def fit_ising_matlab(n, X):
X = _convert_2d_array_matlab(np.array(X).astype(np.int), n)
spikes = Spikes(spikes = X)
learner = Learner(spikes)
learner.learn_from_spikes(spikes, window_size = 1)
J = learner.network.J
theta = learner.network.theta
return J.ravel().tolist(), theta.ravel().tolist()
def sample_from_model(self,
J=None,
theta=None,
trials=None,
reshape=False):
"""
Returns new spikes object with iid Ising spike trains:
(with Ising model determined by learning with MPF)
.. warning:
MIGHT NOT BE WORKING PROPERLY!
Parameters
----------
J : Type, optional
Description (default None)
theta : Type, optional
Description (default None)
trials : Type, optional
Description (default None)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
trials = trials or range(self._original_spikes.T)
X = np.zeros(
(len(trials), self._original_spikes.N, self._original_spikes.M))
learner = Learner(spikes=self._original_spikes)
no_net = False
if J is None or theta is None:
no_net = True
for c, t in enumerate(trials):
if no_net:
learner.learn_from_spikes(window_size=1, trials=[t])
J = learner._network.J
theta = learner._network.theta
X[c, :, :] = sample_from_ising(J, theta, self._original_spikes.M)
return Spikes(spikes=X)
def sample_from_model(self, J=None, theta=None, trials=None, reshape=False):
"""
Returns new spikes object with iid Ising spike trains:
(with Ising model determined by learning with MPF)
.. warning:
MIGHT NOT BE WORKING PROPERLY!
Parameters
----------
J : Type, optional
Description (default None)
theta : Type, optional
Description (default None)
trials : Type, optional
Description (default None)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
trials = trials or range(self._original_spikes.T)
X = np.zeros((len(trials), self._original_spikes.N, self._original_spikes.M))
learner = Learner(spikes=self._original_spikes)
no_net = False
if J is None or theta is None:
no_net = True
for c, t in enumerate(trials):
if no_net:
learner.learn_from_spikes(window_size=1, trials=[t])
J = learner._network.J
theta = learner._network.theta
X[c, :, :] = sample_from_ising(J, theta, self._original_spikes.M)
return Spikes(spikes=X)
def fit(self, trials=None, remove_zeros=True, reshape=False):
"""
Missing documentation
Parameters
----------
trials : Type, optional
Description (default None)
remove_zeros : bool, optional
Remove all 0 training patterns (default True)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
# TODO: take care of remove_zeros
self._sample_spikes = self.sample_from_model(trials=trials, reshape=reshape)
self._learner = Learner(spikes=self._sample_spikes)
self._learner.learn_from_spikes(remove_zeros=remove_zeros)
def fit(self, trials=None, remove_zeros=False, reshape=False):
"""
Missing documentation
Parameters
----------
trials : Type, optional
Description (default None)
remove_zeros : bool, optional
Description (default False)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
# TODO: take care of remove_zeros
self._sample_spikes = self.sample_from_model(trials=trials, reshape=reshape)
self._learner = Learner(spikes=self._sample_spikes)
self._learner.learn_from_spikes(remove_zeros=remove_zeros)
class SpikeModel(Restoreable, object):
"""
Generic model of spikes (and stimulus).
