def setUp(self):
"""Sets up test fixtures"""
self.optimizer = LocalBestPSO
self.n_particles = 40
self.dimensions = 20
self.options = {
'c1': [1, 5],
'c2': [6, 10],
'k': [11, 15],
'w': [0.4, 0.9],
'p': 1
}
self.bounds = (np.array([-5, -5]), np.array([5, 5]))
self.iters = 10
self.n_selection_iters = 100
self.objective_func = sphere_func
self.g = RandomSearch(self.optimizer,
self.n_particles,
self.dimensions,
self.options,
self.objective_func,
self.iters,
self.n_selection_iters,
self.bounds,
velocity_clamp=None)
self.g_unbound = RandomSearch(self.optimizer,
self.n_particles,
self.dimensions,
self.options,
self.objective_func,
self.iters,
self.n_selection_iters,
bounds=None,
velocity_clamp=None)
def test_optimizer_type_fail(self):
"""Test that :code:`optimizer` of type :code:`string` raises
:code:`TypeError`"""
bad_optimizer = 'LocalBestPSO' # a string instead of a class object
with self.assertRaises(TypeError):
g = RandomSearch(bad_optimizer,
self.n_particles,
self.dimensions,
self.options,
self.objective_func,
self.iters,
self.n_selection_iters,
bounds=None,
velocity_clamp=None)
def test_n_selection_iters_type_fail(self):
"""Test that :code:`n_selection_iters` of type :code:`float` raises
:code:`TypeError`"""
bad_n_selection_iters = 100.0 # should be an int
with self.assertRaises(TypeError):
g = RandomSearch(self.optimizer,
self.n_particles,
self.dimensions,
self.options,
self.objective_func,
self.iters,
bad_n_selection_iters,
self.bounds,
velocity_clamp=None)
def main(image, data):
'''
Parametes:
image: path to image file
data: path to file that contain the message
p: QKrawtchoukMatrix p parameter
q: QKrawtchoukMatrix q parameter
'''
with open(data, 'r') as file:
message = file.read()
file.close()
cover_work = imageio.imread(image)
options = {
'c1': [0.1, 3],
'c2': [0.1, 3],
'w': [0.6, 2],
'k': [0, 10],
'p': 1
}
g = RandomSearch(
GlobalBestPSO,
n_particles=20,
dimensions=2,
options=options,
objective_func=lambda x: qKrawtchouk_hide(x, cover_work, message),
iters=10,
n_selection_iters=10,
bounds=([0, 0], [1, 1]))
best_score, best_options = g.search()
dct_score = dct_insert(cover_work, message)
logging.info('Score: {}'.format(best_score))
logging.info('DCT score {}'.format(dct_score))
logging.info(best_options)
def random_unbounded():
"""Returns a RandomSearch instance without bounds"""
options = {
'c1': [1, 5],
'c2': [6, 10],
'k': [11, 15],
'w': [0.4, 0.9],
'p': 1
}
return RandomSearch(LocalBestPSO,
n_particles=40,
dimensions=20,
options=options,
objective_func=sphere_func,
iters=10,
n_selection_iters=100,
bounds=None)
def random_bounded():
"""Returns a RandomSearch instance with bounds"""
bounds = (np.array([-5, -5]), np.array([5, 5]))
options = {
'c1': [1, 5],
'c2': [6, 10],
'k': [11, 15],
'w': [0.4, 0.9],
'p': 1
}
return RandomSearch(LocalBestPSO,
n_particles=40,
dimensions=20,
options=options,
objective_func=sphere_func,
iters=10,
n_selection_iters=100,
bounds=bounds)
def random_unbounded():
"""Returns a RandomSearch instance without bounds"""
options = {
"c1": [1, 5],
"c2": [6, 10],
"k": [11, 15],
"w": [0.4, 0.9],
"p": 1,
}
return RandomSearch(
LocalBestPSO,
n_particles=40,
dimensions=20,
options=options,
objective_func=sphere_func,
iters=10,
n_selection_iters=100,
bounds=None,
)
def random_bounded():
"""Returns a RandomSearch instance with bounds"""
bounds = (np.array([-5, -5]), np.array([5, 5]))
options = {
"c1": [1, 5],
"c2": [6, 10],
"k": [11, 15],
"w": [0.4, 0.9],
"p": 1,
}
return RandomSearch(
LocalBestPSO,
n_particles=40,
dimensions=20,
options=options,
objective_func=sphere,
iters=10,
n_selection_iters=100,
bounds=bounds,
)