Add frontend and backend part for job to materialize a volume annotation

docs/jobs.md

@@ -7,6 +7,7 @@ Example workflows:
 - [converting datasets on upload](./datasets.md#uploading-through-the-web-browser)
 - [automated analysis](./automated_analysis.md), e.g., nuclei inferal 
 - [mesh file (pre)-computation](./mesh_visualization.md)
+- [applying a merger mode annotation](./volume_annotation.md#proof_reading_and_merging_segments)
 - Tiff export of a bounding box
 - downsampling volume annotations
 

docs/volume_annotation.md

@@ -26,7 +26,7 @@ The following interactions and modifiers are available when working with the vol
 - `Change Brush Size`: Changes the size of the brushing tool. 
 - `Overwrite Everything`: When labelling with the brush/trace tool, you can annotate every available voxel without any regard if it was already labelled as a different segment or whether it is unlabelled. This allows you to draw over existing segments.
 - `Only Overwrite Empty Areas`: In contrast to the `Overwrite Everything` modifier, the forces the brush/trace tool to only label voxels without any segment ID ("empty areas"). This is useful when annotating segments that directly touch each other to avoid accidental overwrites.
-- `2D Fill`/ `3D Fill`: Modifies the flood filling tool to work in 2D (in-plane only) or 3D (volumetric fill/re-labelling). 3D flood fill is contraint to a small, regional bounding box for performance reasons. Read more about flood fills below.
+- `2D Fill`/ `3D Fill`: Modifies the flood filling tool to work in 2D (in-plane only) or 3D (volumetric fill/re-labelling). 3D flood fill is constrained to a small, regional bounding box for performance reasons. Read more about flood fills below.
 
 ![Adding labels with the Trace tool](./images/volume_trace.gif)
 ![Adding labels with the Brush tool](./images/volume_brush.gif)
@@ -36,6 +36,15 @@ In the `Segmentation` tab on the right-hand sidepanel, you can see the segment I
 
 The active segment ID under the cursor can be found in the status bar at the bottom of the screen or through the context-sensitive menu on right-click.
 
+### Merging volume annotation with fallback data
+
+After finishing the annotation of a volume layer with a fallback layer, the combined state of these layers can be materialized into a new dataset. For this go to the layer settings in the left border tab. On the top right of the volume layer is the following button:
+
+![Icon to open the materialize volume annotation modal](./images/materialize_volume_annotation_icon.jpg)
+
+This button opens up a modal that starts a long-running job which will materialize the volume annotation.
+
+
 ### Proof-Reading and Merging Segments
 
 webKnossos support proof-reading of segments from automatic segmentation. With "Merger Mode" individual segments (e.g. from over-segmentation) can be combined ("merged") to refine the segmentation and fix split errors. 
@@ -48,7 +57,7 @@ To fix split errors see section on flood filling below.
 
 ![Video: Proof-Reading Volume Annotations](https://www.youtube.com/watch?v=Sq4AuWanK14)
 
-After finishing the proof-reading, a long-running job can be started to apply the merging of segments into a new dataset with the same layers. The job can be started via a modal. The modal can be open by clicking on the button next to the merger mode button.
+After finishing the proof-reading, a long-running job can be started to apply the merging of segments into a new dataset with the same layers. The job can be started via a modal. The modal can be opened by clicking on the button next to the merger mode button.
 
 ![Button to open the Merger mode long-running job modal](./images/start_merger_mode_job_modal_button.jpg)
 ![Modal to start the Merger mode long-running job](./images/start_merger_mode_job_modal.jpg)
@@ -57,9 +66,9 @@ After finishing the proof-reading, a long-running job can be started to apply th
 webKnossos supports volumetric flood fills (3D) to relabel a segment with a new ID. Instead of having the relabel segment slice-by-slice, webKnossos can do this for you. This operation allows you to fix both split and merge errors:
 
 - For split errors: Combine two segments by relabeling one segment with the ID of the other. Since this operation is fairly compute intensive you might be better of with the `Merger Mode`, explained above.
-- For merge errors: You have to manually split two segments at their intersection/border, e.g. a cell boundary. Use the eraser brush and make sure to establish a clear cut between both segments on a slice-by-slice basis. Both segments must not touch any longer. Create a new segment ID from the toolbar and apply to one of the partial segments that you just devided. 
+- For merge errors: You have to manually split two segments at their intersection/border, e.g. a cell boundary. Use the eraser brush and make sure to establish a clear cut between both segments on a slice-by-slice basis. Both segments must not touch any longer. Create a new segment ID from the toolbar and apply to one of the partial segments that you just divided. 
 
