Skip to content

match the 1.4 plugin's channel and camera axis layout #475

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
bls337 merged 3 commits into from
Aug 6, 2026
Merged

match the 1.4 plugin's channel and camera axis layout #475

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
11e7eed
order the channel axis innermost when the controller interleaves chan…
Aug 5, 2026
333e4f2
pack the camera axis channel-fastest to match 1.4
Aug 5, 2026
40edf2b
remove the unused channel event factory that packs the camera axis wrong
Aug 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 5 additions & 4 deletions src/main/java/org/micromanager/lightsheetmanager/model/acquisitions/AcquisitionEngine.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -380,8 +380,9 @@ protected DefaultSummaryMetadata addMMSummaryMetadata(JSONObject summaryMetadata
summaryMetadata.put(PropertyKey.CHANNEL_GROUP.key(), acqSettings_.channels().group());

// one name per position on the store's channel axis; with simultaneous cameras the
// camera index varies fastest (LightSheetEventAdapter.cameras assigns
// channelIndex * numCameras + cameraIndex), so repeat each channel name per camera
// channel index varies fastest, so walk cameras outermost and repeat the whole channel
// list per camera. This loop order is the slot order: reverse one and every image gets
// the wrong name.
final List<String> channelNames = new ArrayList<>();
final List<String> baseChannelNames = new ArrayList<>();
if (acqSettings_.channels().enabled() && acqSettings_.channels().count() > 0) {
Expand All @@ -392,8 +393,8 @@ protected DefaultSummaryMetadata addMMSummaryMetadata(JSONObject summaryMetadata
baseChannelNames.add("Default");
}
if (model_.devices().adapter().numSimultaneousCameras() > 1) {
for (String channelName : baseChannelNames) {
for (CameraBase camera : model_.devices().imagingCameras()) {
for (CameraBase camera : model_.devices().imagingCameras()) {
for (String channelName : baseChannelNames) {
channelNames.add(acqSettings_.channels().enabled()
? channelName + "-" + camera.getDeviceName()
: camera.getDeviceName());
Expand Down
5 changes: 3 additions & 2 deletions ...in/java/org/micromanager/lightsheetmanager/model/acquisitions/AcquisitionEngineScape.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -614,7 +614,8 @@ public void close() {
if (acqSettings_.channels().enabled()) {
currentAcquisition_.submitEventIterator(
LightSheetEventAdapter.createTimelapseMultiChannelVolumeAcqEvents(
baseEvent.copy(), acqSettings_, cameraNames, null));
baseEvent.copy(), acqSettings_, cameraNames,
acqSettings_.channels().used(), null));
} else {
currentAcquisition_.submitEventIterator(
LightSheetEventAdapter.createTimelapseVolumeAcqEvents(
Expand Down Expand Up @@ -663,7 +664,7 @@ public void close() {
currentAcquisition_.submitEventIterator(
LightSheetEventAdapter.createSingleChannelVolumeAcqEvents(
baseEvent.copy(), acqSettings_, cameraNames, null,
channelIndex, used[channelIndex]));
channelIndex, used[channelIndex], used.length));
}
} else {
// SLICE_HW with hardware timepoints off: the controller emits all
Expand Down
173 changes: 105 additions & 68 deletions ...in/java/org/micromanager/lightsheetmanager/model/acquisitions/LightSheetEventAdapter.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,8 +26,12 @@ public final class LightSheetEventAdapter {
// viewer size/display dimensions by them, so the VALUES cannot be changed; an axis MM does not
// know is dropped by TIFF storage and never displayed.
// CAMERA_AXIS is deliberately AcqEngJ's "channel" axis: LSM packs the combined
// (channelIndex * numCameras + cameraIndex) slot into it, which is why the name here says camera
// (channelIndex + cameraIndex * numChannels) slot into it, which is why the name here says camera
// while the value says channel. Note setChannelName() writes this same axis.
// Channel varies fastest, matching what 1.4 writes for one-sided dual-camera acquisitions, so a
// pipeline built to read 1.4 datasets reads ours without re-indexing. The slot naming in
// AcquisitionEngine.addMMSummaryMetadata walks the same order; the two must change together or
// the names label the wrong images.
public static final String TIME_AXIS = "time";
public static final String POSITION_AXIS = "position";
public static final String CAMERA_AXIS = "channel";
Expand All @@ -41,34 +45,51 @@ private LightSheetEventAdapter() {
throw new AssertionError("Utility class; do not instantiate.");
}

