def test_eels_data_is_consistent_when_energy_offset_changes(self):
instrument = InstrumentDevice.Instrument("usim_stem_controller")
instrument.get_scan_data(
scan_base.ScanFrameParameters({
"size": (256, 256),
"pixel_time_us": 1,
"fov_nm": 10
}), 0)
instrument.validate_probe_position()
camera = instrument._get_camera_simulator("eels")
camera_size = camera._camera_shape
camera.noise.enabled = False
readout_area = Geometry.IntRect(origin=Geometry.IntPoint(),
size=camera_size)
binning_shape = Geometry.IntSize(1, 1)
# get the value at 200eV and ZLP offset of 0
instrument.ZLPoffset = 0
d = xd.sum(instrument.get_camera_data("eels", readout_area,
binning_shape, 0.01),
axis=0)
index200_0 = int(
d.dimensional_calibrations[-1].convert_from_calibrated_value(200))
value200_0 = d.data[index200_0]
# get the value at 200eV and ZLP offset of 100
instrument.ZLPoffset = 100
d = xd.sum(instrument.get_camera_data("eels", readout_area,
binning_shape, 0.01),
axis=0)
index200_100 = int(
d.dimensional_calibrations[-1].convert_from_calibrated_value(200))
value200_100 = d.data[index200_100]
self.assertEqual(int(value200_0 / 100), int(value200_100 / 100))
def execute(self, src1, src2, dark_area_region, crop_region, bin_spectrum, gain_image, gain_mode):
try:
xdata = src1.xdata
metadata = src1.metadata.copy()
data_shape = np.array(xdata.data.shape)
dark_area = np.rint(np.array(dark_area_region.bounds) * np.array((data_shape[:2], data_shape[:2]))).astype(np.int)
crop_area = np.rint(np.array(crop_region.bounds) * np.array((data_shape[2:], data_shape[2:]))).astype(np.int)
dark_image = xd.sum(xdata[dark_area[0, 0]:dark_area[0, 0]+dark_area[1, 0],
dark_area[0, 1]:dark_area[0, 1]+dark_area[1, 1],
crop_area[0, 0]:crop_area[0, 0]+crop_area[1, 0],
crop_area[0, 1]:crop_area[0, 1]+crop_area[1, 1]], axis=(0, 1))/(dark_area[1,0]*dark_area[1,1])
self.__new_xdata = xdata[..., crop_area[0, 0]:crop_area[0, 0]+crop_area[1, 0],
crop_area[0, 1]:crop_area[0, 1]+crop_area[1, 1]] - dark_image
current_gain_image_uuid = metadata.get('hardware_source', {}).get('current_gain_image')
current_gain_image = None
if current_gain_image_uuid:
current_gain_image = self.__api.library.get_data_item_by_uuid(uuid.UUID(current_gain_image_uuid))
if metadata.get('hardware_source', {}).get('is_gain_corrected'):
if gain_mode in ('custom', 'off') and current_gain_image:
if current_gain_image.xdata.data_shape == xdata.data_shape[2:]:
self.__new_xdata /= current_gain_image.xdata[crop_area[0, 0]:crop_area[0, 0]+crop_area[1, 0],
crop_area[0, 1]:crop_area[0, 1]+crop_area[1, 1]]
if ((gain_mode == 'auto' and not metadata.get('hardware_source', {}).get('is_gain_corrected') and current_gain_image) or
(gain_mode == 'custom' and gain_image)):
gain_xdata = gain_image[0].xdata if gain_mode == 'custom' else current_gain_image.xdata
if gain_xdata.data_shape == self.__new_xdata.data_shape[2:]:
self.__new_xdata *= gain_xdata
elif gain_xdata.data_shape == xdata.data_shape[2:]:
self.__new_xdata *= gain_xdata[crop_area[0, 0]:crop_area[0, 0]+crop_area[1, 0],
crop_area[0, 1]:crop_area[0, 1]+crop_area[1, 1]]
else:
raise ValueError('Shape of gain image has to match last two dimensions of input data.')
