def __init__(
self,
stop_event: LoggedEvent,
saving_signal: LoggedEvent,
wait_signal: LoggedEvent,
noise_subtraction_on: Event,
camera_queue: ArrayQueue,
saver_queue: ArrayQueue,
max_queue_size=1200,
):
super().__init__(name="dispatcher")
self.stop_event = stop_event
self.saving_signal = saving_signal.new_reference(self.logger)
self.wait_signal = wait_signal.new_reference(self.logger)
self.noise_subtraction_active = noise_subtraction_on.new_reference(
self.logger)
self.camera_queue = camera_queue
self.saver_queue = saver_queue
self.n_planes_queue = Queue()
self.viewer_queue = ArrayQueue(max_mbytes=max_queue_size)
self.calibration_ref_queue = ArrayQueue()
self.volume_buffer = None
self.calibration_ref = None
self.n_planes = 1
self.i_plane = 0
self.first_volume = True
def __init__(
self,
stop_event: LoggedEvent,
is_saving_event: LoggedEvent,
duration_queue: Queue,
max_queue_size=2000,
):
super().__init__(name="saver")
self.stop_event = stop_event.new_reference(self.logger)
self.save_queue = ArrayQueue(max_mbytes=max_queue_size)
self.saving_signal = is_saving_event
self.saver_stopped_signal = LoggedEvent(self.logger,
SashimiEvents.SAVING_STOPPED)
self.saving = False
self.saving_parameter_queue = Queue()
self.save_parameters: Optional[SavingParameters] = SavingParameters()
self.i_in_chunk = 0
self.i_chunk = 0
self.i_plane = 0
self.i_volume = 0
self.n_volumes = 10
self.current_data = None
self.saved_status_queue = Queue()
self.frame_shape = None
self.dtype = np.uint16
self.duration_queue = duration_queue
self.notifier = notifiers[conf["notifier"]]
class ImageReconstructor(Process):
def __init__(self, data_in_queue, stop_event, max_mbytes_queue=300):
super().__init__()
self.data_in_queue = data_in_queue
self.parameter_queue = Queue()
self.stop_event = stop_event
self.output_queue = ArrayQueue(max_mbytes_queue)
self.scanning_parameters = None
self.waveform = None
def run(self):
while not self.stop_event.is_set():
try:
self.scanning_parameters = self.parameter_queue.get(
timeout=0.001)
self.waveform = compute_waveform(self.scanning_parameters)
except Empty:
pass
try:
image = self.data_in_queue.get(timeout=0.001)
self.output_queue.put(
scanning_patterns.reconstruct_image_pattern(
np.roll(image,
self.scanning_parameters.mystery_offset),
*self.waveform,
(self.scanning_parameters.n_x,
self.scanning_parameters.n_y),
self.scanning_parameters.n_bin,
))
except Empty:
pass
def __init__(self, data_in_queue, stop_event, max_mbytes_queue=300):
super().__init__()
self.data_in_queue = data_in_queue
self.parameter_queue = Queue()
self.stop_event = stop_event
self.output_queue = ArrayQueue(max_mbytes_queue)
self.scanning_parameters = None
self.waveform = None
def __init__(self,
experiment_start_event,
duration_queue,
max_queuesize=200):
super().__init__()
self.data_queue = ArrayQueue(max_mbytes=max_queuesize)
self.parameter_queue = Queue()
self.stop_event = Event()
self.experiment_start_event = experiment_start_event
self.scanning_parameters = ScanningParameters()
self.new_parameters = copy(self.scanning_parameters)
self.duration_queue = duration_queue
self.n_frames_queue = Queue()
def __init__(self, *args, signal_start_rec, output_queue_mb=500, **kwargs):
super().__init__(*args, **kwargs)
self.save_queue = ArrayQueue(max_mbytes=output_queue_mb)
self.framestart_queue = Queue()
self.diagnostic_queue = Queue()
self.processing_parameters = MovementDetectionParameters(
).get_clean_values()
self.signal_start_rec = signal_start_rec
self.mem_use = 0
self.diagnostic_params = [
"n_pixels_difference",
"recording_state",
"n_images_in_buffer",
]
def __init__(self, n_items=100, timestamped=False, indexed=False,
n_mbytes=2, wait=0, test_full=False):
super().__init__()
self.source_array = IndexedArrayQueue(max_mbytes==n_mbytes) if indexed \
else (TimestampedArrayQueue(max_mbytes=n_mbytes) if timestamped \
else ArrayQueue(max_mbytes=n_mbytes))