/**
* Build events for a timelapse of volumes whose channels the controller switches slice by slice.
* <p>
* The controller free-runs the time points here, so the axis order below the time axis has to be
* the interleaved order {@link #createChannelPerSliceAcqEvents} describes: the channel axis
* innermost, because the controller switches channel within a slice and AcqEngJ assigns each
* image to the first event matching that camera in sequence order. Ordering the channels above
* the slices instead does not fail, since the event count still matches; nearly every frame is
* filed against the wrong channel and slice.
* <p>
* A single channel takes the same exception as that factory: hardware channel switching is not
* set up for one channel, so its own preset and offset apply and are baked onto the base event,
* where they cannot split the merge.
*
* @param usedChannels the channels the controller was programmed for
*/
public static Iterator<AcquisitionEvent> createTimelapseMultiChannelVolumeAcqEvents(
AcquisitionEvent baseEvent, AcquisitionSettings settings,
String[] cameraDeviceNames,
String[] cameraDeviceNames, ChannelSpec[] usedChannels,
Function<AcquisitionEvent, AcquisitionEvent> eventMonitor) {

if (settings.numTimePoints() <= 1) {
throw new RuntimeException("timelapse selected but only one timepoint");
}
Function<AcquisitionEvent, Iterator<AcquisitionEvent>> timelapse =
timelapse(settings.numTimePoints(), settings.timePointIntervalSec());

if (settings.channels().count() == 1) {
throw new RuntimeException("Expected multiple channels but only one found");
if (usedChannels.length == 1) {
applySingleChannel(baseEvent, usedChannels[0]);
}

Function<AcquisitionEvent, Iterator<AcquisitionEvent>> channels =
channels(settings.channels().used());

Function<AcquisitionEvent, Iterator<AcquisitionEvent>> timelapse =
timelapse(settings.numTimePoints(), settings.timePointIntervalSec());
// Base 0 so cameras() leaves the plain camera index on the axis; channelAxis() runs
// innermost and folds it into the combined slot.
Function<AcquisitionEvent, Iterator<AcquisitionEvent>> cameras =
cameras(cameraDeviceNames, 0);
Function<AcquisitionEvent, Iterator<AcquisitionEvent>> zStack =
zStack(0, settings.volume().slicesPerView());

Function<AcquisitionEvent, Iterator<AcquisitionEvent>> cameras = cameras(cameraDeviceNames);
Function<AcquisitionEvent, Iterator<AcquisitionEvent>> channels =
channelAxis(usedChannels.length);

ArrayList<Function<AcquisitionEvent, Iterator<AcquisitionEvent>>> acqFunctions = new ArrayList<>();
acqFunctions.add(timelapse);
acqFunctions.add(channels);
acqFunctions.add(cameras);
acqFunctions.add(zStack);
acqFunctions.add(channels);
return new AcquisitionEventIterator(baseEvent, acqFunctions, eventMonitor);
}

Expand All @@ -95,26 +116,6 @@ public static Iterator<AcquisitionEvent> createTimelapseVolumeAcqEvents(
return new AcquisitionEventIterator(baseEvent, acqFunctions, eventMonitor);
}

public static Iterator<AcquisitionEvent> createChannelAcqEvents(
AcquisitionEvent baseEvent, AcquisitionSettings settings,
String[] cameraDeviceNames,
Function<AcquisitionEvent, AcquisitionEvent> eventMonitor) {

Function<AcquisitionEvent, Iterator<AcquisitionEvent>> channels =
channels(settings.channels().used());

Function<AcquisitionEvent, Iterator<AcquisitionEvent>> cameras = cameras(cameraDeviceNames);

Function<AcquisitionEvent, Iterator<AcquisitionEvent>> zStack =
zStack(0, settings.volume().slicesPerView());

ArrayList<Function<AcquisitionEvent, Iterator<AcquisitionEvent>>> acqFunctions = new ArrayList<>();
acqFunctions.add(channels);
acqFunctions.add(cameras);
acqFunctions.add(zStack);
return new AcquisitionEventIterator(baseEvent, acqFunctions, eventMonitor);
}

public static Iterator<AcquisitionEvent> createAcqEvents(
AcquisitionEvent baseEvent, AcquisitionSettings settings,
String[] cameraDeviceNames,
Expand All @@ -138,18 +139,21 @@ public static Iterator<AcquisitionEvent> createAcqEvents(
* its own event iterator. AcqEngJ appends a SequenceEnd flush at the end of every
* submitted iterator, guaranteeing exactly one camera sequence + one controller fire per
* channel-volume regardless of whether the channel presets are identical, distinct, or
* property-sequenceable. Composing all channels into a single iterator (the old
* {@link #createChannelAcqEvents}) instead lets AcqEngJ merge identical-preset channels into one
* sequence, firing the controller once and collapsing the channel dimension.
* property-sequenceable. Composing all channels into a single iterator instead lets AcqEngJ merge
* identical-preset channels into one sequence, firing the controller once and collapsing the
* channel dimension.
*
* @param channelIndex zero-based index of this channel among the used channels
* @param channel the channel to acquire
* @param numUsedChannels how many channels the submitting loop is iterating; taken from the
* caller's captured array rather than re-read here, so the slot stride
* cannot change between one channel's submission and the next
*/
public static Iterator<AcquisitionEvent> createSingleChannelVolumeAcqEvents(
AcquisitionEvent baseEvent, AcquisitionSettings settings,
String[] cameraDeviceNames,
Function<AcquisitionEvent, AcquisitionEvent> eventMonitor,
int channelIndex, ChannelSpec channel) {
int channelIndex, ChannelSpec channel, int numUsedChannels) {

