Added support to show the transform handler for the selected object

When using the ObjectDetection effect, it's now possible to select one detected object and update it's properties through it's transform handler.

Author: BrennoCaldato

src/windows/video_widget.py

@@ -310,20 +310,20 @@ def paintEvent(self, event, *args):
                 # Get properties of clip at current frame
                 raw_properties_effect = json.loads(self.transforming_effect_object.PropertiesJSON(clip_frame_number))
                 for effect_key in raw_properties_effect.keys():
-                    if effect_key.startswith("box_id"):
+                    if effect_key.startswith("x1-"):
                         # Get current tracked object index
                         tracked_obj_idx = effect_key.split("-", 2)[1]
                         # Check if the tracked object is visible in this frame
-                        visible = raw_properties_effect.get('visible-'+tracked_obj_idx).get('value')
+                        visible = raw_properties_effect['visible-'+tracked_obj_idx]['value']
                         if visible:
                             # Get current bounding box values
-                            r = raw_properties_effect.get('rotation-'+tracked_obj_idx).get('value')
-                            x1 = raw_properties_effect.get('x1-'+tracked_obj_idx).get('value')
-                            y1 = raw_properties_effect.get('y1-'+tracked_obj_idx).get('value')
-                            x2 = raw_properties_effect.get('x2-'+tracked_obj_idx).get('value')
-                            y2 = raw_properties_effect.get('y2-'+tracked_obj_idx).get('value')
+                            rotation = raw_properties_effect['rotation-'+tracked_obj_idx]['value']
+                            x1 = raw_properties_effect['x1-'+tracked_obj_idx]['value']
+                            y1 = raw_properties_effect['y1-'+tracked_obj_idx]['value']
+                            x2 = raw_properties_effect['x2-'+tracked_obj_idx]['value']
+                            y2 = raw_properties_effect['y2-'+tracked_obj_idx]['value']
                             self.drawTransformHandler(painter, sx, sy, source_width, source_height, origin_x, origin_y,
-                                x1, y1, x2, y2, r)
+                                x1, y1, x2, y2, rotation)
             else:
                 self.drawTransformHandler(painter, sx, sy, source_width, source_height, origin_x, origin_y)
 
@@ -828,7 +828,7 @@ def mouseMoveEvent(self, event):
             raw_properties = json.loads(self.transforming_effect_object.PropertiesJSON(clip_frame_number))
 
             for raw_property_key in raw_properties.keys():
-                if raw_property_key.startswith("box_id"):
+                if raw_property_key.startswith("x1-"):
                     # Get current tracked object index
                     tracked_obj_idx = raw_property_key.split("-", 2)[1]
                     # Check if the tracked object is visible in this frame

src/windows/views/properties_tableview.py

@@ -419,6 +419,34 @@ def contextMenuEvent(self, event):
                 self.choices = []
                 self.menu_reset = False
 
+            # Handle selected object options (ObjectDetection effect)
+            if property_key == "selected_object_index" and not self.choices:
+                # Add visible objects as choices - initialize with the first visible object
+                object_index_choices = [{
+                    "name": str(0),
+                    "value": str(0),
+                    "selected": False,
+                    "icon": QIcon()
+                }]
+                # Get all visible object's indexes
+                timeline_instance = get_app().window.timeline_sync.timeline
+                # Instantiate the effect
+                effect = timeline_instance.GetClipEffect(clip_id)
+                # Get effect's properties JSON
+                effect_properties = json.loads(effect.PropertiesJSON(frame_number))
+                for effect_key in effect_properties.keys():
+                    if effect_key.startswith("visible-"):
+                        if effect_properties[effect_key]["value"]:
+                            # Get visible object index
+                            object_index = effect_key.split("-", 2)[1]
+                            object_index_choices.append({
+                                "name": object_index,
+                                "value": object_index,
+                                "selected": False,
+                                "icon": QIcon()
+                            })
+                self.choices.append({"name": _("Detected Objects"), "value": object_index_choices, "selected": False, "icon": None})
+
             # Handle clip attach options
             if property_key == "attached_id" and not self.choices:
                 # Add all Clips as choices - initialize with None
@@ -457,20 +485,21 @@ def contextMenuEvent(self, event):
                             # Iterate through the effect properties
                             for effect_key in effect.keys():
                                 # Check if the effect has the "box_id" property, i.e., is the Tracker or ObjectDetection effect
-                                if effect_key.startswith("box_id"):
-                                    # Get effect's JSON properties for this frame
-                                    effect_instance = timeline_instance.GetClipEffect(effect["id"])
-                                    raw_properties_effect = json.loads(effect_instance.PropertiesJSON(frame_number))
+                                if effect_key.startswith("box_id-"):
                                     # Get current tracked object index
                                     tracked_obj_idx = effect_key.split("-", 2)[1]
-                                    # Check if the tracked object is visible in this frame
+                                    effect_instance = timeline_instance.GetClipEffect(effect["id"])
+                                    raw_properties_effect = json.loads(effect_instance.PropertiesJSON(frame_number))
                                     visible = raw_properties_effect.get('visible-'+tracked_obj_idx).get('value')
+                                    # Check if the tracked object is visible in this frame
                                     if visible:
                                         # Get the Tracked Object properties
-                                        x1 = raw_properties_effect.get('x1-'+tracked_obj_idx).get('value')
-                                        y1 = raw_properties_effect.get('y1-'+tracked_obj_idx).get('value')
-                                        x2 = raw_properties_effect.get('x2-'+tracked_obj_idx).get('value')
-                                        y2 = raw_properties_effect.get('y2-'+tracked_obj_idx).get('value')
+                                        tracked_obj_id = effect['box_id-'+tracked_obj_idx]
+                                        tracked_obj_properties = json.loads(timeline_instance.GetTrackedObjectValues(tracked_obj_id, 0))
+                                        x1 = tracked_obj_properties['x1']
+                                        y1 = tracked_obj_properties['y1']
+                                        x2 = tracked_obj_properties['x2']
+                                        y2 = tracked_obj_properties['y2']
                                         # Get the tracked object's icon from the clip's icon
                                         tracked_object_icon = icon_pixmap.copy(QRect(x1*icon_size, y1*icon_size, (x2-x1)*icon_size, (y2-y1)*icon_size/2)).scaled(icon_size, icon_size)
                                         tracked_objects.append({"name": effect[effect_key],