self.n_items = n_items
self.wait = wait
self.test_full = test_full
def __init__(self, diagnostics=False):
super().__init__()
if diagnostics:
self.sequence_queue = Queue()
self.sequence_diagram = SequenceDiagram(self.sequence_queue,
"main")
self.experiment_start_event = Event()
self.scanning_settings = ScanningSettings()
self.experiment_settings = ExperimentSettings()
self.pause_after = False
self.parameter_tree = ParameterTree()
self.parameter_tree.add(self.scanning_settings)
self.parameter_tree.add(self.experiment_settings)
self.end_event = Event()
self.external_sync = ZMQcomm()
self.duration_queue = Queue()
self.scanner = Scanner(self.experiment_start_event,
duration_queue=self.duration_queue)
self.scanning_parameters = None
self.reconstructor = ImageReconstructor(self.scanner.data_queue,
self.scanner.stop_event)
self.save_queue = ArrayQueue(max_mbytes=800)
self.saver = StackSaver(self.scanner.stop_event, self.save_queue,
self.scanner.n_frames_queue)
self.save_status: Optional[SavingStatus] = None
self.motors = dict()
self.motors["x"] = MotorControl("COM6", axes="x")
self.motors["y"] = MotorControl("COM6", axes="y")
self.motors["z"] = MotorControl("COM6", axes="z")
self.power_controller = LaserPowerControl()
self.scanning_settings.sig_param_changed.connect(self.send_scan_params)
self.scanning_settings.sig_param_changed.connect(self.send_save_params)
self.scanner.start()
self.reconstructor.start()
self.saver.start()
self.open_setup()
self.paused = False
def __init__(
self,
stop_event: LoggedEvent,
wait_event: LoggedEvent,
exp_trigger_event: LoggedEvent,
camera_id=0,
max_queue_size=1200,
n_fps_frames=20,
):
super().__init__(name="camera")
# Queue to communicate
self.triggered_frame_rate_queue = Queue()
self.parameter_queue = Queue()
self.stop_event = stop_event.new_reference(self.logger)
self.wait_event = wait_event.new_reference(self.logger)
self.experiment_trigger_event = exp_trigger_event.new_reference(
self.logger)
self.image_queue = ArrayQueue(max_mbytes=max_queue_size)
self.camera_id = camera_id
self.camera = None
self.parameters = CamParameters()
self.framerate_rec = FramerateRecorder(n_fps_frames=n_fps_frames)
self.was_waiting = False
def __init__(
self,
stop_event: LoggedEvent,
waiting_event: LoggedEvent,
restart_event: LoggedEvent,
start_experiment_from_scanner=False,
n_samples_waveform=10000,
sample_rate=40000,
):
""""""
super().__init__(name="scanner")
self.stop_event = stop_event.new_reference(self.logger)
self.restart_event = restart_event.new_reference(self.logger)
self.wait_signal = waiting_event.new_reference(self.logger)
self.parameter_queue = Queue()
self.waveform_queue = ArrayQueue(max_mbytes=100)
self.n_samples = n_samples_waveform
self.sample_rate = sample_rate
self.parameters = ScanParameters()
self.start_experiment_from_scanner = start_experiment_from_scanner
class VolumeDispatcher(LoggingProcess):
"""
Parameters
----------
stop_event
saving_signal
wait_signal
noise_subtraction_on
camera_queue
saver_queue
max_queue_size
"""
def __init__(
self,
stop_event: LoggedEvent,
saving_signal: LoggedEvent,
wait_signal: LoggedEvent,
noise_subtraction_on: Event,
camera_queue: ArrayQueue,
saver_queue: ArrayQueue,
max_queue_size=1200,
):
super().__init__(name="dispatcher")
self.stop_event = stop_event
self.saving_signal = saving_signal.new_reference(self.logger)
self.wait_signal = wait_signal.new_reference(self.logger)
self.noise_subtraction_active = noise_subtraction_on.new_reference(
self.logger)
self.camera_queue = camera_queue
self.saver_queue = saver_queue
self.n_planes_queue = Queue()
self.viewer_queue = ArrayQueue(max_mbytes=max_queue_size)
self.calibration_ref_queue = ArrayQueue()
self.volume_buffer = None
self.calibration_ref = None
self.n_planes = 1
self.i_plane = 0
self.first_volume = True
def run(self):
self.logger.log_message("started")
while not self.stop_event.is_set():