Parameters
----------
spikes : Type, optional
Description (default None)
stimulus : Type, optional
Description (default None)
window_size : int, optional
Description (default 1)
learner : Type, optional
Description (default None)
Parameters
spikes: spikes to model
stimulus: corresp stimulus if existent
window_size: length of time window in binary bins
Returns
-------
Value : Type
Description
"""
_SAVE_ATTRIBUTES_V1 = ['_window_size', '_learn_time']
_SAVE_VERSION = 1
_SAVE_TYPE = 'SpikeModel'
_INTERNAL_OBJECTS = zip([
Spikes, Spikes, Spikes, PatternsRaw, PatternsHopfield, Stimulus,
Learner
], [
'_original_spikes', '_sample_spikes', '_hopfield_spikes',
'_raw_patterns', '_hopfield_patterns', '_stimulus', '_learner'
], [
'spikes_original', 'spikes_sample', 'spikes_hopfield', 'patterns_raw',
'patterns_hopfield', 'stimulus', 'learner'
])
def __init__(self,
spikes=None,
stimulus=None,
window_size=1,
learner=None):
object.__init__(self)
Restoreable.__init__(self)
self._stimulus = stimulus
self._window_size = window_size
self._learner = learner or None
self._original_spikes = spikes
self._learn_time = None
self._sample_spikes = None
self._raw_patterns = None
self._hopfield_patterns = None
self._hopfield_spikes = None
@property
def stimulus(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._stimulus
@property
def window_size(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._window_size
@property
def learner(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._learner
@property
def original_spikes(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._original_spikes
@property
def learn_time(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._learn_time
@property
def sample_spikes(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._sample_spikes
@property
def raw_patterns(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._raw_patterns
@property
def hopfield_patterns(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._hopfield_patterns
@property
def hopfield_spikes(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._hopfield_spikes
def fit(self, trials=None, remove_zeros=False, reshape=False):
"""
Missing documentation
Parameters
----------
trials : Type, optional
Description (default None)
remove_zeros : bool, optional
Description (default False)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
# TODO: take care of remove_zeros
self._sample_spikes = self.sample_from_model(trials=trials,
reshape=reshape)
self._learner = Learner(spikes=self._sample_spikes)
self._learner.learn_from_spikes(remove_zeros=remove_zeros)
def chomp(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
hdlog.debug("Chomping samples from model")
self._raw_patterns = PatternsRaw(save_sequence=True)
self._raw_patterns.chomp_spikes(spikes=self._sample_spikes)
hdlog.info("Raw: %d-bit, %d patterns" %
(self._sample_spikes.N, len(self._raw_patterns)))
hdlog.debug(
"Chomping dynamics (from network learned on the samples) applied to samples"
)
self._hopfield_patterns = PatternsHopfield(learner=self._learner,
save_sequence=True)
self._hopfield_patterns.chomp_spikes(spikes=self._sample_spikes)
hdlog.info("Hopfield: %d-bit, %d patterns" %
(self._sample_spikes.N, len(self._hopfield_patterns)))
# print "Before dynamics:"
# print self.sample_spikes.spikes
# print "Applied dynamics:"
self._hopfield_spikes = self._hopfield_patterns.apply_dynamics(
spikes=self._sample_spikes, reshape=True)
# ma_err = np.abs(self.sample_spikes.spikes - hop_model_spikes).mean()
# print hop_model_spikes
# print "Mean prediction: %1.4f/1.0 (vs guess zero: %1.4f)" % (
# (1 - ma_err), 1 - np.abs(self.sample_spikes.spikes).mean())
# # distortion
# self.sample_spikes
def distinct_patterns_over_windows(self,
window_sizes=None,
trials=None,
save_couplings=False,
remove_zeros=False):
"""
Returns tuple: counts, entropies [, couplings]
counts, entropies: arrays of size 2 x T x WSizes
(0: empirical from model sample, 1: dynamics from learned model on sample)
Parameters
----------
window_sizes : Type, optional
Description (default None)
trials : Type, optional
Description (default None)
save_couplings : bool, optional
Description (default False)
remove_zeros : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
if window_sizes is None:
window_sizes = [1]
trials = trials or range(self._original_spikes.T)
counts = np.zeros((2, len(trials), len(window_sizes)))
#entropies = np.zeros((2, len(trials), len(window_sizes)))
couplings = {}
tot_learn_time = 0
for ws, window_size in enumerate(window_sizes):
couplings[window_size] = []
for c, trial in enumerate(trials):
hdlog.info("Trial %d | ws %d" % (trial, window_size))
self._window_size = window_size
t = now()
self.fit(trials=[trial], remove_zeros=remove_zeros)
diff = now() - t
hdlog.info("[%1.3f min]" % (diff / 60.))
tot_learn_time += diff
if save_couplings:
couplings[ws].append(self._learner.network.J.copy())
self.chomp()
#entropies[0, c, ws] = self._raw_patterns.entropy()
counts[0, c, ws] = len(self._raw_patterns)
#entropies[1, c, ws] = self._hopfield_patterns.entropy()
counts[1, c, ws] = len(self._hopfield_patterns)
hdlog.info("Total learn time: %1.3f mins" % (tot_learn_time / 60.))
