def create_pixel_histogram(cspad_det,dark=None,Nmax=None):
cc = 0
npatt = len(cspad_dat.time[cc])
if Nmax==None or Nmax>npatt:
Nmax = npatt
chunks = cspad_dat.chunks(Nmax=Nmax)
histbins = []
histograms = []
for ch in chunks[cc]:
data = ch.data
if not dark==None:
data = data-dark
if histbins==[]:
datshp = np.shape(data[0])
Nel = np.prod(datshp)
histbins = np.empty(datshp,dtype=np.object_)
data = data.reshape(datshp[0],-1)
ind = 0
for n in range(Nel):
ind = np.ravel_index(n,datshp)
histbins[ind] = histEdges(data[ind])
for n in range(Nel):
ind = np.ravel_index(n,datshp)
if histograms ==[]:
histograms = np.empty(datshp,dtype=np.object_)
histpgrams[ind] = np.histogram(data[ind],histbins[ind])
else:
histograms[ind] += np.histogram(data[ind],histbins[ind])
return histograms,histbins
def predict(self, data):
label_ids = []
for inp in self.scaler.transform(data):
label_ids.append(np.ravel_index(compute_bmu(inp),(self.dim, self.dim)))
if self.labels != None:
return self.labels[label_ids]
else:
return np.array(label_ids)