-Due to perfomance reasons, 3D flood-fills only work in a small, local bounding box. webKnossos will add bounding box around the affected area. In order to truly propagate the new segment ID(s) throughout a whole dataset, you can trigger a webKnossos job to apply this change globally. From the `BBox` tab in the right-hand menu, press the "Globalize Flood-Fill" button. Make sure to do all local fill operations first and apply them all at once.
+Due to performance reasons, 3D flood-fills only work in a small, local bounding box. webKnossos will add bounding box around the affected area. In order to truly propagate the new segment ID(s) throughout a whole dataset, you can trigger a webKnossos job to apply this change globally. From the `BBox` tab in the right-hand menu, press the "Globalize Flood-Fill" button. Make sure to do all local fill operations first and apply them all at once.
 
 Check the `Processing Jobs` page from the `Admin` menu at the top of the screen to track progress or cancel the operation. The finished, processed dataset will appear as new dataset in your dashboard.
 

frontend/javascripts/oxalis/view/action-bar/toolbar_view.tsx

@@ -142,8 +142,12 @@ function RadioButtonWithTooltip({
         disabled={disabled}
         {...props}
         onClick={(evt) => {
-          if (document.activeElement) (document.activeElement as HTMLElement).blur();
-          if (onClick) onClick(evt);
+          if (document.activeElement) {
+            (document.activeElement as HTMLElement).blur();
+          }
+          if (onClick) {
+            onClick(evt);
+          }
         }}
       />
     </Tooltip>
@@ -226,10 +230,12 @@ function AdditionalSkeletonModesButtons() {
   const isMergerModeEnabled = useSelector(
     (state: OxalisState) => state.temporaryConfiguration.isMergerModeEnabled,
   );
-  const [showApplyMergerModeModal, setShowApplyMergerModeModal] = useState<boolean>(false);
+  const [showMaterializeVolumeAnnotationModal, setShowMaterializeVolumeAnnotationModal] =
+    useState<boolean>(false);
   const isNewNodeNewTreeModeOn = useSelector(
     (state: OxalisState) => state.userConfiguration.newNodeNewTree,
   );
+
   const toggleNewNodeNewTreeMode = () =>
     dispatch(updateUserSettingAction("newNodeNewTree", !isNewNodeNewTreeModeOn));
 
@@ -266,17 +272,19 @@ function AdditionalSkeletonModesButtons() {
         </ButtonComponent>
       </Tooltip>
       {features().jobsEnabled && isMergerModeEnabled && (
-        <Tooltip title="Apply this merger mode tracing.">
+        <Tooltip title="Materialize this merger mode annotation into a new dataset.">
           <ButtonComponent
             style={narrowButtonStyle}
-            onClick={() => setShowApplyMergerModeModal(true)}
+            onClick={() => setShowMaterializeVolumeAnnotationModal(true)}
           >
             <ExportOutlined />
           </ButtonComponent>
         </Tooltip>
       )}
-      {features().jobsEnabled && showApplyMergerModeModal && (
-        <MaterializeVolumeAnnotationModal handleClose={() => setShowApplyMergerModeModal(false)} />
+      {features().jobsEnabled && showMaterializeVolumeAnnotationModal && (
+        <MaterializeVolumeAnnotationModal
+          handleClose={() => setShowMaterializeVolumeAnnotationModal(false)}
+        />
       )}
     </React.Fragment>
   );

frontend/javascripts/oxalis/view/left-border-tabs/layer_settings_tab.tsx

@@ -474,7 +474,9 @@ class DatasetSettings extends React.PureComponent<DatasetSettingsProps, State> {
             {hasHistogram && !isDisabled ? this.getEditMinMaxButton(layerName, isInEditMode) : null}
           </div>
           <div className="flex-item">
-            {isVolumeTracing && !isDisabled ? this.getMergeWithFallbackLayerButton(layer) : null}
+            {isVolumeTracing && !isDisabled && maybeFallbackLayer != null
+              ? this.getMergeWithFallbackLayerButton(layer)
+              : null}
           </div>
           <div className="flex-item">
             {this.getFindDataButton(layerName, isDisabled, isColorLayer, maybeVolumeTracing)}