// Bake this one channel's config into the base event (what channels().next() does per channel).
baseEvent.setConfigGroup(channel.getGroup());
Expand All @@ -169,13 +173,14 @@ public static Iterator<AcquisitionEvent> createSingleChannelVolumeAcqEvents(
}
baseEvent.setZ(baseEvent.getZIndex(), zPos);

// cameras() would now derive the same base from CAMERA_AXIS (setChannelName above put the
// channel index there); passing it explicitly keeps this path independent of that coupling.
final int channelAxisBase = isUsingMultipleCameras
? channelIndex * cameraDeviceNames.length : channelIndex;
// No channel stage follows here, so cameras() writes the finished slot: this channel's index
// as the base, stepping a whole channel count per camera. cameras() would otherwise derive
// the base from CAMERA_AXIS (setChannelName above put the channel index there); passing it
// explicitly keeps this path independent of that coupling.
final int cameraStride = isUsingMultipleCameras ? numUsedChannels : 1;

Function<AcquisitionEvent, Iterator<AcquisitionEvent>> cameras =
cameras(cameraDeviceNames, channelAxisBase);
cameras(cameraDeviceNames, channelIndex, cameraStride);
Function<AcquisitionEvent, Iterator<AcquisitionEvent>> zStack =
zStack(0, settings.volume().slicesPerView());

Expand Down Expand Up @@ -219,22 +224,7 @@ public static Iterator<AcquisitionEvent> createChannelPerSliceAcqEvents(
Function<AcquisitionEvent, AcquisitionEvent> eventMonitor) {

if (usedChannels.length == 1) {
final ChannelSpec channel = usedChannels[0];
baseEvent.setConfigGroup(channel.getGroup());
baseEvent.setConfigPreset(channel.getName());

// Channel z-offset: mirror channels(), apply the offset to the current stage/z position.
double zPos;
if (baseEvent.getZPosition() == null) {
try {
zPos = Engine.getCore().getPosition() + channel.getOffset();
} catch (Exception e) {
throw new RuntimeException(e);
}
} else {
zPos = baseEvent.getZPosition() + channel.getOffset();
}
baseEvent.setZ(baseEvent.getZIndex(), zPos);
applySingleChannel(baseEvent, usedChannels[0]);
}

// Base 0 so cameras() leaves the plain camera index on the axis; channelAxis() runs
Expand All @@ -244,7 +234,7 @@ public static Iterator<AcquisitionEvent> createChannelPerSliceAcqEvents(
Function<AcquisitionEvent, Iterator<AcquisitionEvent>> zStack =
zStack(0, settings.volume().slicesPerView());
Function<AcquisitionEvent, Iterator<AcquisitionEvent>> channels =
channelAxis(usedChannels.length, cameraDeviceNames.length);
channelAxis(usedChannels.length);

ArrayList<Function<AcquisitionEvent, Iterator<AcquisitionEvent>>> acqFunctions = new ArrayList<>();
acqFunctions.add(cameras);
Expand All @@ -253,6 +243,28 @@ public static Iterator<AcquisitionEvent> createChannelPerSliceAcqEvents(
return new AcquisitionEventIterator(baseEvent, acqFunctions, eventMonitor);
}

/**
* Bake one channel's preset and z-offset onto the base event, as {@link #channels} stamps them
* per event on the software path.
*/
private static void applySingleChannel(AcquisitionEvent baseEvent, ChannelSpec channel) {
baseEvent.setConfigGroup(channel.getGroup());
baseEvent.setConfigPreset(channel.getName());

// Channel z-offset: mirror channels(), apply the offset to the current stage/z position.
double zPos;
if (baseEvent.getZPosition() == null) {
try {
zPos = Engine.getCore().getPosition() + channel.getOffset();
} catch (Exception e) {
throw new RuntimeException(e);
}
} else {
zPos = baseEvent.getZPosition() + channel.getOffset();
}
baseEvent.setZ(baseEvent.getZIndex(), zPos);
}