self.receive_options()
self.get_frame()
self.close_log()
def process_frame(self, current_frame):
if self.calibration_ref is not None and self.noise_subtraction_active.is_set(
):
current_frame = neg_dif(current_frame, self.calibration_ref)
if self.first_volume or self.volume_buffer.shape[
1:3] != current_frame.shape:
self.volume_buffer = np.empty(
(self.n_planes, *current_frame.shape), dtype=np.uint16)
self.first_volume = False
self.logger.log_message(f"received plane {self.i_plane}")
self.volume_buffer[self.i_plane, :, :] = current_frame
self.i_plane += 1
if self.i_plane == self.n_planes:
self.fill_queues()
self.i_plane = 0
def fill_queues(self):
if self.viewer_queue.queue.qsize() < 3:
self.viewer_queue.put(self.volume_buffer)
else:
pass # volume has been dropped from the viewer
if self.saving_signal.is_set():
self.saver_queue.put(self.volume_buffer)
def get_frame(self):
if self.wait_signal.is_set():
self.logger.log_message("wait starting")
self.saver_queue.clear()
while self.wait_signal.is_set():
try:
_ = self.camera_queue.get(timeout=TIMEOUT_S)
except Empty:
pass
self.logger.log_message("wait over")
self.i_plane = 0
try:
current_frame = self.camera_queue.get(timeout=TIMEOUT_S)
self.process_frame(current_frame)
except Empty:
pass
def receive_options(self):
# Get number of planes:
n_planes = get_last_parameters(self.n_planes_queue, timeout=TIMEOUT_S)
if n_planes is not None:
self.n_planes = n_planes
self.reset()
# Get flat noise image to subtract:
calibration_ref = get_last_parameters(self.calibration_ref_queue,
timeout=TIMEOUT_S)
if calibration_ref is not None:
self.calibration_ref = calibration_ref
def reset(self):
self.first_volume = True
self.i_plane = 0
class StackSaver(LoggingProcess):
def __init__(
self,
stop_event: LoggedEvent,
is_saving_event: LoggedEvent,
duration_queue: Queue,
max_queue_size=2000,
):
super().__init__(name="saver")
self.stop_event = stop_event.new_reference(self.logger)
self.save_queue = ArrayQueue(max_mbytes=max_queue_size)
self.saving_signal = is_saving_event
self.saver_stopped_signal = LoggedEvent(self.logger,
SashimiEvents.SAVING_STOPPED)
self.saving = False
self.saving_parameter_queue = Queue()
self.save_parameters: Optional[SavingParameters] = SavingParameters()
self.i_in_chunk = 0
self.i_chunk = 0
self.i_plane = 0
self.i_volume = 0
self.n_volumes = 10
self.current_data = None
self.saved_status_queue = Queue()
self.frame_shape = None
self.dtype = np.uint16
self.duration_queue = duration_queue
self.notifier = notifiers[conf["notifier"]]
def run(self):
self.logger.log_message("started")
while not self.stop_event.is_set():
if self.saving_signal.is_set(
) and self.save_parameters is not None:
self.save_loop()
else:
self.receive_save_parameters()
self.close_log()
def save_loop(self):
notifier = self.notifier("lightsheet", **conf["notifier_options"])
# remove files if some are found at the save location
Path(self.save_parameters.output_dir).mkdir(parents=True,
exist_ok=True)
if (Path(self.save_parameters.output_dir) / "original" /
"stack_metadata.json").is_file():
shutil.rmtree(Path(self.save_parameters.output_dir) / "original")
(Path(self.save_parameters.output_dir) / "original").mkdir(
parents=True, exist_ok=True)
self.i_in_chunk = 0
self.i_chunk = 0
self.i_volume = 0
self.current_data = None
while (self.i_volume < self.n_volumes and self.saving_signal.is_set()
and not self.stop_event.is_set()):
self.receive_save_parameters()
try:
frame = self.save_queue.get(timeout=0.01)
self.logger.log_message("received volume")
self.fill_dataset(frame)
except Empty:
pass
if self.i_volume > 0:
if self.i_in_chunk != 0:
self.save_chunk()
self.update_saved_status_queue()
self.finalize_dataset()
self.current_data = None
if self.saving_signal.is_set():
notifier.notify()
self.saving_signal.clear()
self.saver_stopped_signal.set()
# Update status queue by resetting it:
self.update_saved_status_queue()
