def create_plot_for_measurements(path, dir=None):
if dir is None:
full_path = os.path.join(path, 'plots_upper_limit')
elif type(dir) is str:
full_path = os.path.join(path, dir)
if not os.path.exists(full_path):
os.mkdir(full_path)
i = 0
yell_colors = [(0.4, 0.4, 0, c) for c in np.linspace(0, 1, 1000)]
cmapred = mcolors.LinearSegmentedColormap.from_list('mycmap',
yell_colors,
N=5)
green_colors = [(173 / 255, 1, 47 / 255, c)
for c in np.linspace(0, 1, 1000)]
cmapblue = mcolors.LinearSegmentedColormap.from_list('mycmap',
green_colors,
N=5)
fp = 1000
f = np.linspace(0, 60000, fp)
t = np.linspace(0, 2.4, int(fp * 2.4))
Sxx = np.full((len(f), len(t)), 0)
plt.figure(figsize=(12, 6))
plt.pcolormesh(t, f, Sxx)
# fig = plt.figure(figsize=(8,8))
# ax = fig.add_subplot(111)
for file in os.listdir(path):
# if i == 4:
# break
if '.csv' in file:
time, signal = open_file(os.path.join(path, file))
# time = get_real_time(time)
if time is not None:
f, t, Sxx = get_stft(time, signal)
png_file = os.path.join(full_path,
file.replace('.csv', '.png'))
#
# plt.plot(Sxx)
# ax.pcolormesh(t, f, np.abs(Sxx), rasterized=
plt.contourf(t, f, np.abs(Sxx), cmap=cmapred)
plt.pcolormesh(t, f, np.abs(Sxx), cmap=cmapblue)
# plt.hold(True)
print(f" {png_file} plotted.")
# i += 1
# x0, x1 = ax.get_xlim()
# y0, y1 = ax.get_ylim()
# ax.imshow(img, extent=[x0, x1, y0, y1], aspect='auto')
# fig.savefig('/tmp/test.png')
plt.title('Spektrogram transformaty Gabora')
plt.ylabel('Częstotliwość [Hz]')
plt.xlabel('Czas [s]')
upper_limit = 60000
plt.ylim((0, upper_limit))
plt.savefig('full_stft_all_mes3.png')
# plt.show()
return
def create_plots_for_measuerements(path, dir=None):
if dir is None:
full_path = os.path.join(path, 'plots_upper_limit')
elif type(dir) is str:
full_path = os.path.join(path, dir)
if not os.path.exists(full_path):
os.mkdir(full_path)
for file in os.listdir(path):
if '.csv' in file:
time, signal = open_file(os.path.join(path, file))
# time = get_real_time(time)
if time is not None:
# f, t, Sxx = get_stft(time, signal)
png_file = os.path.join(full_path,
file.replace('.csv', '.png'))
plot_stft(time, signal, save=png_file, upper_limit=65000)
print(f"File {png_file} is saved.")
import numpy as np
from random import randint
import file_reader
import neural_network_keras
from neural_network import NeuralNetwork, get_accuracy, error
data_test = file_reader.open_file('mnist_test.csv')
data_train = file_reader.open_file('mnist_train.csv')
training_data = file_reader.normalize_data(data_train)
training_labels = file_reader.normalize_labels(data_train)
one_hot_train_labels = file_reader.label_to_one_hot(training_labels)
test_data = file_reader.normalize_data(data_test)
test_labels = file_reader.normalize_labels(data_test)
one_hot_test_labels = file_reader.label_to_one_hot(test_labels)
# print(one_hot_train_labels)
#
print('INITIALIZE NETWORK')
model = NeuralNetwork(training_data, one_hot_train_labels, 128, 0.5)
print('Number of Epochs: ')
epochs = int(input())
print('1.) Use built from scratch network')
print('2.) Use keras network')
option = int(input())
if option == 1:
min_dis = min_local
assign_to = i
final_cluster[assign_to][1].append(p)
#print("remining points:",len(initial_points)-total)
total_points = 0
for f in final_cluster:
total_points += len(f[1])
assert (total_points == len(initial_points)
), "diff length between final cluster and initial points!"
return final_cluster
'=============================================='
initial_points = fr.open_file()
initial_cluster = {}
initial_cluster[0] = initial_points
vi.show_scatter(initial_cluster)
sample_points = sample(initial_points)
small_cluster = {}
small_cluster[0] = sample_points
vi.show_scatter(small_cluster)
k = int(sys.argv[2])
sample_cluster = bottom_up(sample_points, k)
cluster_dicts = {}
if temp_dis < min_dis:
min_dis = temp_dis
tree0 = tree
tree1 = next_tree
n0 = tree0[0]
n1 = tree1[0]
sumx = n0[2] + n1[2]
sumy = n0[3] + n1[3]
sumn = n0[4] + n1[4]
new_node = (sumx / sumn, sumy / sumn, sumx, sumy, sumn)
new_tree = [new_node, tree0, tree1]
cluster_trees.remove(tree0)
cluster_trees.remove(tree1)
cluster_trees.append(new_tree)
return cluster_trees
def top_down():
pass
'===================================================================================='
test = [(0, 0), (1, 2), (2, 1), (4, 1), (5, 0), (5, 3)]
read = fr.open_file()
final = bottom_up(read, 20)
vi.show_scatter(final)