f,
datainfo.datafiles[0],
sensor,
nclusters,
globalc=args.globalclust)
data = datainfo.get_peaks_resample_PCA(f,
datainfo.datafiles[0],
sensor)
for i in range(nclusters):
variance[i] = np.std(data[labels == i], axis=0)
lsignals = []
mhisto = np.zeros((len(datainfo.datafiles) // batches, nclusters))
cbatch = [c for c in enumerate(datainfo.datafiles)]
lbatch = batchify(cbatch, batches)
for nf, btch in enumerate(lbatch):
npeaks = 0
histo = np.zeros(nclusters)
for _, dfile in btch:
labels = datainfo.compute_peaks_labels(
f, dfile, sensor, nclusters, globalc=args.globalclust)
npeaks += len(labels)
for i in labels:
mhisto[nf, i] += 1.0
mhisto[nf] /= npeaks
matplotlib.rcParams.update({'font.size': 20})
fig = plt.figure()
fig.suptitle(sensorname, fontsize=50)
f = datainfo.open_experiment_data(mode='r')
for sensor, nclusters in zip(datainfo.sensors, datainfo.clusters):
print(sensor, nclusters)
if args.globalclust:
centroids = datainfo.get_peaks_global_clustering_centroids(f, sensor, nclusters)
else:
centroids = datainfo.get_peaks_clustering_centroids(f, datainfo.datafiles[0], sensor, nclusters)
lsignals = []
mhisto = np.zeros((len(datainfo.datafiles)//batches, nclusters))
cbatch = [c for c in enumerate(datainfo.datafiles)]
lbatch = batchify(cbatch, batches)
for nf, btch in enumerate(lbatch):
npeaks = 0
histo = np.zeros(nclusters)
for _, dfile in btch:
labels = datainfo.compute_peaks_labels(f, dfile, sensor, nclusters, globalc=args.globalclust)
npeaks += len(labels)
for i in labels:
mhisto[nf, i] += 1.0
mhisto[nf] /= npeaks
matplotlib.rcParams.update({'font.size': 25})
fig = plt.figure()
fig.set_figwidth(24)
fig.set_figheight(12)
lexperiments = ['e150514']
mbasal = 'meanfirst' # 'alternative'
altsmooth = False
args.wavy = True
args.extra = True
print('Begin Smoothing: ', time.ctime())
for expname in lexperiments:
datainfo = experiments[expname]
if not args.extra:
lsensors = datainfo.sensors
else:
lsensors = datainfo.extrasensors
batches = batchify([i for i in product(datainfo.datafiles, lsensors)], njobs)
if 'recenter' in datainfo.peaks_smooth:
# If recenter is true a subwindow of the data has to be indicated to be able to re-crop the signal
recenter = datainfo.peaks_smooth['recenter']
wtsel = datainfo.peaks_smooth['wtsel']
else:
recenter = False
wtsel = None
for batch in batches:
# Paralelize PCA computation
res = Parallel(n_jobs=-1)(
delayed(do_the_job)(dfile, sensor, recenter=False, wtsel=None, clean=False, mbasal=mbasal, alt_smooth=altsmooth, wavy=args.wavy) for dfile, sensor in batch)
mbasal = 'meanfirst' # 'globalmeanfirst' # 'alternative'
args.altsmooth = False
args.wavy = False
args.extra = False
args.pca = 0.98
print('Begin Smoothing: ', time.ctime())
for expname in lexperiments:
datainfo = experiments[expname]
if not args.extra:
lsensors = datainfo.sensors
else:
lsensors = datainfo.extrasensors
batches = batchify([i for i in product(datainfo.datafiles, lsensors)],
njobs)
if 'recenter' in datainfo.peaks_smooth:
# If recenter is true a subwindow of the data has to be indicated to be able to re-crop the signal
recenter = datainfo.peaks_smooth['recenter']
wtsel = datainfo.peaks_smooth['wtsel']
else:
recenter = False
wtsel = None
for batch in batches:
# Paralelize PCA computation
res = Parallel(n_jobs=-1)(
delayed(do_the_job)(dfile,
sensor,
recenter=False,