def dashboard():
data = None
d = Read_Data()
d = d.read()
if(('user' in session and session['user'] == params['admin_user'])):
return render_template('dashboard.html',data = d)
elif('user' in session and session['user'] == params['username']):
if(request.method == 'POST'):
name = request.form.get('Name')
email = request.form.get('Email')
msg = request.form.get('Msg')
data = {'Name':name,'Email':email,'Messages':msg}
check_file = False
try:
df = pd.read_csv("Messages.csv",index_col=0)
check_file = True
except:
df = pd.DataFrame(data,index = [0])
df.to_csv("Messages.csv")
if(check_file):
df = df.append(data,ignore_index = True)
df.to_csv("Messages.csv")
usr = params['username']
return render_template('user_dashboard.html',data = d,usr = usr)
msg = None
if(request.method == 'POST'):
usrname = request.form.get('usrname')
password = request.form.get('pass')
if(usrname == params['admin_user'] and password == params['admin_pass']):
session['user'] = usrname
return render_template('dashboard.html',data = d)
if(usrname == params['username'] and password == params['password']):
session['user'] = usrname
usr = usrname
return render_template('user_dashboard.html',data = d,usr = usr)
else:
msg = 0
return render_template('Login.html',msg = msg)
return render_template('Login.html',msg = msg)
def grid_search(nb_epoch=200):
global NUM_CLASSES, input_shape
file_path = "./gridSearch/fft_params/"
nfft_try = [int(4096), int(2048), int(1024)]
overlap_try = [0.9, 0.5, 0.75]
brange_try = [8, 16]
K.clear_session()
# Grid Search
for nfft_val in nfft_try:
for overlap_val in overlap_try:
for brange_val in brange_try:
# Define file name to store results
fname = file_path + "Exp_" + str(nfft_val) + "_" + str(overlap_val) + "_" + str(brange_val)
# Read and format data - all gestures
gd = Read_Data.GestureData(gest_set=1)
print("Reading data")
x, y, user, input_shape, lab_enc = gd.compile_data(nfft=nfft_val, overlap=overlap_val,
brange=brange_val, keras_format=True,
plot_spectogram=False,
baseline_format=False)
NUM_CLASSES = len(lab_enc.classes_)
print("NFFT_Val: ", nfft_val, "Overlap_Val: ", overlap_val, "Brange_Val:", brange_val)
print("Train the model")
train_val_hist = loso_gridSearch_cv(x, y, user,
lab_enc,
batch_size=64,
nb_epoch=nb_epoch,
file_path=fname)
plot_train_hist(train_val_hist, file_path=fname)
K.clear_session()
def airware_baseline_data():
gd = Read_Data.GestureData(gest_set=1)
x, y, user, lab_enc = gd.compile_data(nfft=4096,
overlap=0.5,
brange=16,
keras_format=False,
plot_spectogram=False,
baseline_format=True)
return x, y, user, lab_enc
def airware_data(gest_set=1):
param_list = HyperParams()
gd = Read_Data.GestureData(gest_set=gest_set)
x, y, user, input_shape, lab_enc = gd.compile_data(
nfft=param_list.NFFT_VAL,
overlap=param_list.OVERLAP,
brange=param_list.BRANGE,
max_seconds=2.5,
keras_format=True,
plot_spectogram=False,
baseline_format=False)
num_classes = len(lab_enc.classes_)
param_list.input_shape = input_shape
param_list.num_classes = num_classes
return x, y, user, lab_enc, param_list
def airware_data():
param_list = HyperParams()
gd = Read_Data.GestureData(gest_set=1)
x, y, user, input_shape, lab_enc = gd.compile_data(
nfft=param_list.NFFT_VAL,
overlap=param_list.OVERLAP,
brange=param_list.BRANGE,
max_seconds=2.5,
keras_format=True,
plot_spectogram=False,
baseline_format=False)
x_train, x_test, y_train, y_test = train_test_split(x,
y,
test_size=0.3,
stratify=None,
random_state=234)
num_classes = len(lab_enc.classes_)
param_list.input_shape = input_shape
param_list.num_classes = num_classes
return x_train, x_test, y_train, y_test, param_list
torch.cuda.manual_seed_all(seed)
torch.backends.cudnn.enabled = False
torch.backends.cudnn.benchmark = False
torch.backends.cudnn.deterministic = True
seed_index = seed_index + 1
print('Reading Data. with seeds: ' + str(seed))
data = Read_Data(column_names=args,
data_name=args.data_name,
data_file=args.data_file,
train_ratio=args.training_ratio,
test_val_ratio=args.test_val_ratio,
med_flag=args.medical_data,
weight_flag=args.weight_class,
seed=seed,
upos=args.upos,
umed=args.umed,
uad=args.uad,
qk=args.question_knowledge,
cluster_size=args.cluster_number,
wordnet=args.wordnet,
class_number=args.class_number)
print('Building Embedding Matrix')
embedding = Create_Embedding(
file_path=args.embd_file,
embd_file_mimic=args.embd_file_mimic,
word2index=data.word2index,
custom=args.embedding_flag,