
ecg_count = 0
num_yzjump += 1
xUline.setPos(300)
xLline.setPos(300)
constrainedy = [None, None]
constrainedx = [20, 179]
if rotors_found > 0:
current_ecg_y_pos = (pred_rotor_y[-1] +
100) % 200
else:
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
prev_y_vector = y_vector
current_ecg_y_pos -= y_vector
# WRAPPING AROUND THE BOUNDRY CONDITION
if current_ecg_y_pos > 199 or current_ecg_y_pos < 0:
current_ecg_y_pos %= 200
# Y LOOP FORM CHECK
if current_ecg_y_pos in y_short_memory:
previousR = '******'
state = 0
ecg_count = 0
num_Yloops += 1
xUline.setPos(300)
xLline.setPos(300)
constrainedy = [None, None]
constrainedx = [20, 179]
if rotors_found > 0:
current_ecg_y_pos = (current_ecg_y_pos +
100) % 200
else:
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
del y_short_memory
y_short_memory = []
del vsign_short_memory
vsign_short_memory = []
total_sign_info = []
vconsistent = []
hconsistent = []
bconsistent = []
x_history = []
y_history = []
# LOOKING FOR THE ROTORS Y AXIS
if state == 1:
sample = sample.reshape(
1, -1) # Get deprication warning if this is not done.
hsign = sample[0, :][-2] # Gets the h sign
# sample_ = sample[0, :][0:-3].reshape(1, -1) # Takes a sample without sign information
x_probarg = reg_predictor(
x_classifier_full.predict_proba(sample)[0, :]) # Prob map
x_class_value = x_class.predict(sample)[0]
# POSITIVE CLASSIFICATION FOR X
if x_class_value == 1:
previousR = "******" % (current_ecg_x_pos,
current_ecg_y_pos)
state = 0
ecg_count = 0
num_xpos_class += 1
pred_rotor_y.append(current_ecg_y_pos)
pred_rotor_x.append(current_ecg_x_pos)
if rotors_found == N_rotors:
xUline.setPos(300)
xLline.setPos(300)
pUline.setPos(-300)
pLline.setPos(-300)
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
rotors_found = 0
perminant_constraints = []
else:
rotors_found += 1
upper = current_ecg_y_pos - 5
upper %= 200
lower = current_ecg_y_pos + 5
lower %= 200
perminant_constraints.append([lower, upper])
xUline.setPos(300)
xLline.setPos(300)
pUline.setPos(upper)
pLline.setPos(lower)
xvec, yvec = relative_vectors(x_history, y_history,
current_ecg_x_pos,
current_ecg_y_pos)
xvec = np.array(xvec)
yvec = np.array(yvec)
for i in range(len(x_history)):
vconsistent.append(
check_signs(xvec[i],
yvec[i],
total_sign_info[i][0],
vsign_check,
thr=0.05))
hconsistent.append(
check_signs(xvec[i],
yvec[i],
total_sign_info[i][1],
hsign_check,
thr=0.02))
bconsistent.append(
check_bsign(total_sign_info[i][-2],
total_sign_info[i][-1]))
vconsistent = np.array(vconsistent)[
np.absolute(yvec) > 4]
hconsistent = np.array(hconsistent)[
np.absolute(yvec) > 4]
bconsistent = np.array(bconsistent)[
np.absolute(yvec) > 4]
total_sign_info = np.array(total_sign_info)
tsign = total_sign_info[np.absolute(yvec) > 4]
yvec_use = yvec[np.absolute(yvec) > 4]
xvec_use = xvec[np.absolute(yvec) > 4]
special_y = (yvec_use[vconsistent == False] +
current_ecg_y_pos) % 200
special_vsign = tsign[:, 0][vconsistent == False]
current_ecg_y_pos = (current_ecg_y_pos + 100) % 200
special_state = True
constrainedy = [None, None]
constrainedx = [20, 179]
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
total_sign_info = []
vconsistent = []
hconsistent = []
bconsistent = []
x_history = []
y_history = []
# NEGATIVE CLASSIFICATION FOR X
if x_class_value == 0:
x_short_memory.append(current_ecg_x_pos)
hsign_short_memory.append(hsign)
constrainedx, hsign_short_memory = constrained_finder(
prev_x_vector, hsign_short_memory, current_ecg_x_pos,
constrainedx, 'y', perminant_constraints)
# CONSTRAINED CONDITION FOR X
if condistance(
constrainedx
) == 0: # Row is constrained to be have distance 1, take position.