/**
* Build events for a volume whose channels the controller switches volume by volume.
* <p>
Expand Down Expand Up @@ -289,14 +301,13 @@ public static Iterator<AcquisitionEvent> createChannelPerVolumeAcqEvents(
* Deliberately sets no config group or preset, and no channel z-offset: the controller owns
* channel switching here, and the offset stays wherever the user's own preset left it.
* Runs innermost, after {@link #cameras}, so the value already on the axis is the camera index;
* it is replaced with the combined slot {@code channelIndex * numCameras + cameraIndex} that
* it is replaced with the combined slot {@code channelIndex + cameraIndex * numChannels} that
* sizes and names the datastore.
*
* @param numChannels the number of channels the controller was programmed for
* @param numCameras the number of cameras this event fans out over
*/
public static Function<AcquisitionEvent, Iterator<AcquisitionEvent>> channelAxis(
int numChannels, int numCameras) {
int numChannels) {
return (AcquisitionEvent event) -> new Iterator<>() {

private int channelIndex_ = 0;
Expand All @@ -320,7 +331,7 @@ public AcquisitionEvent next() {
}
}

channelEvent.setAxisPosition(CAMERA_AXIS, channelIndex_ * numCameras + cameraIndex);
channelEvent.setAxisPosition(CAMERA_AXIS, channelIndex_ + cameraIndex * numChannels);
channelIndex_++;
return channelEvent;
}
Expand All @@ -342,17 +353,40 @@ public static Function<AcquisitionEvent, Iterator<AcquisitionEvent>> cameras(Str
* <b>{@code CAMERA_AXIS} is AcqEngJ's {@code "channel"} axis</b> ({@code AcqEngMetadata.CHANNEL_AXIS}):
* {@code AcquisitionEvent.setChannelName(s)} compiles to {@code setAxisPosition("channel", s)}. So the
* channel index and the camera index share one axis, and the coordinate written here is the combined
* slot {@code channelIndex * numCameras + cameraIndex} that {@code addMMSummaryMetadata} names and
* slot {@code channelIndex + cameraIndex * numChannels} that {@code addMMSummaryMetadata} names and
* sizes the datastore for. Do not treat a value already present on that axis as a camera base unless
* you put it there: on the {@code channels()}-composed paths it is the raw channel index written by
* {@code setChannelName}.
* <p>
* This overload steps one slot per camera, which is only the whole coordinate where a single
* channel is in play; a multi-channel caller wants the stride overload below.
*
* @param channelAxisBase the channel-axis base for this fan-out, or {@code null} to derive it from
* the event: the channel index {@code channels()} left on the axis, times
* the camera count when using multiple cameras
* the event: the channel index {@code channels()} left on the axis
*/
public static Function<AcquisitionEvent, Iterator<AcquisitionEvent>> cameras(
String[] cameraDeviceNames, Integer channelAxisBase) {
return cameras(cameraDeviceNames, channelAxisBase, 1);
}

/**
* Fan an event out over the cameras, advancing the {@link #CAMERA_AXIS} coordinate by
* {@code cameraStride} per camera.
* <p>
* Channel varies fastest on that axis, so a camera step is a step of {@code numChannels}, not of
* one. Two shapes use this:
* <ul>
* <li>composed with {@link #channelAxis}, which supplies the channel term itself: pass base 0
* and stride 1, leaving the plain camera index for it to fold in.</li>
* <li>alone, on the per-channel software path, where no channel stage follows: pass the channel
* index as the base and {@code numChannels} as the stride, giving the whole slot here.</li>
* </ul>
*
* @param cameraStride how far the slot advances per camera; {@code numChannels} when this call
* produces the finished coordinate, {@code 1} when {@link #channelAxis} does
*/
public static Function<AcquisitionEvent, Iterator<AcquisitionEvent>> cameras(
String[] cameraDeviceNames, Integer channelAxisBase, int cameraStride) {
return (AcquisitionEvent event) -> new Iterator<>() {

private int cameraIndex_ = 0;
Expand Down Expand Up @@ -386,11 +420,14 @@ public AcquisitionEvent next() {
// ignore => number already assigned
}
}
baseIndex = isUsingMultipleCameras
? parsed * cameraDeviceNames_.length : parsed;
// The channel index is the base as it stands: channel varies fastest, so it is
// the camera step that scales, and that is cameraStride's job. A multi-channel
// caller of this derive path must therefore pass numChannels as the stride;
// today every caller that reaches it has channels disabled, so this is 0.
baseIndex = parsed;
}

cameraEvent.setAxisPosition(CAMERA_AXIS, baseIndex + cameraIndex_);
cameraEvent.setAxisPosition(CAMERA_AXIS, baseIndex + cameraIndex_ * cameraStride);
cameraIndex_++;
return cameraEvent;
}
Expand Down
Loading