del gain_xdata
if bin_spectrum:
self.__new_xdata = xd.sum(self.__new_xdata, axis=2)
metadata['nion.dark_correction_4d.parameters'] = {'src1': src1._data_item.write_to_dict(),
'src2': src2._data_item.write_to_dict(),
'dark_area_region': dark_area_region._graphic.write_to_dict(),
'crop_region': crop_region._graphic.write_to_dict(),
'bottom_dark_region_region': crop_region._graphic.write_to_dict(),
'bin_spectrum': bin_spectrum,
'gain_image': gain_image[0].data_item.write_to_dict() if gain_image else None,
'gain_mode': gain_mode}
self.__new_xdata.metadata.update(metadata)
except Exception as e:
print(str(e))
import traceback
traceback.print_exc()
def test_eels_data_camera_current_is_consistent(self):
instrument = InstrumentDevice.Instrument("usim_stem_controller")
# set up the scan context; these are here temporarily until the scan context architecture is fully implemented
instrument._update_scan_context(Geometry.IntSize(256, 256),
Geometry.FloatPoint(), 10, 0.0)
instrument._set_scan_context_probe_position(
instrument.scan_context, Geometry.FloatPoint(0.5, 0.5))
# grab scan data
instrument.get_scan_data(
scan_base.ScanFrameParameters({
"size": (256, 256),
"pixel_time_us": 1,
"fov_nm": 10
}), 0)
instrument.validate_probe_position()
camera = instrument._get_camera_simulator("eels")
camera_size = camera._camera_shape
camera.noise.enabled = False
readout_area = Geometry.IntRect(origin=Geometry.IntPoint(),
size=camera_size)
binning_shape = Geometry.IntSize(1, 1)
# get the value at 200eV and ZLP offset of 0
instrument.ZLPoffset = -20
exposure_s = 0.01
d = xd.sum(instrument.get_camera_data("eels", readout_area,
binning_shape, exposure_s),
axis=0).data
# confirm it is a reasonable value
camera_current_pA = numpy.sum(
d) / exposure_s / instrument.counts_per_electron / 6.242e18 * 1e12
# print(f"current {camera_current_pA :#.2f}pA")
self.assertTrue(190 < camera_current_pA < 210)
def test_eels_data_thickness_is_consistent(self):
instrument = InstrumentDevice.Instrument("usim_stem_controller")
# use the flake sample
instrument.sample_index = 0
# set up the scan context; these are here temporarily until the scan context architecture is fully implemented
instrument._update_scan_context(Geometry.IntSize(256, 256),
Geometry.FloatPoint(), 10, 0.0)
instrument._set_scan_context_probe_position(
instrument.scan_context, Geometry.FloatPoint(0.5, 0.5))
# grab scan data
instrument.get_scan_data(
scan_base.ScanFrameParameters({
"size": (256, 256),
"pixel_time_us": 1,
"fov_nm": 10
}), 0)
instrument.validate_probe_position()
camera = instrument._get_camera_simulator("eels")
camera_size = camera._camera_shape
camera.noise.enabled = False
readout_area = Geometry.IntRect(origin=Geometry.IntPoint(),
size=camera_size)
binning_shape = Geometry.IntSize(1, 1)
# get the value at 200eV and ZLP offset of 0
instrument.ZLPoffset = -20
d = xd.sum(instrument.get_camera_data("eels", readout_area,
binning_shape, 0.01),
axis=0).data
# confirm it is a reasonable value
# print(measure_thickness(d))
self.assertTrue(0.40 < measure_thickness(d) < 1.00)
def execute(self, src):
try:
self.__new_xdata = xd.sum(src.xdata, axis=(2, 3))
except Exception as e:
import traceback
traceback.print_exc()
print(str(e))
raise
def acquire_multi_eels(interactive, api):
# first grab the stem controller object by asking the Registry
stem_controller = Registry.get_component("stem_controller")
# establish the EELS camera object and stop it if it is playing
eels_camera = stem_controller.eels_camera
eels_camera.stop_playing()
print(eels_camera.hardware_source_id)
# this table represents the acquisitions to be performed
# each entry is energy offset, exposure (milliseconds), and the number of frames to integrate
table = [
# energy offset, exposure(ms), N frames
(0, 100, 2),
(10, 100, 2),
#(250, 1000, 10),
#(0, 100, 5),
]
# this is the list of integrated spectra that will be the result of this script
spectra = list()
# this algorithm handles dark subtraction specially - so dark subtraction and gain normalization should
# be disabled in the camera settings; this algorithm will handle dark subtraction itself.
do_dark = True
do_gain = False
print("start taking data")
energy_offset_control = "EELS_MagneticShift_Offset" # for hardware EELS
# energy_offset_control = "EELS_MagneticShift_Offset" # for simulator
tolerance_factor_from_nominal = 1.0
timeout_for_confirmation_ms = 3000
for energy_offset_ev, exposure_ms, frame_count in table:
# for each table entry, set the drift tube loss to the energy offset
stem_controller.SetValAndConfirm(energy_offset_control,
energy_offset_ev,
tolerance_factor_from_nominal,
timeout_for_confirmation_ms)
# configure the camera to have the desired exposure
frame_parameters = eels_camera.get_current_frame_parameters()
frame_parameters["exposure_ms"] = exposure_ms
eels_camera.set_current_frame_parameters(frame_parameters)
# disable blanker
stem_controller.SetValAndConfirm("C_Blank", 0,
tolerance_factor_from_nominal,
timeout_for_confirmation_ms)
# acquire a sequence of images and discard it; this ensures a steady state
eels_camera.grab_sequence_prepare(frame_count)
eels_camera.grab_sequence(frame_count)
# acquire a sequence of images again, but now integrate the acquired images into a single image
eels_camera.grab_sequence_prepare(frame_count)
xdata = eels_camera.grab_sequence(frame_count)[0]
print(f"grabbed data of shape {xdata.data_shape}")
# extract the calibration info
counts_per_electron = xdata.metadata.get("hardware_source",
dict()).get(
"counts_per_electron", 1)
exposure_ms = xdata.metadata.get("hardware_source",
dict()).get("exposure", 1)
intensity_scale = xdata.intensity_calibration.scale / counts_per_electron / xdata.dimensional_calibrations[