self.i_in_chunk = 0
self.i_chunk = 0
self.i_volume = 0
self.save_parameters = None
def fill_dataset(self, volume):
if self.current_data is None:
self.calculate_optimal_size(volume)
self.current_data = np.empty(
(self.save_parameters.chunk_size, *volume.shape),
dtype=self.dtype,
)
self.current_data[self.i_in_chunk, :, :, :] = volume
self.i_volume += 1
self.i_in_chunk += 1
self.update_saved_status_queue()
if self.i_in_chunk == self.save_parameters.chunk_size:
self.save_chunk()
def update_saved_status_queue(self):
self.saved_status_queue.put(
SavingStatus(
target_params=self.save_parameters,
i_in_chunk=self.i_in_chunk,
i_chunk=self.i_chunk,
i_volume=self.i_volume,
n_volumes=self.n_volumes,
))
def finalize_dataset(self):
self.logger.log_message("finished saving")
with open(
(Path(self.save_parameters.output_dir) / "original" /
"stack_metadata.json"),
"w",
) as f:
json.dump(
{
"shape_full": (
self.n_volumes,
*self.current_data.shape[1:],
),
"shape_block": (
self.save_parameters.chunk_size,
*self.current_data.shape[1:],
),
"crop_start": [0, 0, 0, 0],
"crop_end": [0, 0, 0, 0],
"padding": [0, 0, 0, 0],
"voxel_size":
self.save_parameters.voxel_size,
},
f,
)
def save_chunk(self):
self.logger.log_message("saved chunk")
fl.save(
Path(self.save_parameters.output_dir) /
"original/{:04d}.h5".format(self.i_chunk),
{"stack_4D": self.current_data[:self.i_in_chunk, :, :, :]},
compression="blosc",
)
self.i_in_chunk = 0
self.i_chunk += 1
def calculate_optimal_size(self, volume):
if self.dtype == np.uint16:
array_megabytes = (2 * volume.shape[0] * volume.shape[1] *
volume.shape[2] / 1048576)
else:
raise TypeError(
"Saving data type not supported. Only uint16 is supported")
self.save_parameters.chunk_size = int(
self.save_parameters.optimal_chunk_MB_RAM / array_megabytes)
def receive_save_parameters(self):
"""Receive parameters on the stack shape from the `State` obj and new duration
from either the `EsternalTrigger` or the `State` if triggering is disabled.
"""
# Get parameters:
parameters = get_last_parameters(self.saving_parameter_queue)
if parameters is not None:
self.save_parameters = parameters
# Get duration and update number of volumes:
new_duration = get_last_parameters(self.duration_queue)
if new_duration is not None:
self.n_volumes = int(
np.ceil(self.save_parameters.volumerate * new_duration))
class CameraProcess(LoggingProcess):
"""Process that handles the setting of new parameters to and the acquisition of
frames from the camera.
Parameters
----------
stop_event
wait_event
exp_trigger_event
camera_id
max_queue_size
n_fps_frames
"""
def __init__(
self,
stop_event: LoggedEvent,
wait_event: LoggedEvent,
exp_trigger_event: LoggedEvent,
camera_id=0,
max_queue_size=1200,
n_fps_frames=20,
):
super().__init__(name="camera")
# Queue to communicate
self.triggered_frame_rate_queue = Queue()
self.parameter_queue = Queue()
self.stop_event = stop_event.new_reference(self.logger)
self.wait_event = wait_event.new_reference(self.logger)
self.experiment_trigger_event = exp_trigger_event.new_reference(
self.logger)
self.image_queue = ArrayQueue(max_mbytes=max_queue_size)
self.camera_id = camera_id
self.camera = None
self.parameters = CamParameters()
self.framerate_rec = FramerateRecorder(n_fps_frames=n_fps_frames)
self.was_waiting = False
def initialize_camera(self):
if conf["scopeless"]:
self.camera = camera_class_dict["mock"]()
else:
self.camera = camera_class_dict[conf["camera"]["name"]](
camera_id=conf["camera"]["id"],
max_sensor_resolution=tuple(
conf["camera"]["max_sensor_resolution"]),
)
def run(self):
self.logger.log_message("started")
self.initialize_camera()
self.run_camera()
self.camera.shutdown()
self.logger.close()
def run_camera(self):
"""Main run for the camera. Depending on whether the camera mode is PAUSED or not,
either the self.pause_loop or self.camera_loop are executed.
