def test_measure_focus(self):
"""
Test MeasureFocus
"""
data = hdf5.read_data(os.path.dirname(__file__) + "/grid_10x10.h5")
C, T, Z, Y, X = data[0].shape
data[0].shape = Y, X
input = data[0]
prev_res = autofocus.MeasureSEMFocus(input)
for i in range(1, 10, 1):
blur = ndimage.gaussian_filter(input, sigma=i)
res = autofocus.MeasureSEMFocus(blur)
self.assertGreater(prev_res, res)
prev_res = res
def read_timelapse(infn, emfn, fmfn):
"""
infn (str): pattern for input filename
emfn: sem output filename
fmfn: fluorescence output filename
"""
infiles = sorted(glob.glob(infn))
if not infiles:
raise ValueError("No file fitting '%s'" % (infn, ))
emdata = {} # timestamp -> tuple of info (X/Y, overlay X/Y)
fmdata = {} # timestamp -> tuple of info (X/Y)
emda_prev = emda0 = fmda_prev = fmda0 = None
for i, infl in enumerate(infiles):
logging.info("Processing %s (%d/%d)", infl, i + 1, len(infiles))
try:
# Read the file
reader = dataio.find_fittest_converter(infl)
das = reader.read_data(infl)
# Read the metadata (we expect one fluo image and one SEM image)
fmpos = empos = emda = fmda = fmfoc = emfoc = None
for da in das:
if model.MD_IN_WL in da.metadata: # Fluo image
fmpos = da.metadata[
model.
MD_POS] # this one has the overlay translation included
fmdate = da.metadata[model.MD_ACQ_DATE]
fmda = da
fmpxs = da.metadata[model.MD_PIXEL_SIZE]
fmfoc = autofocus.MeasureOpticalFocus(da)
else: # SEM
empos = da.metadata[model.MD_POS]
emdate = da.metadata[model.MD_ACQ_DATE]
emda = da
empxs = da.metadata[model.MD_PIXEL_SIZE]
emfoc = autofocus.MeasureSEMFocus(da)
# Overlay translation
ovlpos = fmpos[0] - empos[0], fmpos[1] - empos[1]
# Compute drift from first image and previous image
if i == 0:
emda0 = emda
fmda0 = fmda
emdriftm = 0, 0
empdriftm = 0, 0
fmdriftm = 0, 0
fmpdriftm = 0, 0
else:
emdrift = drift.CalculateDrift(emda0, emda, 10) # in pixels
emdriftm = emdrift[0] * empxs[0], emdrift[1] * empxs[1]
logging.info("Computed total EM drift of %s px = %s m",
emdrift, emdriftm)
empdrift = drift.CalculateDrift(emda_prev, emda,
10) # in pixels
empdriftm = empdrift[0] * empxs[0], empdrift[1] * empxs[1]
logging.info("Computed previous EM drift of %s px = %s m",
empdrift, empdriftm)
fmdrift = drift.CalculateDrift(fmda0, fmda, 10) # in pixels
fmdriftm = fmdrift[0] * fmpxs[0], fmdrift[1] * fmpxs[1]
logging.info("Computed total FM drift of %s px = %s m",
fmdrift, fmdriftm)
fmpdrift = drift.CalculateDrift(fmda_prev, fmda,
10) # in pixels
fmpdriftm = fmpdrift[0] * fmpxs[0], fmpdrift[1] * fmpxs[1]
logging.info("Computed previous FM drift of %s px = %s m",
fmpdrift, fmpdriftm)
emdata[emdate] = (empos[0], empos[1], ovlpos[0], ovlpos[1],
emdriftm[0], emdriftm[1], empdriftm[0],
empdriftm[1], emfoc)
fmdata[fmdate] = (fmpos[0], fmpos[1], fmdriftm[0], fmdriftm[1],
fmpdriftm[0], fmpdriftm[1], fmfoc)
emda_prev = emda
fmda_prev = fmda
except KeyboardInterrupt:
logging.info("Closing after only %d images processed", i)
return
except Exception:
logging.exception("Failed to read %s", infl)
# export the data
logging.info("Exporting the data...")
export_csv(
emdata, emfn,
"timestamp, posx, posy, overlay x, overlay y, total drift x, total drift y, prev drift x, prev drift y, focus level\n"
)
export_csv(
fmdata, fmfn,
"timestamp, posx, posy, total drift x, total drift y, prev drift x, prev drift y, focus level\n"
)
return 0