-1].scale / exposure_ms / frame_count
# now sum the data in the sequence/time dimension. use xd.sum to automatically handle metadata such as calibration.
xdata = xd.sum(xdata, 0)
# if dark subtraction is enabled, perform another similar acquisition with blanker enabled and subtract it
if do_dark:
# enable blanker
stem_controller.SetValAndConfirm("C_Blank", 1,
tolerance_factor_from_nominal,
timeout_for_confirmation_ms)
# acquire a sequence of images and discard it; this ensures a steady state
eels_camera.grab_sequence_prepare(frame_count)
eels_camera.grab_sequence(frame_count)
# acquire a sequence of images again, but now integrate the acquired images into a single image
eels_camera.grab_sequence_prepare(frame_count)
dark_xdata = eels_camera.grab_sequence(frame_count)[0]
# sum it and subtract it from xdata
dark_xdata = xd.sum(dark_xdata, 0)
xdata = xdata - dark_xdata
print(f"subtracted dark data of shape {dark_xdata.data_shape}")
if do_gain:
# divide out the gain
gain_uuid = uuid.uuid4(
) # fill this in with the actual gain image uuid
gain = interactive.document_controller.document_model.get_data_item_by_uuid(
gain_uuid)
if gain is not None:
xdata = xdata / gain.xdata
# next sum the 2d data into a 1d spectrum by collapsing the y-axis (0th dimension)
# also configure the intensity calibration and title.
spectrum = xd.sum(xdata, 0)
spectrum.data_metadata._set_intensity_calibration(
Calibration.Calibration(scale=intensity_scale, units="e/eV/s"))
spectrum.data_metadata._set_metadata({
"title":
f"{energy_offset_ev}eV {int(exposure_ms*1000)}ms [x{frame_count}]"
})
# add it to the list of spectra
spectra.append(spectrum)
# disable blanking and return drift tube loss to 0.0eV
stem_controller.SetValAndConfirm("C_Blank", 0,
tolerance_factor_from_nominal,
timeout_for_confirmation_ms)
stem_controller.SetValAndConfirm(energy_offset_control, 0,
tolerance_factor_from_nominal,
timeout_for_confirmation_ms)
print("finished taking data")
# when multi display is available, we can combine the spectra into a single line plot display without
# padding the data; but for now, we need to use a single master data item where each row is the same length.
if len(spectra) > 0:
# define the padded spectra list
padded_spectra = list()
# extract calibration info
ev_per_channel = spectra[0].dimensional_calibrations[-1].scale
units = spectra[0].dimensional_calibrations[-1].units
min_ev = min([
spectrum.dimensional_calibrations[-1].convert_to_calibrated_value(
0) for spectrum in spectra
])
max_ev = max([
spectrum.dimensional_calibrations[-1].convert_to_calibrated_value(
spectrum.data_shape[-1]) for spectrum in spectra
])
# calculate what the length of the padded data will be
data_length = int((max_ev - min_ev) / ev_per_channel)
# for each spectra, pad it out to the appropriate length, putting the actual data in the proper range
for spectrum in spectra:
energy_offset_ev = int(
(spectrum.dimensional_calibrations[-1].
convert_to_calibrated_value(0) - min_ev) / ev_per_channel)
calibration_factor = spectrum.intensity_calibration.scale / spectra[
0].intensity_calibration.scale
data = numpy.zeros((data_length, ))
data[energy_offset_ev:energy_offset_ev +
spectrum.data_shape[-1]] = spectrum.data * calibration_factor
padded_spectrum = DataAndMetadata.new_data_and_metadata(
data, spectrum.intensity_calibration,
[Calibration.Calibration(min_ev, ev_per_channel, units)])
padded_spectra.append(padded_spectrum)
# stack all of the padded data together for display
master_xdata = xd.vstack(padded_spectra)
# show the data
window = api.application.document_windows[0]
data_item = api.library.create_data_item_from_data_and_metadata(
master_xdata)
legends = [s.metadata["title"] for s in spectra]
data_item.title = f"MultiEELS ({', '.join(legends)})"
window.display_data_item(data_item)
print("finished")
def execute(self, src):
self.__new_xdata = xd.sum(src.xdata, axis=(0, 1))