"""
while not self.stop_event.is_set():
# removed if statement based on camera mode: toggeling between modes is done in loops.
self.pause_loop()
self.camera_loop()
def pause_loop(self):
"""Camera idle loop, just wait until parameters are updated. Check them, and if
CameraMode.PAUSED is still set, return here.
"""
self.logger.log_message("Paused aquisition")
self.camera.stop_acquisition()
while (not self.stop_event.is_set()
and self.parameters.camera_mode == CameraMode.PAUSED):
try:
new_parameters = self.parameter_queue.get(timeout=0.001)
if new_parameters != self.parameters:
self.update_parameters(new_parameters, stop_start=False)
if self.parameters.camera_mode != CameraMode.PAUSED:
break
except Empty:
pass
def camera_loop(self):
"""Camera running loop, grab frames and set new parameters if available."""
self.logger.log_message("Started acquisition")
self.camera.start_acquisition()
while (not self.stop_event.is_set()
and self.parameters.camera_mode != CameraMode.PAUSED):
is_waiting = self.wait_event.is_set()
frames = self.camera.get_frames()
# if no frames are received (either this loop is in between frames
# or we are in the waining period)
if frames:
for frame in frames:
self.logger.log_message("received frame of shape " +
str(frame.shape))
# this means this is the first frame received since
# the waiting period is over, the signal has to be sent that
# saving can start
if self.was_waiting and not is_waiting:
self.experiment_trigger_event.set()
# TODO do not crash here if queue is full
self.image_queue.put(frame)
self.was_waiting = is_waiting
self.update_framerate()
# Empty parameters queue and set new parameters with the most recent value
new_parameters = get_last_parameters(self.parameter_queue,
timeout=0.001)
if new_parameters is not None:
if new_parameters.camera_mode == CameraMode.ABORT or (
new_parameters != self.parameters):
self.update_parameters(new_parameters)
def update_parameters(self, new_parameters, stop_start=True):
""" "Set new parameters and stop and start the camera to make sure all changes take place."""
self.parameters = new_parameters
if stop_start:
self.camera.stop_acquisition()
# In general, ROI and binning are a bit funny in their interactions, and need to be handled
# carefully in the specific camera interfaces.
for attribute in ["binning", "roi", "exposure_time", "trigger_mode"]:
setattr(self.camera, attribute, getattr(self.parameters,
attribute))
if stop_start:
self.camera.start_acquisition()
self.framerate_rec.restart()
self.logger.log_message("Updated parameters " + str(self.parameters))
def update_framerate(self):
self.framerate_rec.update_framerate()
if self.framerate_rec.i_fps == 0:
self.triggered_frame_rate_queue.put(
self.framerate_rec.current_framerate)
def close_camera(self):
self.camera.shutdown()
class Scanner(Process):
def __init__(self,
experiment_start_event,
duration_queue,
max_queuesize=200):
super().__init__()
self.data_queue = ArrayQueue(max_mbytes=max_queuesize)
self.parameter_queue = Queue()
self.stop_event = Event()
self.experiment_start_event = experiment_start_event
self.scanning_parameters = ScanningParameters()
self.new_parameters = copy(self.scanning_parameters)
self.duration_queue = duration_queue
self.n_frames_queue = Queue()
def run(self):
self.compute_scan_parameters()
self.run_scanning()
def compute_scan_parameters(self):
self.extent_x = (
-self.scanning_parameters.voltage_x,
self.scanning_parameters.voltage_x,
)
self.extent_y = (
-self.scanning_parameters.voltage_y,
self.scanning_parameters.voltage_y,
)
self.n_x = self.scanning_parameters.n_x
self.n_y = self.scanning_parameters.n_y
self.raw_x, self.raw_y = compute_waveform(self.scanning_parameters)
self.pos_x = (self.raw_x *
((self.extent_x[1] - self.extent_x[0]) / self.n_x) +
self.extent_x[0])
self.pos_y = (self.raw_y *
((self.extent_y[1] - self.extent_y[0]) / self.n_y) +
self.extent_y[0])
self.n_bin = self.scanning_parameters.n_bin
self.n_samples_out = len(self.raw_x)
self.n_samples_in = self.n_samples_out * self.n_bin
self.sample_rate_out = self.scanning_parameters.sample_rate_out
self.plane_duration = self.n_samples_out / self.sample_rate_out
self.sample_rate_in = self.n_bin * self.sample_rate_out