self._learn_time = tot_learn_time
if save_couplings:
return counts, couplings
return counts
def sample_from_model(self, trials=None, reshape=False):
"""
Missing documentation
Parameters
----------
trials : Type, optional
Description (default None)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
return self._original_spikes.to_windowed(window_size=self._window_size,
trials=trials,
reshape=reshape)
def save(self, folder_name='spikes_model'):
"""
saves as npz's: network, params, spikes file_name
Parameters
----------
folder_name : str, optional
Description (default 'spikes_model')
Returns
-------
Value : Type
Description
"""
super(SpikeModel, self)._save('spikes_model.npz',
self._SAVE_ATTRIBUTES_V1,
self._SAVE_VERSION,
has_internal=True,
folder_name=folder_name,
internal_objects=self._INTERNAL_OBJECTS)
@classmethod
def load(cls, folder_name='spikes_model', load_extra=False):
"""
Missing documentation
Parameters
----------
folder_name : str, optional
Description (default 'spikes_model')
load_extra : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
# TODO: document
return super(SpikeModel,
cls)._load('spikes_model.npz',
has_internal=True,
folder_name=folder_name,
internal_objects=cls._INTERNAL_OBJECTS,
load_extra=load_extra)
def _load_v1(self, contents, load_extra=False):
# internal function to load v1 file format
hdlog.debug('Loading SpikeModel, format version 1')
return Restoreable._load_attributes(self, contents,
self._SAVE_ATTRIBUTES_V1)
# representation
def __repr__(self):
return '<SpikeModel: {s}, window size {ws}>'.\
format(s=repr(self.original_spikes), ws=self.window_size)
def test_patterns_hopfield(self):
file_contents = np.load(os.path.join(os.path.dirname(__file__), 'test_data/tiny_spikes.npz'))
spikes = Spikes(file_contents[file_contents.keys()[0]])
learner = Learner(spikes)
learner.learn_from_spikes(spikes)
patterns = PatternsHopfield(learner=learner)
patterns.chomp_spikes(spikes)
# print spikes.spikes
self.assertEqual(len(patterns), 3)
# print "%d fixed-points (entropy H = %1.3f):" % (len(patterns), patterns.entropy())
# print map(patterns.pattern_for_key, patterns.counts.keys())
patterns.save(os.path.join(self.TMP_PATH, 'patterns'))
patterns2 = PatternsHopfield.load(os.path.join(self.TMP_PATH, 'patterns'))
self.assertTrue(isinstance(patterns2, PatternsHopfield))
self.assertEqual(len(patterns2), 3)
self.assertEqual(len(patterns2.mtas), 3)
self.assertEqual(len(patterns2.mtas_raw), 3)
learner.learn_from_spikes(spikes, window_size=3)
patterns = PatternsHopfield(learner=learner)
patterns.chomp_spikes(spikes, window_size=3)
# print spikes.spikes
# print patterns.counts
self.assertEqual(len(patterns), 4)
# print "%d fixed-points (entropy H = %1.3f):" % (len(patterns), patterns.entropy())
# for x in patterns.list_patterns(): print x
spikes_arr1 = np.array([[1, 0, 1], [0, 0, 1], [0, 1, 0]])
spikes = Spikes(spikes=spikes_arr1)
learner = Learner(spikes)
learner.learn_from_spikes(spikes)
# test recording fixed-points
file_contents = np.load(os.path.join(os.path.dirname(__file__), 'test_data/spikes_trials.npz'))
spikes = Spikes(file_contents[file_contents.keys()[0]])
learner = Learner(spikes)
learner.learn_from_spikes(spikes)
patterns = PatternsHopfield(learner, save_sequence=True)
patterns.chomp_spikes(spikes)
self.assertEqual(patterns._sequence, [0, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1])
file_contents = np.load(os.path.join(os.path.dirname(__file__), 'test_data/spikes_trials.npz'))
spikes = Spikes(file_contents[file_contents.keys()[0]])
learner = Learner(spikes)
learner.learn_from_spikes(spikes, window_size=2)
patterns = PatternsHopfield(learner, save_sequence=True)
patterns.chomp_spikes(spikes, window_size=2)
# print patterns.mtas
# print patterns.sequence
# for x in patterns.list_patterns(): print x
# print spikes.spikes
self.assertEqual(patterns._sequence, [0, 1, 2, 3, 0, 1, 4, 5, 6, 5, 7, 3])
# self.assertTrue(np.mean(patterns.pattern_to_binary_matrix(1) == [[0, 0], [0, 1], [1, 0]]))
# self.assertTrue(np.mean(patterns.pattern_to_mta_matrix(1) == [[0, 0], [0, 1], [1, .5]]))
hdlog.info(spikes._spikes)
hdlog.info(patterns.pattern_to_trial_raster(3))
class SpikeModel(Restoreable, object):
"""
Generic model of spikes (and stimulus).
