Batch of fixes (#1403)

* Fixed bug when job cannot be opened

* Fixed bug when deactivated shape is still highlighted

* Fixed Error: 'AttributeError: 'tuple' object has no attribute 'read'

* Fixed: wrong semi-automatic segmentation near edges of an image

* Updated changelog

Co-authored-by: Nikita Manovich <40690625+nmanovic@users.noreply.github.com>

CHANGELOG.md

@@ -44,6 +44,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - VOC format exports Upper case labels correctly in lower case (https://github.com/opencv/cvat/pull/1379)
 - Fixed polygon exporting bug in COCO dataset (https://github.com/opencv/cvat/issues/1387)
 - Task creation from remote files (https://github.com/opencv/cvat/pull/1392)
+- Job cannot be opened in some cases when the previous job was failed during opening (https://github.com/opencv/cvat/issues/1403)
+- Deactivated shape is still highlighted on the canvas (https://github.com/opencv/cvat/issues/1403)
+- AttributeError: 'tuple' object has no attribute 'read' in ReID algorithm (https://github.com/opencv/cvat/issues/1403)
+- Wrong semi-automatic segmentation near edges of an image (https://github.com/opencv/cvat/issues/1403)
 - Git repos paths (https://github.com/opencv/cvat/pull/1400)
 
 ### Security

cvat-ui/src/actions/annotation-actions.ts

@@ -923,7 +923,9 @@ export function getJobAsync(
 
             dispatch({
                 type: AnnotationActionTypes.GET_JOB,
-                payload: {},
+                payload: {
+                    requestedId: jid,
+                },
             });
 
             const loadJobEvent = await logger.log(

cvat-ui/src/components/annotation-page/standard-workspace/canvas-wrapper.tsx

@@ -160,9 +160,6 @@ export default class CanvasWrapperComponent extends React.PureComponent<Props> {
         if (prevProps.activatedStateID !== null
             && prevProps.activatedStateID !== activatedStateID) {
             canvasInstance.activate(null);
-        }
-
-        if (activatedStateID) {
             const el = window.document.getElementById(`cvat_canvas_shape_${prevProps.activatedStateID}`);
             if (el) {
                 (el as any).instance.fill({ opacity: opacity / 100 });

cvat-ui/src/containers/annotation-page/annotation-page.tsx

@@ -33,6 +33,7 @@ function mapStateToProps(state: CombinedState, own: OwnProps): StateToProps {
     const {
         annotation: {
             job: {
+                requestedId,
                 instance: job,
                 fetching,
             },
@@ -41,7 +42,7 @@ function mapStateToProps(state: CombinedState, own: OwnProps): StateToProps {
     } = state;
 
     return {
-        job: !job || jobID === job.id ? job : null,
+        job: jobID === requestedId ? job : null,
         fetching,
         workspace,
     };

cvat-ui/src/reducers/annotation-reducer.ts

@@ -35,6 +35,7 @@ const defaultState: AnnotationState = {
     },
     job: {
         labels: [],
+        requestedId: null,
         instance: null,
         attributes: {},
         fetching: false,
@@ -105,6 +106,8 @@ export default (state = defaultState, action: AnyAction): AnnotationState => {
                 ...state,
                 job: {
                     ...state.job,
+                    instance: null,
+                    requestedId: action.payload.requestedId,
                     fetching: true,
                 },
             };

cvat-ui/src/reducers/interfaces.ts

@@ -317,6 +317,7 @@ export interface AnnotationState {
     };
     job: {
         labels: any[];
+        requestedId: number | null;
         instance: any | null | undefined;
         attributes: Record<number, any[]>;
         fetching: boolean;

cvat/apps/dextr_segmentation/dextr.py

@@ -48,11 +48,19 @@ class DEXTR_HANDLER:
         image = PIL.Image.open(image[0])
         numpy_image = np.array(image)
         points = np.asarray([[int(p["x"]), int(p["y"])] for p in points], dtype=int)
+
+        # Padding mustn't be more than the closest distance to an edge of an image
+        [width, height] = numpy_image.shape[:2]
+        x_values = points[:, 0]
+        y_values = points[:, 1]
+        [min_x, max_x] = [np.min(x_values), np.max(x_values)]
+        [min_y, max_y] = [np.min(y_values), np.max(y_values)]
+        padding = min(min_x, min_y, width - max_x, height - max_y, _DEXTR_PADDING)
         bounding_box = (
-            max(min(points[:, 0]) - _DEXTR_PADDING, 0),
-            max(min(points[:, 1]) - _DEXTR_PADDING, 0),
-            min(max(points[:, 0]) + _DEXTR_PADDING, numpy_image.shape[1] - 1),
-            min(max(points[:, 1]) + _DEXTR_PADDING, numpy_image.shape[0] - 1)
+            max(min(points[:, 0]) - padding, 0),
+            max(min(points[:, 1]) - padding, 0),
+            min(max(points[:, 0]) + padding, numpy_image.shape[1] - 1),
+            min(max(points[:, 1]) + padding, numpy_image.shape[0] - 1)
         )
 
         # Prepare an image
@@ -61,7 +69,7 @@ class DEXTR_HANDLER:
             interpolation = cv2.INTER_CUBIC).astype(np.float32)
 
         # Make a heatmap
-        points = points - [min(points[:, 0]), min(points[:, 1])] + [_DEXTR_PADDING, _DEXTR_PADDING]
+        points = points - [min(points[:, 0]), min(points[:, 1])] + [padding, padding]
         points = (points * [_DEXTR_SIZE / numpy_cropped.shape[1], _DEXTR_SIZE / numpy_cropped.shape[0]]).astype(int)
         heatmap = np.zeros(shape=resized.shape[:2], dtype=np.float64)
         for point in points:

cvat/apps/reid/reid.py

@@ -150,7 +150,7 @@ class ReID:
         job = rq.get_current_job()
         box_tracks = {}
 
-        next_image = cv2.imdecode(numpy.fromstring(next(self.__frame_iter).read(), numpy.uint8), cv2.IMREAD_COLOR)
+        next_image = cv2.imdecode(numpy.fromstring((next(self.__frame_iter)[0]).read(), numpy.uint8), cv2.IMREAD_COLOR)
         for idx, (cur_frame, next_frame) in enumerate(list(zip(frames[:-1], frames[1:]))):
             job.refresh()
             if "cancel" in job.meta:
@@ -172,7 +172,7 @@ class ReID:
                 continue
 
             cur_image = next_image
-            next_image = cv2.imdecode(numpy.fromstring(next(self.__frame_iter).read(), numpy.uint8), cv2.IMREAD_COLOR)
+            next_image = cv2.imdecode(numpy.fromstring((next(self.__frame_iter)[0]).read(), numpy.uint8), cv2.IMREAD_COLOR)
             difference_matrix = self.__compute_difference_matrix(cur_boxes, next_boxes, cur_image, next_image)
             cur_idxs, next_idxs = linear_sum_assignment(difference_matrix)
             for idx, cur_idx in enumerate(cur_idxs):