self.write_signals = np.stack([self.pos_x, self.pos_y], 0)
self.read_buffer = np.zeros((2, self.n_samples_in))
self.mystery_offset = self.scanning_parameters.mystery_offset
def setup_tasks(self, read_task, write_task, shutter_task):
# Configure the channels
read_task.ai_channels.add_ai_voltage_chan(
"Dev1/ai0:1", min_val=-1, max_val=1
) # channels are 0: green PMT, 1 x galvo pos 2 y galvo pos
write_task.ao_channels.add_ao_voltage_chan("Dev1/ao0:1",
min_val=-10,
max_val=10)
shutter_task.do_channels.add_do_chan(
"Dev1/port0/line1", line_grouping=LineGrouping.CHAN_PER_LINE)
# Set the timing of both to the onboard clock so that they are synchronised
read_task.timing.cfg_samp_clk_timing(
rate=self.sample_rate_in,
source="OnboardClock",
active_edge=Edge.RISING,
sample_mode=AcquisitionType.CONTINUOUS,
samps_per_chan=self.n_samples_in,
)
write_task.timing.cfg_samp_clk_timing(
rate=self.sample_rate_out,
source="OnboardClock",
active_edge=Edge.RISING,
sample_mode=AcquisitionType.CONTINUOUS,
samps_per_chan=self.n_samples_out,
)
# This is necessary to synchronise reading and wrting
read_task.triggers.start_trigger.cfg_dig_edge_start_trig(
"/Dev1/ao/StartTrigger", Edge.RISING)
def check_start_plane(self):
if self.scanning_parameters.scanning_state == ScanningState.EXPERIMENT_RUNNING:
while not self.experiment_start_event.is_set():
sleep(0.00001)
def calculate_duration(self):
try:
duration = self.duration_queue.get(timeout=0.0001)
self.scanning_parameters.n_frames = (int(
ceil(duration / frame_duration(self.scanning_parameters))) + 1)
self.n_frames_queue.put(self.scanning_parameters.n_frames)
except Empty:
pass
def scan_loop(self, read_task, write_task):
writer = AnalogMultiChannelWriter(write_task.out_stream)
reader = AnalogMultiChannelReader(read_task.in_stream)
first_write = True
i_acquired = 0
while not self.stop_event.is_set() and (
not self.scanning_parameters.scanning_state
== ScanningState.EXPERIMENT_RUNNING
or i_acquired < self.scanning_parameters.n_frames):
# The first write has to be defined before the task starts
try:
writer.write_many_sample(self.write_signals)
if i_acquired == 0:
self.check_start_plane()
if first_write:
read_task.start()
write_task.start()
first_write = False
reader.read_many_sample(
self.read_buffer,
number_of_samples_per_channel=self.n_samples_in,
timeout=1,
)
i_acquired += 1
except nidaqmx.DaqError as e:
print(e)
break
self.data_queue.put(self.read_buffer[0, :])
# if new parameters have been received and changed, update
# them, breaking out of the loop if the experiment is not running
try:
self.new_parameters = self.parameter_queue.get(timeout=0.0001)
if self.new_parameters != self.scanning_parameters and (
self.scanning_parameters.scanning_state !=
ScanningState.EXPERIMENT_RUNNING
or self.new_parameters.scanning_state
== ScanningState.PREVIEW):
break
except Empty:
pass
# calculate duration
self.calculate_duration()
def pause_loop(self):
while not self.stop_event.is_set():
try:
self.new_parameters = self.parameter_queue.get(timeout=0.001)
if self.new_parameters != self.scanning_parameters and (
self.scanning_parameters.scanning_state !=
ScanningState.EXPERIMENT_RUNNING
or self.new_parameters.scanning_state
== ScanningState.PREVIEW):
break
except Empty:
pass
def toggle_shutter(self, shutter_task):
shutter_task.write(False, auto_start=True)
shutter_task.write(True, auto_start=True)
shutter_task.write(False, auto_start=True)
sleep(0.05)
def run_scanning(self):
while not self.stop_event.is_set():
toggle_shutter = False
if (self.new_parameters.scanning_state == ScanningState.PAUSED
and self.scanning_parameters.scanning_state !=
ScanningState.PAUSED
) or (
self.new_parameters.scanning_state != ScanningState.PAUSED
and self.scanning_parameters.scanning_state
== ScanningState.PAUSED):
toggle_shutter = True
self.scanning_parameters = self.new_parameters
self.compute_scan_parameters()
with Task() as write_task, Task() as read_task, Task(
) as shutter_task:
self.setup_tasks(read_task, write_task, shutter_task)
if self.scanning_parameters.reset_shutter or toggle_shutter:
self.toggle_shutter(shutter_task)
if self.scanning_parameters.scanning_state == ScanningState.PAUSED:
self.pause_loop()
else:
self.scan_loop(read_task, write_task)
class ExperimentState(QObject):
