def _cfcFiltSuro(xPha, xAmp, surJob, self):
"""SUP: Get the cfc and surrogates
The function return:
- The unormalized cfc
- All the surrogates (for pvalue)
- The mean of surrogates (for normalization)
- The deviation of surrogates (for normalization)
"""
# Check input variables :
npts, ntrial = xPha.shape
W = self._window
nwin = len(W)
# Get the filter for phase/amplitude properties :
phaMeth = self._pha.get(self._sf, self._pha.f, self._npts)
ampMeth = self._amp.get(self._sf, self._amp.f, self._npts)
# Filt the phase and amplitude :
xPha = self._pha.apply(xPha, phaMeth)
xAmp = self._amp.apply(xAmp, ampMeth)
# Extract phase of amplitude for PLV method:
if self.Id[0] in ['4']:
for a in range(xAmp.shape[0]):
for t in range(xAmp.shape[2]):
xAmp[a, :, t] = np.angle(hilbert(np.ravel(xAmp[a, :, t])))
# 2D loop trick :
claIdx, listWin, listTrial = list2index(nwin, ntrial)
# Get the unormalized cfc :
uCfc = [
_cfcGet(np.squeeze(xPha[:, W[k[0]][0]:W[k[0]][1], k[1]]),
np.squeeze(xAmp[:, W[k[0]][0]:W[k[0]][1], k[1]]), self.Id,
self._nbins) for k in claIdx
]
uCfc = np.array(groupInList(uCfc, listWin))
# Run surogates on each window :
if (self.n_perm != 0) and (self.Id[0] is not '5') and (self.Id[1]
is not '0'):
Suro = Parallel(n_jobs=surJob)(
delayed(_cfcGetSuro)(xPha[:,
k[0]:k[1], :], xAmp[:,
k[0]:k[1], :], self.
Id, self.n_perm, self._nbins, self._matricial)
for k in self._window)
mSuro = [np.mean(k, 3) for k in Suro]
stdSuro = [np.std(k, 3) for k in Suro]
else:
Suro, mSuro, stdSuro = None, None, None
return uCfc, Suro, mSuro, stdSuro
def _fit(x, y, clf, cv, mf, grp, center, n_jobs):
"""Sub function for fitting
"""
# Check the inputs size :
x, y = checkXY(x, y, mf, grp, center)
rep, nfeat = len(cv), len(x)
# Tricks : construct a list of tuple containing the index of
# (repetitions,features) & loop on it. Optimal for parallel computing :
claIdx, listRep, listFeat = list2index(rep, nfeat)
# Run the classification :
cvs = Parallel(n_jobs=n_jobs)(
delayed(_cvscore)(x[k[1]], y, clone(clf), cv[k[0]]) for k in claIdx)
da, y_true, y_pred = zip(*cvs)
# Reconstruct elements :
da = np.array(groupInList(da, listFeat))
y_true = groupInList(y_true, listFeat)
y_pred = groupInList(y_pred, listFeat)
return da, x, y, y_true, y_pred
def _fit(x, y, clf, cv, mf, grp, center, n_jobs):
"""Sub function for fitting
"""
# Check the inputs size :
x, y = checkXY(x, y, mf, grp, center)
rep, nfeat = len(cv), len(x)
# Tricks : construct a list of tuple containing the index of
# (repetitions,features) & loop on it. Optimal for parallel computing :
claIdx, listRep, listFeat = list2index(rep, nfeat)
# Run the classification :
cvs = Parallel(n_jobs=n_jobs)(delayed(_cvscore)(
x[k[1]], y, clone(clf), cv[k[0]]) for k in claIdx)
da, y_true, y_pred = zip(*cvs)
# Reconstruct elements :
da = np.array(groupInList(da, listFeat))
y_true = groupInList(y_true, listFeat)
y_pred = groupInList(y_pred, listFeat)
return da, x, y, y_true, y_pred
def _cfcFiltSuro(xPha, xAmp, surJob, self):
"""SUP: Get the cfc and surrogates
The function return:
- The unormalized cfc
- All the surrogates (for pvalue)
- The mean of surrogates (for normalization)
- The deviation of surrogates (for normalization)
"""
# Check input variables :
npts, ntrial = xPha.shape
W = self._window
nwin = len(W)
# Get the filter for phase/amplitude properties :
phaMeth = self._pha.get(self._sf, self._pha.f, self._npts)
ampMeth = self._amp.get(self._sf, self._amp.f, self._npts)
# Filt the phase and amplitude :
xPha = self._pha.apply(xPha, phaMeth)
xAmp = self._amp.apply(xAmp, ampMeth)
# Extract phase of amplitude for PLV method:
if self.Id[0] in ['4']:
for a in range(xAmp.shape[0]):
for t in range(xAmp.shape[2]):
xAmp[a, :, t] = np.angle(hilbert(np.ravel(xAmp[a, :, t])))
# 2D loop trick :
claIdx, listWin, listTrial = list2index(nwin, ntrial)
# Get the unormalized cfc :
uCfc = [_cfcGet(np.squeeze(xPha[:, W[k[0]][0]:W[k[0]][1], k[1]]),
np.squeeze(xAmp[:, W[k[0]][0]:W[k[0]][1], k[1]]),
self.Id, self._nbins) for k in claIdx]
uCfc = np.array(groupInList(uCfc, listWin))
# Run surogates on each window :
if (self.n_perm != 0) and (self.Id[0] is not '5') and (self.Id[1] is not '0'):
Suro = Parallel(n_jobs=surJob)(delayed(_cfcGetSuro)(
xPha[:, k[0]:k[1], :], xAmp[:, k[0]:k[1], :],
self.Id, self.n_perm, self._nbins, self._matricial) for k in self._window)
mSuro = [np.mean(k, 3) for k in Suro]
stdSuro = [np.std(k, 3) for k in Suro]
else:
Suro, mSuro, stdSuro = None, None, None
return uCfc, Suro, mSuro, stdSuro