previousR = "******" % (
current_ecg_x_pos, current_ecg_y_pos)
state = 0
ecg_count = 0
num_xconstraint += 1
xUline.setPos(300)
xLline.setPos(300)
constrainedy = [None, None]
constrainedx = [20, 179]
if rotors_found > 0:
current_ecg_y_pos = (pred_rotor_y[-1] + 100) % 200
else:
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
total_sign_info = []
x_history = []
y_history = []
vconsistent = []
hconsistent = []
bconsistent = []
# MOVING THE PROBE IN THE X AXIS
else:
likelyp = prediction(x_probarg,
vector_constraint=vecdistance(
current_ecg_x_pos,
constrainedx),
axis='y')
x_vector = x_classifier_full.classes_[likelyp]
# IF THE PREDICTED JUMP IS ZERO
if x_vector == 0: # If the predicted X jump is 0
previousR = "******" % (
current_ecg_x_pos, current_ecg_y_pos)
state = 0
ecg_count = 0
num_xzjump += 1
xUline.setPos(300)
xLline.setPos(300)
constrainedy = [None, None]
constrainedx = [20, 179]
if rotors_found > 0:
current_ecg_y_pos = (pred_rotor_y[-1] +
100) % 200
else:
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
total_sign_info = []
x_history = []
y_history = []
vconsistent = []
hconsistent = []
bconsistent = []
prev_x_vector = x_vector
current_ecg_x_pos -= x_vector
# WRAPPING AROUND THE BOUNDRY
if current_ecg_x_pos > 199 or current_ecg_x_pos < 0:
current_ecg_x_pos %= 200
# X LOOP FORM CHECK
if current_ecg_x_pos in x_short_memory:
previousR = '******' % current_ecg_y_pos
state = 0
ecg_count = 0
num_Xloops += 1
xUline.setPos(300)
xLline.setPos(300)
constrainedy = [None, None]
constrainedx = [20, 179]
if rotors_found > 0:
current_ecg_y_pos = (pred_rotor_y[-1] +
100) % 200
else:
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
yUline.setPos(constrainedx[1])
yLline.setPos(constrainedx[0])
del x_short_memory
x_short_memory = []
del hsign_short_memory
hsign_short_memory = []
total_sign_info = []
x_history = []
y_history = []
vconsistent = []
hconsistent = []
bconsistent = []
# UPDATING LINES AND PREPARING FOR NEW MEASURMENT
ecg_processing.reset_singlegrid(
(current_ecg_y_pos, current_ecg_x_pos))
if constrainedy[0] is not None:
xLline.setPos(constrainedy[0])
if constrainedy[1] is not None:
xUline.setPos(constrainedy[1])
yLline.setPos(constrainedx[0])
yUline.setPos(constrainedx[1])
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
del process_list
process_list = []
update_label_text(cp_x_pos, cp_y_pos, cp_x_pos2, cp_y_pos2,
current_ecg_x_pos, current_ecg_y_pos, ecg_count,
previousR, constrainedy, constrainedx,
perminant_constraints, num_ypos_class,
num_xpos_class, num_Yloops, num_Xloops,
num_yconstraint, num_xconstraint, num_yzjump,
num_xzjump)
# time.sleep(1/120.) # gives more stable fps.
img.setImage(data.T) # puts animation grid on image.