sig_scanning_changed = pyqtSignal()
def __init__(self, diagnostics=False):
super().__init__()
if diagnostics:
self.sequence_queue = Queue()
self.sequence_diagram = SequenceDiagram(self.sequence_queue,
"main")
self.experiment_start_event = Event()
self.scanning_settings = ScanningSettings()
self.experiment_settings = ExperimentSettings()
self.pause_after = False
self.parameter_tree = ParameterTree()
self.parameter_tree.add(self.scanning_settings)
self.parameter_tree.add(self.experiment_settings)
self.end_event = Event()
self.external_sync = ZMQcomm()
self.duration_queue = Queue()
self.scanner = Scanner(self.experiment_start_event,
duration_queue=self.duration_queue)
self.scanning_parameters = None
self.reconstructor = ImageReconstructor(self.scanner.data_queue,
self.scanner.stop_event)
self.save_queue = ArrayQueue(max_mbytes=800)
self.saver = StackSaver(self.scanner.stop_event, self.save_queue,
self.scanner.n_frames_queue)
self.save_status: Optional[SavingStatus] = None
self.motors = dict()
self.motors["x"] = MotorControl("COM6", axes="x")
self.motors["y"] = MotorControl("COM6", axes="y")
self.motors["z"] = MotorControl("COM6", axes="z")
self.power_controller = LaserPowerControl()
self.scanning_settings.sig_param_changed.connect(self.send_scan_params)
self.scanning_settings.sig_param_changed.connect(self.send_save_params)
self.scanner.start()
self.reconstructor.start()
self.saver.start()
self.open_setup()
self.paused = False
@property
def saving(self):
return self.saver.saving_signal.is_set()
def open_setup(self):
self.send_scan_params()
def start_experiment(self, first_plane=True):
duration = self.external_sync.send(self.parameter_tree.serialize())
if duration is None:
self.restart_scanning()
return False
self.duration_queue.put(duration)
params_to_send = convert_params(self.scanning_settings)
params_to_send.scanning_state = ScanningState.EXPERIMENT_RUNNING
self.scanner.parameter_queue.put(params_to_send)
if first_plane:
self.send_save_params()
self.saver.saving_signal.set()
self.experiment_start_event.set()
return True
def end_experiment(self, force=False):
self.experiment_start_event.clear()
if not force and self.save_status.i_z + 1 < self.save_status.target_params.n_z:
self.advance_plane()
else:
self.saver.saving_signal.clear()
if self.pause_after:
self.pause_scanning()
else:
self.restart_scanning()
def restart_scanning(self):
params_to_send = convert_params(self.scanning_settings)
params_to_send.scanning_state = ScanningState.PREVIEW
self.scanner.parameter_queue.put(params_to_send)
self.paused = False
def pause_scanning(self):
params_to_send = convert_params(self.scanning_settings)
params_to_send.scanning_state = ScanningState.PAUSED
self.scanner.parameter_queue.put(params_to_send)
self.paused = True
def advance_plane(self):
self.motors["z"].move_rel(self.experiment_settings.dz / 1000)
sleep(0.2)
self.start_experiment(first_plane=False)
def close_setup(self):
""" Cleanup on programe close:
end all parallel processes, close all communication channels
"""
for motor in self.motors.values():
motor.end_session()
self.power_controller.terminate_connection()
self.scanner.stop_event.set()
self.end_event.set()
self.scanner.join()
self.reconstructor.join()
def get_image(self):
try:
image = -self.reconstructor.output_queue.get(timeout=0.001)
if self.saver.saving_signal.is_set():
if (self.save_status is not None and self.save_status.i_t + 1
== self.save_status.target_params.n_t):
self.end_experiment()
self.save_queue.put(image)
return image
except Empty:
return None
def send_scan_params(self):
self.scanning_parameters = convert_params(self.scanning_settings)
self.power_controller.move_abs(self.scanning_settings.laser_power)
self.scanner.parameter_queue.put(self.scanning_parameters)
self.reconstructor.parameter_queue.put(self.scanning_parameters)
self.sig_scanning_changed.emit()
def send_save_params(self):
self.saver.saving_parameter_queue.put(
SavingParameters(
output_dir=Path(self.experiment_settings.save_dir),
plane_size=(self.scanning_parameters.n_x,
self.scanning_parameters.n_y),
n_z=self.experiment_settings.n_planes,
notification_email=self.experiment_settings.notification_email,
notification_frequency=self.experiment_settings.