Parameters
----------
spikes : Type, optional
Description (default None)
stimulus : Type, optional
Description (default None)
window_size : int, optional
Description (default 1)
learner : Type, optional
Description (default None)
Parameters
spikes: spikes to model
stimulus: corresp stimulus if existent
window_size: length of time window in binary bins
Returns
-------
Value : Type
Description
"""
_SAVE_ATTRIBUTES_V1 = ['_window_size', '_learn_time']
_SAVE_VERSION = 1
_SAVE_TYPE = 'SpikeModel'
_INTERNAL_OBJECTS = zip([Spikes, Spikes, Spikes, PatternsRaw, PatternsHopfield, Stimulus, Learner],
['_original_spikes', '_sample_spikes', '_hopfield_spikes',
'_raw_patterns', '_hopfield_patterns', '_stimulus', '_learner'],
['spikes_original', 'spikes_sample', 'spikes_hopfield',
'patterns_raw', 'patterns_hopfield', 'stimulus', 'learner'])
def __init__(self, spikes=None, stimulus=None, window_size=1, learner=None):
object.__init__(self)
Restoreable.__init__(self)
self._stimulus = stimulus
self._window_size = window_size
self._learner = learner or None
self._original_spikes = spikes
self._learn_time = None
self._sample_spikes = None
self._raw_patterns = None
self._hopfield_patterns = None
self._hopfield_spikes = None
@property
def stimulus(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._stimulus
@property
def window_size(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._window_size
@property
def learner(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._learner
@property
def original_spikes(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._original_spikes
@property
def learn_time(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._learn_time
@property
def sample_spikes(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._sample_spikes
@property
def raw_patterns(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._raw_patterns
@property
def hopfield_patterns(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._hopfield_patterns
@property
def hopfield_spikes(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
return self._hopfield_spikes
def fit(self, trials=None, remove_zeros=False, reshape=False):
"""
Missing documentation
Parameters
----------
trials : Type, optional
Description (default None)
remove_zeros : bool, optional
Description (default False)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
# TODO: take care of remove_zeros
self._sample_spikes = self.sample_from_model(trials=trials, reshape=reshape)
self._learner = Learner(spikes=self._sample_spikes)
self._learner.learn_from_spikes(remove_zeros=remove_zeros)
def chomp(self):
"""
Missing documentation
Returns
-------
Value : Type
Description
"""
hdlog.debug("Chomping samples from model")
self._raw_patterns = PatternsRaw(save_sequence=True)
self._raw_patterns.chomp_spikes(spikes=self._sample_spikes)
hdlog.info("Raw: %d-bit, %d patterns" % (
self._sample_spikes.N, len(self._raw_patterns)))
hdlog.debug("Chomping dynamics (from network learned on the samples) applied to samples")
self._hopfield_patterns = PatternsHopfield(learner=self._learner, save_sequence=True)
self._hopfield_patterns.chomp_spikes(spikes=self._sample_spikes)
hdlog.info("Hopfield: %d-bit, %d patterns" % (
self._sample_spikes.N, len(self._hopfield_patterns)))
# print "Before dynamics:"
# print self.sample_spikes.spikes
# print "Applied dynamics:"
self._hopfield_spikes = self._hopfield_patterns.apply_dynamics(spikes=self._sample_spikes, reshape=True)
# ma_err = np.abs(self.sample_spikes.spikes - hop_model_spikes).mean()
# print hop_model_spikes
# print "Mean prediction: %1.4f/1.0 (vs guess zero: %1.4f)" % (
# (1 - ma_err), 1 - np.abs(self.sample_spikes.spikes).mean())
# # distortion
# self.sample_spikes
def distinct_patterns_over_windows(self, window_sizes=None, trials=None, save_couplings=False, remove_zeros=False):
"""
Returns tuple: counts, entropies [, couplings]
counts, entropies: arrays of size 2 x T x WSizes
(0: empirical from model sample, 1: dynamics from learned model on sample)
Parameters
----------
window_sizes : Type, optional
Description (default None)
trials : Type, optional
Description (default None)
save_couplings : bool, optional
Description (default False)
remove_zeros : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
if window_sizes is None:
window_sizes = [1]
trials = trials or range(self._original_spikes.T)
counts = np.zeros((2, len(trials), len(window_sizes)))
#entropies = np.zeros((2, len(trials), len(window_sizes)))
couplings = {}
tot_learn_time = 0
for ws, window_size in enumerate(window_sizes):
couplings[window_size] = []
for c, trial in enumerate(trials):
hdlog.info("Trial %d | ws %d" % (trial, window_size))
self._window_size = window_size
t = now()
self.fit(trials=[trial], remove_zeros=remove_zeros)
diff = now() - t
hdlog.info("[%1.3f min]" % (diff / 60.))