# Stuff to do with time and fps.
QtCore.QTimer.singleShot(1, update_data)
now = ptime.time()
fps2 = 1.0 / (now - updateTime)
updateTime = now
fps = fps * 0.9 + fps2 * 0.1
def update_data():
global updateTime, fps, ptr1, process_list, ECG_start_flag, state, y_regress_treshold
global current_ecg_y_pos, current_ecg_x_pos, y_short_memory, x_short_memory, ecg_count, previousR
# resets the tissue if rotor is found or loop is reached. (Needs work
# (have to reset the time before a measurment is taken.))
# if tissue_reset:
# a = ps.Heart(nu=0.2, delta=0.0, fakedata=True)
# # Randomises the rotor x,y position
# cp_x_pos = randint(0, 199)
# cp_y_pos = randint(0, 199)
# a.set_pulse(60, [[cp_y_pos], [cp_x_pos]])
# update_label_text(cp_x_pos, cp_y_pos, current_ecg_x_pos, current_ecg_y_pos, ecg_count)
# tissue_reset = False
data = a.propagate(ecg=True)
data = ani_convert(data,
shape=a.shape,
rp=a.rp,
animation_grid=animation_grid)
# Initial Crosshair drawing
if ptr1 == 0:
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
# If flag triggered, then start taking measurments.
if ECG_start_flag:
process_list.append(copy.deepcopy(data))
ptr1 += 1
if ptr1 >= stability_time:
if not ECG_start_flag:
ECG_start_flag = True
if ptr1 % process_length == 0 and ptr1 != stability_time:
if state == 0:
# ECG Recording and feature gathering
sample = rt_ecg_gathering(process_list)
ecg_count += 1
# Get deprication warning if this is not done.
sample = sample.reshape(1, -1)
y_class_value = y_estimator.predict(sample)[0]
y_vector = int(y_regress.predict(sample)[0])
if y_class_value == 1:
# Change to state 1 for y axis regression/classification.
state = 1
del y_short_memory
y_short_memory = []
x_class_value = x_class.predict(sample)[0]
if x_class_value == 1:
previousR = "******"
# reseting the process.
del y_short_memory
y_short_memory = []
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
state = 0
ecg_count = 0
if y_class_value == 0:
y_short_memory.append(current_ecg_y_pos)
current_ecg_y_pos -= y_vector
if current_ecg_y_pos > 199 or current_ecg_y_pos < 0:
current_ecg_y_pos %= 200
if current_ecg_y_pos in y_short_memory:
previousR = "******"
ecg_count = 0
current_ecg_x_pos = randint(30, 169)
current_ecg_y_pos = randint(0, 199)
if state == 1:
# ECG Recording and feature gathering
sample = rt_ecg_gathering(process_list)
print sample[441:445]
ecg_count += 1
# Get deprication warning if this is not done.
sample = sample.reshape(1, -1)
x_class_value = x_class.predict(sample)[0]
x_vector = int(x_regress.predict(sample)[0])
if x_class_value == 1:
previousR = "******"
del x_short_memory
x_short_memory = []
# reseting the process.
current_ecg_y_pos = randint(0, 199)
current_ecg_x_pos = randint(20, 179)
state = 0
ecg_count = 0
if x_class_value == 0:
x_short_memory.append(current_ecg_x_pos)
current_ecg_x_pos -= x_vector
if current_ecg_x_pos > 199 or current_ecg_x_pos < 0:
current_ecg_x_pos %= 200
if current_ecg_x_pos in x_short_memory:
previousR = '******'
ecg_count = 0
del x_short_memory
x_short_memory = []
current_ecg_x_pos = randint(20, 179)
current_ecg_y_pos = randint(0, 199)
state = 0
ecg_processing.reset_singlegrid(
(current_ecg_y_pos, current_ecg_x_pos))
vLine.setPos(current_ecg_x_pos + 0.5)
hLine.setPos(current_ecg_y_pos + 0.5)
del process_list
process_list = []
update_label_text(cp_x_pos, cp_y_pos, current_ecg_x_pos,
current_ecg_y_pos, ecg_count, previousR)
# time.sleep(1/120.) # gives more stable fps.
img.setImage(data.T) # puts animation grid on image.
# Stuff to do with time and fps.
QtCore.QTimer.singleShot(1, update_data)
now = ptime.time()
fps2 = 1.0 / (now - updateTime)
updateTime = now
fps = fps * 0.9 + fps2 * 0.1