notify_every_n_planes))
def get_save_status(self) -> Optional[SavingStatus]:
try:
self.save_status = self.saver.saved_status_queue.get(timeout=0.001)
return self.save_status
except Empty:
pass
return None
class MovingFrameDispatcher(FrameDispatcher):
""" """
def __init__(self, *args, signal_start_rec, output_queue_mb=500, **kwargs):
super().__init__(*args, **kwargs)
self.save_queue = ArrayQueue(max_mbytes=output_queue_mb)
self.framestart_queue = Queue()
self.diagnostic_queue = Queue()
self.processing_parameters = MovementDetectionParameters(
).get_clean_values()
self.signal_start_rec = signal_start_rec
self.mem_use = 0
self.diagnostic_params = [
"n_pixels_difference",
"recording_state",
"n_images_in_buffer",
]
def run(self):
""" """
t, frame_0 = self.frame_queue.get(timeout=10)
n_previous_compare = 3
image_crop = slice(
self.processing_parameters["frame_margin"],
-self.processing_parameters["frame_margin"],
)
previous_ims = np.zeros(
(n_previous_compare, ) + frame_0[image_crop].shape, dtype=np.uint8)
image_buffer = deque()
record_counter = 0
i_frame = 0
recording_state = False
i_recorded = 0
while not self.finished_signal.is_set():
# Gets the processing parameters from their queue
if self.processing_parameter_queue is not None:
try:
self.processing_parameters = self.processing_parameter_queue.get(
timeout=0.00001)
except Empty:
pass
try:
current_time, current_frame = self.frame_queue.get(
timeout=0.001)
# process frames as they come, threshold them to roughly
# find the fish (e.g. eyes)
_, current_frame_thresh = cv2.threshold(
cv2.boxFilter(current_frame[image_crop], -1, (3, 3)),
self.processing_parameters["fish_threshold"],
255,
cv2.THRESH_BINARY,
)
# compare the thresholded frame to the previous ones,
# if there are enough differences
# because the fish moves, start recording to file
if i_frame >= n_previous_compare:
difsum = _compare_to_previous(current_frame_thresh,
previous_ims)
# put the difference in the diagnostic queue so that
# the threshold can be set correctly
if np.all(difsum > self.
processing_parameters["motion_threshold_n_pix"]):
record_counter = self.processing_parameters[
"n_next_save"]
if record_counter > 0:
if self.signal_start_rec.is_set(
) and self.mem_use < 0.9:
if not recording_state:
while image_buffer:
time, im = image_buffer.popleft()
self.framestart_queue.put(time)
self.save_queue.put(im)
i_recorded += 1
self.save_queue.put(current_frame)
self.framestart_queue.put(current_time)
i_recorded += 1
recording_state = True
record_counter -= 1
else:
recording_state = False
image_buffer.append((current_time, current_frame))
if (len(image_buffer) >
self.processing_parameters["n_previous_save"]):
image_buffer.popleft()
self.diagnostic_queue.put((
current_time,
(
difsum[i_frame % n_previous_compare],
recording_state,
len(image_buffer),
),
))
i_frame += 1
previous_ims[i_frame %
n_previous_compare, :, :] = current_frame_thresh
# calculate the framerate and send frame to gui
self.update_framerate()
if self.processing_parameters["show_thresholded"]:
self.send_to_gui(current_frame_thresh)
else:
self.send_to_gui(current_frame)
except Empty:
pass
