def start(self):
if self.activities is not None:
for body in self.activities['body']['activities']:
for act in body:
value = body[act]
if act == 'hr_average':
val = value
self.setDriver('ST', val)
if act == 'hr_min':
val = value
self.setDriver('GV0', val)
if act == 'hr_max':
val = value
self.setDriver('GV1', val)
if act == 'hr_zone_0':
val = utils.seconds_to_minutes(value)
self.setDriver('GV2', val)
if act == 'hr_zone_1':
val = utils.seconds_to_minutes(value)
self.setDriver('GV3', val)
if act == 'hr_zone_2':
val = utils.seconds_to_minutes(value)
self.setDriver('GV4', val)
if act == 'hr_zone_3':
val = utils.seconds_to_minutes(value)
self.setDriver('GV5', val)
def start(self):
if self.devices is not None:
for dev in self.devices['body']['devices']:
if dev['type'] == "Activity Tracker":
value = dev['battery']
if value == "low":
battery = 1
elif value == "medium":
battery = 2
elif value == "high":
battery = 3
else:
battery = 0
self.setDriver('BATLVL', battery)
if self.activities is not None:
for body in self.activities['body']['activities']:
for act in body:
value = body[act]
if act == 'steps':
val = value
self.setDriver('ST', val)
if act == 'distance':
val = utils.meters_to_mile(value)
self.setDriver('GV0', val)
if act == 'elevation':
if value == 0:
self.setDriver('GV1', value, force=True)
else:
val = value
self.setDriver('GV1', val)
if act == 'soft':
val = utils.seconds_to_minutes(value)
self.setDriver('GV2', val)
if act == 'moderate':
val = utils.seconds_to_minutes(value)
self.setDriver('GV3', val)
if act == 'intense':
val = utils.seconds_to_minutes(value)
self.setDriver('GV4', val)
if act == 'active':
val = utils.seconds_to_minutes(value)
self.setDriver('GV5', val)
if act == 'calories':
val = round(value, 2)
self.setDriver('GV6', val)
if act == 'totalcalories':
val = round(value, 2)
self.setDriver('GV7', val)
def start(self):
if self.sleep is not None:
for series in self.sleep['body']['series']:
model = series['model']
if model == 16:
for entry in series['data']:
value = series['data'][entry]
if entry == "sleep_score":
val = value
self.setDriver('ST', val)
if entry == "lightsleepduration":
val = utils.seconds_to_minutes(value)
self.setDriver('GV0', val)
if entry == "deepsleepduration":
val = utils.seconds_to_minutes(value)
self.setDriver('GV1', val)
if entry == "wakeupcount":
val = value
self.setDriver('GV2', val)
if entry == "wakeupduration":
val = utils.seconds_to_minutes(value)
self.setDriver('GV3', val)
if entry == "durationtosleep":
val = utils.seconds_to_minutes(value)
self.setDriver('GV4', val)
if entry == "durationtowakeup":
val = utils.seconds_to_minutes(value)
self.setDriver('GV5', val)
if entry == "hr_average":
val = value
self.setDriver('GV6', val)
if entry == "hr_min":
val = value
self.setDriver('GV7', val)
if entry == "hr_max":
val = value
self.setDriver('GV8', val)
def start(self):
if self.sleep is not None:
for series in self.sleep['body']['series']:
model = series['model']
if model == 32:
for entry in series['data']:
value = series['data'][entry]
if entry == "sleep_score":
val = value
self.setDriver('ST', val)
if entry == "lightsleepduration":
val = utils.seconds_to_minutes(value)
self.setDriver('GV0', val)
if entry == "deepsleepduration":
val = utils.seconds_to_minutes(value)
self.setDriver('GV1', val)
if entry == "remsleepduration":
val = utils.seconds_to_minutes(value)
self.setDriver('GV2', val)
if entry == "wakeupcount":
val = value
self.setDriver('GV3', val)
if entry == "wakeupduration":
val = utils.seconds_to_minutes(value)
self.setDriver('GV4', val)
if entry == "durationtosleep":
val = utils.seconds_to_minutes(value)
self.setDriver('GV5', val)
if entry == "durationtowakeup":
val = utils.seconds_to_minutes(value)
self.setDriver('GV6', val)
if entry == "hr_average":
val = value
self.setDriver('GV7', val)
if entry == "hr_min":
val = value
self.setDriver('GV8', val)
if entry == "hr_max":
val = value
self.setDriver('GV9', val)
if entry == "rr_average":
val = value
self.setDriver('GV10', val)
if entry == "rr_min":
val = value
self.setDriver('GV11', val)
if entry == "rr_max":
val = value
self.setDriver('GV12', val)
if entry == "breathing_disturbances_intensity":
if value <= 30:
val = 1
elif value > 30 <= 60:
val = 2
elif value > 60:
val = 3
else:
val = 0
self.setDriver('GV13', val)
if entry == "snoring":
val = utils.seconds_to_minutes(value)
self.setDriver('GV14', val)
if entry == "snoringepisodecount":
val = value
self.setDriver('GV15', val)
f1_weighted = f1_score(np.argmax(y_test, axis=1),
np.argmax(pred, axis=1),
labels=None,
average='weighted',
sample_weight=None)
logger.append_log("F1 score (weighted) " + str(f1_weighted))
logger.append_log("F1 score (macro) " + str(
f1_score(np.argmax(y_test, axis=1),
np.argmax(pred, axis=1),
labels=None,
average='macro',
sample_weight=None)))
logger.append_log("F1 score (micro) " + str(
f1_score(np.argmax(y_test, axis=1),
np.argmax(pred, axis=1),
labels=None,
average='micro',
sample_weight=None))
) # weighted and micro preferred in case of imbalance
for i, row in enumerate(conf_mat):
logger.append_log("precision of class {} = {}".format(
i, np.round(row[i] / np.sum(row), 2)))
if is_last_batch:
break
logger.append_log("Complete training finished in {}".format(
seconds_to_minutes(time.time() - start_time)))
logger.flush()
plot_history(history, count)
min_arg = np.argmin(np.array(history.history['val_loss']))
logger.append_log("Best val_loss is {} and corresponding train_loss is {}".format(
history.history['val_loss'][min_arg], history.history['loss'][min_arg]))
test_res = model.evaluate(x_test, y_test, verbose=0)
logger.append_log("keras evaluate result =" + str(test_res)+", metrics:"+str(model.metrics_names))
pred = model.predict(x_test)
check_baseline(np.argmax(pred, axis=1), np.argmax(y_test, axis=1))
conf_mat = confusion_matrix(np.argmax(y_test, axis=1), np.argmax(pred, axis=1))
logger.append_log('\n'+str(conf_mat))
labels = [0, 1, 2]
f1_weighted = f1_score(np.argmax(y_test, axis=1), np.argmax(pred, axis=1), labels=None,
average='weighted', sample_weight=None)
logger.append_log("F1 score (weighted) " + str(f1_weighted))
logger.append_log(
"F1 score (macro) " + str(f1_score(np.argmax(y_test, axis=1), np.argmax(pred, axis=1), labels=None,
average='macro', sample_weight=None)))
logger.append_log(
"F1 score (micro) " + str(f1_score(np.argmax(y_test, axis=1), np.argmax(pred, axis=1), labels=None,
average='micro',
sample_weight=None))) # weighted and micro preferred in case of imbalance
for i, row in enumerate(conf_mat):
logger.append_log("precision of class {} = {}".format(i, np.round(row[i] / np.sum(row), 2)))
if is_last_batch:
break
logger.append_log("Complete training finished in {}".format(seconds_to_minutes(time.time() - start_time)))
logger.flush()