frontend/javascripts/oxalis/view/right-border-tabs/starting_job_modals.tsx

@@ -9,7 +9,11 @@ import {
 } from "admin/admin_rest_api";
 import { useSelector } from "react-redux";
 import { DatasetNameFormItem } from "admin/dataset/dataset_components";
-import { getColorLayers, getSegmentationLayers } from "oxalis/model/accessors/dataset_accessor";
+import {
+  getColorLayers,
+  getSegmentationLayers,
+  getDataLayers,
+} from "oxalis/model/accessors/dataset_accessor";
 import {
   getReadableNameByVolumeTracingId,
   getActiveSegmentationTracingLayer,
@@ -23,10 +27,16 @@ import { computeArrayFromBoundingBox, rgbToHex } from "libs/utils";
 import { getBaseSegmentationName } from "oxalis/view/right-border-tabs/segments_tab/segments_view_helper";
 
 const { ThinSpace } = Unicode;
-const jobNameToImagePath = {
+const enum JobNames {
+  NEURON_INFERRAL = "neuron inferral",
+  NUCLEI_INFERRAL = "nuclei inferral",
+  MATERIALIZE_VOLUME_ANNOTATION = "materialize volume annotation",
+  GLOBALIZE_FLODDFILLS = "globalization of the floodfill operation(s)",
+}
+const jobNameToImagePath: Record<JobNames, string | null> = {
   "neuron inferral": "neuron_inferral_example.jpg",
   "nuclei inferral": "nuclei_inferral_example.jpg",
-  "materializing this volume annotation": "materialize_volume_annotation_example.jpg",
+  "materialize volume annotation": "materialize_volume_annotation_example.jpg",
   "globalization of the floodfill operation(s)": null,
 };
 type Props = {
@@ -177,14 +187,12 @@ function BoundingBoxSelectionFormItem({
 
 type OutputSegmentationLayerNameProps = {
   hasOutputSegmentationLayer: boolean;
-  layers: APIDataLayer[];
-  additionalAllowedNames: string[];
+  notAllowedLayerNames: string[];
 };
 
 export function OutputSegmentationLayerNameFormItem({
   hasOutputSegmentationLayer,
-  layers,
-  additionalAllowedNames,
+  notAllowedLayerNames,
 }: OutputSegmentationLayerNameProps) {
   return (
     <Form.Item
@@ -200,10 +208,7 @@ export function OutputSegmentationLayerNameFormItem({
         },
         {
           validator: async (_rule, newOutputLayerName) => {
-            if (
-              layers.some((layer) => layer.name === newOutputLayerName) &&
-              !additionalAllowedNames.includes(newOutputLayerName)
-            ) {
+            if (notAllowedLayerNames.includes(newOutputLayerName)) {
               const reason =
                 "This name is already used by another segmentation layer of this dataset.";
               return Promise.reject(reason);
@@ -232,6 +237,7 @@ function StartingJobModal(props: StartingJobModalProps) {
   const tracing = useSelector((state: OxalisState) => state.tracing);
   const activeUser = useSelector((state: OxalisState) => state.activeUser);
   const layers = chooseSegmentationLayer ? getSegmentationLayers(dataset) : getColorLayers(dataset);
+  const allLayer = getDataLayers(dataset);
 