tot_learn_time += diff
if save_couplings:
couplings[ws].append(self._learner.network.J.copy())
self.chomp()
#entropies[0, c, ws] = self._raw_patterns.entropy()
counts[0, c, ws] = len(self._raw_patterns)
#entropies[1, c, ws] = self._hopfield_patterns.entropy()
counts[1, c, ws] = len(self._hopfield_patterns)
hdlog.info("Total learn time: %1.3f mins" % (tot_learn_time / 60.))
self._learn_time = tot_learn_time
if save_couplings:
return counts, couplings
return counts
def sample_from_model(self, trials=None, reshape=False):
"""
Missing documentation
Parameters
----------
trials : Type, optional
Description (default None)
reshape : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
return self._original_spikes.to_windowed(window_size=self._window_size, trials=trials, reshape=reshape)
def save(self, folder_name='spikes_model'):
"""
saves as npz's: network, params, spikes file_name
Parameters
----------
folder_name : str, optional
Description (default 'spikes_model')
Returns
-------
Value : Type
Description
"""
super(SpikeModel, self)._save(
'spikes_model.npz', self._SAVE_ATTRIBUTES_V1, self._SAVE_VERSION,
has_internal=True, folder_name=folder_name, internal_objects=self._INTERNAL_OBJECTS)
@classmethod
def load(cls, folder_name='spikes_model', load_extra=False):
"""
Missing documentation
Parameters
----------
folder_name : str, optional
Description (default 'spikes_model')
load_extra : bool, optional
Description (default False)
Returns
-------
Value : Type
Description
"""
# TODO: document
return super(SpikeModel, cls)._load('spikes_model.npz', has_internal=True,
folder_name=folder_name,
internal_objects=cls._INTERNAL_OBJECTS,
load_extra=load_extra)
def _load_v1(self, contents, load_extra=False):
# internal function to load v1 file format
hdlog.debug('Loading SpikeModel, format version 1')
return Restoreable._load_attributes(self, contents, self._SAVE_ATTRIBUTES_V1)
# representation
def __repr__(self):
return '<SpikeModel: {s}, window size {ws}>'.\
format(s=repr(self.original_spikes), ws=self.window_size)
def test_basic(self):
file_contents = np.load(os.path.join(os.path.dirname(__file__), 'test_data/tiny_spikes.npz'))
spikes = Spikes(file_contents[file_contents.keys()[0]])
learner = Learner(spikes)
self.assertEqual(learner._spikes.N, 3)
learner.learn_from_spikes()
self.assertTrue(learner._network.J.mean() != 0.)
learner.learn_from_spikes(spikes)
self.assertTrue(learner._network.J.mean() != 0.)
learner.learn_from_spikes(spikes, window_size=3)
self.assertTrue(learner._network.J.mean() != 0.)
self.assertTrue(learner._network.J.shape == (9, 9))
learner._params['hi'] = 'chris'
learner.save(os.path.join(self.TMP_PATH, 'learner'))
learner2 = Learner.load(os.path.join(self.TMP_PATH, 'learner'))
self.assertEqual(learner2.params['hi'], 'chris')
self.assertEqual(learner2.window_size, 3)
self.assertTrue(learner2.network.J.mean() != 0.)
self.assertTrue(learner2.network.J.shape == (9, 9))
def test_basic(self):
file_contents = np.load(
os.path.join(os.path.dirname(__file__),
'test_data/tiny_spikes.npz'))
spikes = Spikes(file_contents[file_contents.keys()[0]])
learner = Learner(spikes)
self.assertEqual(learner._spikes.N, 3)
learner.learn_from_spikes()
self.assertTrue(learner._network.J.mean() != 0.)
learner.learn_from_spikes(spikes)
self.assertTrue(learner._network.J.mean() != 0.)
learner.learn_from_spikes(spikes, window_size=3)
self.assertTrue(learner._network.J.mean() != 0.)
self.assertTrue(learner._network.J.shape == (9, 9))
learner._params['hi'] = 'chris'
learner.save(os.path.join(self.TMP_PATH, 'learner'))
learner2 = Learner.load(os.path.join(self.TMP_PATH, 'learner'))
self.assertEqual(learner2.params['hi'], 'chris')
self.assertEqual(learner2.window_size, 3)
self.assertTrue(learner2.network.J.mean() != 0.)
self.assertTrue(learner2.network.J.shape == (9, 9))