Geometry

app.ui.qt.widgets.preview.layer_bbox.geometry

View source code

Geometry and coordinate-space helpers for bbox overlay interactions.

This module converts between image space and widget space, and provides proximity lookup for selecting the nearest bbox around the cursor.

See Also

• BBoxState — tracks selected and edited geometry
• get_drag_mode_for_rect_at_pos — detects handle hits

Layer: Presentation. Context: Contains concrete Qt presentation components and supporting utilities.

image_rect_to_widget_space(x1: int, y1: int, x2: int, y2: int, pixmap_rect: QRect, image_size: QSize | None) -> QRect

Convert image-space coordinates into a widget-space rectangle.

PARAMETER	DESCRIPTION
x1	Left image-space coordinate. TYPE: int
y1	Top image-space coordinate. TYPE: int
x2	Right image-space coordinate. TYPE: int
y2	Bottom image-space coordinate. TYPE: int
pixmap_rect	Displayed image rectangle in widget coordinates. TYPE: QRect
image_size	Original image dimensions, or None if unavailable. TYPE: QSize | None

RETURNS	DESCRIPTION
QRect	Normalized widget-space rectangle, or an empty rectangle if conversion is unavailable.

widget_rect_to_image_space(rect: QRect, pixmap_rect: QRect, image_size: QSize | None) -> BBoxXYXYTuple

Convert a widget-space rectangle into clamped image-space coordinates.

PARAMETER	DESCRIPTION
rect	Bounding rectangle in widget coordinates. TYPE: QRect
pixmap_rect	Displayed image rectangle in widget coordinates. TYPE: QRect
image_size	Original image dimensions, or None if unavailable. TYPE: QSize | None

RETURNS	DESCRIPTION
BBoxXYXYTuple	(x1, y1, x2, y2) coordinates clamped to the image bounds, or zeros if conversion is unavailable.

find_bbox_at_pos(active_bboxes: dict[str, BBoxXYXYTuple], pos: QPoint, pixmap_rect: QRect, image_size: QSize | None, snap_radius_sq: int = 2500) -> tuple[str | None, QRect | None]

Find a bbox at the cursor or snap to the nearest bbox center within radius.

PARAMETER	DESCRIPTION
active_bboxes	Box keys mapped to image-space coordinate tuples. TYPE: dict[str, BBoxXYXYTuple]
pos	Cursor position in widget coordinates. TYPE: QPoint
pixmap_rect	Displayed image rectangle in widget coordinates. TYPE: QRect
image_size	Original image dimensions, or None if unavailable. TYPE: QSize | None
snap_radius_sq	Maximum squared widget-pixel distance from a box center for snapping. TYPE: int DEFAULT: 2500

RETURNS	DESCRIPTION
tuple[str | None, QRect | None]	(box_key, widget_rect) for a containing or nearby box; (None, None) if none qualifies.

image_rect_to_widget_space_with_transforms(x1: int, y1: int, x2: int, y2: int, pixmap_rect: QRect, image_size: QSize | None, zoom: float = 1.0, pan_x: float = 0.0, pan_y: float = 0.0) -> QRect

Convert image-space box coordinates using the displayed image rectangle.

The video widget already applies zoom/pan to the displayed image, so the overlay should use that same pixmap rect rather than re-scaling the bbox rectangle again.

PARAMETER	DESCRIPTION
x1	Left image-space coordinate. TYPE: int
y1	Top image-space coordinate. TYPE: int
x2	Right image-space coordinate. TYPE: int
y2	Bottom image-space coordinate. TYPE: int
pixmap_rect	Displayed image rectangle in widget coordinates. TYPE: QRect
image_size	Original image dimensions, or None if unavailable. TYPE: QSize | None
zoom	Retained for API compatibility; coordinate conversion uses pixmap_rect. TYPE: float DEFAULT: 1.0
pan_x	Retained for API compatibility; coordinate conversion uses pixmap_rect. TYPE: float DEFAULT: 0.0
pan_y	Retained for API compatibility; coordinate conversion uses pixmap_rect. TYPE: float DEFAULT: 0.0

RETURNS	DESCRIPTION
QRect	Normalized widget-space rectangle.

find_bbox_at_pos_with_transforms(active_bboxes: dict[str, BBoxXYXYTuple], pos: QPoint, pixmap_rect: QRect, image_size: QSize | None, zoom: float = 1.0, pan_x: float = 0.0, pan_y: float = 0.0, snap_radius_sq: int = 2500) -> tuple[str | None, QRect | None]

Find a bbox at the cursor using the viewport-aligned widget-space rect.

PARAMETER	DESCRIPTION
active_bboxes	Box keys mapped to image-space coordinate tuples. TYPE: dict[str, BBoxXYXYTuple]
pos	Cursor position in widget coordinates. TYPE: QPoint
pixmap_rect	Displayed image rectangle in widget coordinates. TYPE: QRect
image_size	Original image dimensions, or None if unavailable. TYPE: QSize | None
zoom	Retained for API compatibility; the supplied pixmap_rect already reflects zoom. TYPE: float DEFAULT: 1.0
pan_x	Retained for API compatibility; the supplied pixmap_rect already reflects panning. TYPE: float DEFAULT: 0.0
pan_y	Retained for API compatibility; the supplied pixmap_rect already reflects panning. TYPE: float DEFAULT: 0.0
snap_radius_sq	Maximum squared widget-pixel distance from a box center for snapping. TYPE: int DEFAULT: 2500

RETURNS	DESCRIPTION
tuple[str | None, QRect | None]	(box_key, widget_rect) for a containing or nearby box; (None, None) if none qualifies.

find_bbox_keys_intersecting_rect(active_bboxes: dict[str, BBoxXYXYTuple], selection_rect: QRect, pixmap_rect: QRect, image_size: QSize | None) -> list[str]

Return box keys whose widget-space bounds intersect a marquee rectangle.

PARAMETER	DESCRIPTION
active_bboxes	Box keys mapped to image-space coordinate tuples. TYPE: dict[str, BBoxXYXYTuple]
selection_rect	Marquee rectangle in widget coordinates. TYPE: QRect
pixmap_rect	Displayed image rectangle in widget coordinates. TYPE: QRect
image_size	Original image dimensions, or None if unavailable. TYPE: QSize | None

RETURNS	DESCRIPTION
list[str]	Keys of boxes whose widget-space rectangles intersect the marquee.