   const startJob = async ({
     layerName,
@@ -297,7 +303,19 @@ function StartingJobModal(props: StartingJobModalProps) {
     initialOutputSegmentationLayerName || "segmentation"
   }_corrected`;
   // TODO: Other jobs also have an output segmentation layer. The names for these jobs should also be configurable.
-  const hasOutputSegmentationLayer = jobName === "materializing this volume annotation";
+  const hasOutputSegmentationLayer = jobName === JobNames.MATERIALIZE_VOLUME_ANNOTATION;
+  const notAllowedOutputLayerNames = allLayer
+    .filter((layer) => {
+      // Filtering out the layer that is currently selected as this name is allowed
+      // because the output layer overwrites the selected layer.
+      const currentSelectedVolumeLayerName = form.getFieldValue("layerName") || initialLayerName;
+      return (
+        getReadableNameOfVolumeLayer(layer, tracing) !== currentSelectedVolumeLayerName &&
+        layer.name !== currentSelectedVolumeLayerName
+      );
+    })
+    .map((layer) => getReadableNameOfVolumeLayer(layer, tracing) || layer.name);
+
   return (
     <Modal title={title} onCancel={handleClose} visible width={700} footer={null}>
       {description}
@@ -337,8 +355,7 @@ function StartingJobModal(props: StartingJobModalProps) {
         />
         <OutputSegmentationLayerNameFormItem
           hasOutputSegmentationLayer={hasOutputSegmentationLayer}
-          layers={layers}
-          additionalAllowedNames={[form.getFieldValue("layerName")]}
+          notAllowedLayerNames={notAllowedOutputLayerNames}
         />
         <BoundingBoxSelectionFormItem
           isBoundingBoxConfigurable={isBoundingBoxConfigurable}
@@ -359,7 +376,7 @@ export function NucleiInferralModal({ handleClose }: Props) {
   return (
     <StartingJobModal
       handleClose={handleClose}
-      jobName="nuclei inferral"
+      jobName={JobNames.NUCLEI_INFERRAL}
       title="Start a Nuclei Inferral"
       suggestedDatasetSuffix="with_nuclei"
       jobApiCall={async ({ newDatasetName, selectedLayer: colorLayer }) =>
@@ -395,7 +412,7 @@ export function NeuronInferralModal({ handleClose }: Props) {
   return (
     <StartingJobModal
       handleClose={handleClose}
-      jobName="neuron inferral"
+      jobName={JobNames.NEURON_INFERRAL}
       title="Start a Neuron Inferral"
       suggestedDatasetSuffix="with_reconstructed_neurons"
       isBoundingBoxConfigurable
@@ -446,17 +463,24 @@ export function MaterializeVolumeAnnotationModal({
   const tracing = useSelector((state: OxalisState) => state.tracing);
   const activeSegmentationTracingLayer = useSelector(getActiveSegmentationTracingLayer);
   const fixedSelectedLayer = selectedVolumeLayer || activeSegmentationTracingLayer;
+  const readableVolumeLayerName =
+    fixedSelectedLayer && getReadableNameOfVolumeLayer(fixedSelectedLayer, tracing);
+  const hasFallbackLayer =
+    fixedSelectedLayer && "tracingId" in fixedSelectedLayer
+      ? fixedSelectedLayer.fallbackLayer != null
+      : false;
   const isMergerModeEnabled = useSelector(
     (state: OxalisState) => state.temporaryConfiguration.isMergerModeEnabled,
   );
   let description = (
     <p>
       Start a job that takes the current state of this volume annotation and materializes it into a
-      new dataset. All annotations done on the
-      {` "${fixedSelectedLayer && getReadableNameOfVolumeLayer(fixedSelectedLayer, tracing)}" `}
-      volume layer will be merged with the data of the fallback layer.
+      new dataset.
+      {hasFallbackLayer
+        ? ` All annotations done on the "${readableVolumeLayerName}" volume layer will be merged with the data of the fallback layer. `
+        : null}
       {isMergerModeEnabled
-        ? " Additionally, the skeletons will be used to merge segments as merger mode is active. "
+        ? " Since merger mode is currently active, in the new output dataset the segments connected via skeleton nodes will be merged. "
         : " "}
       Please enter the name of the output dataset and the output segmentation layer.
     </p>
@@ -465,8 +489,9 @@ export function MaterializeVolumeAnnotationModal({
     description = (
       <p>
         Start a job that takes the current state of this merger mode tracing and materializes it
-        into a new dataset. In the new output dataset the segments connected via skeleton nodes will
-        be merged. Please enter the name of the output dataset and the output segmentation layer.
+        into a new dataset. Since merger mode is currently active, in the new output dataset the
+        segments connected via skeleton nodes will be merged. Please enter the name of the output
+        dataset and the output segmentation layer.
       </p>
     );
   }
@@ -475,7 +500,7 @@ export function MaterializeVolumeAnnotationModal({
     <StartingJobModal
       handleClose={handleClose}
       title="Start Materializing this Volume Annotation"
-      jobName="materializing this volume annotation"
+      jobName={JobNames.MATERIALIZE_VOLUME_ANNOTATION}
       suggestedDatasetSuffix="with_merged_segmentation"
       chooseSegmentationLayer
       fixedSelectedLayer={fixedSelectedLayer}
@@ -513,7 +538,7 @@ export function StartGlobalizeFloodfillsModal({ handleClose }: Props) {
     <StartingJobModal
       handleClose={handleClose}
       title="Start Globalizing of the Floodfill Operation(s)"
-      jobName="globalization of the floodfill operation(s)"
+      jobName={JobNames.GLOBALIZE_FLODDFILLS}
       suggestedDatasetSuffix="with_floodfills"
       chooseSegmentationLayer
       jobApiCall={async ({ newDatasetName, selectedLayer: segmentationLayer }) => {