I'm implementing an image viewer on an embedded platform. The hardware is a sort of tablet and has a touch screen as input device. The Qt version I'm using is 5.4.3.
The QGraphicsView is used to display a QGraphicsScene which contains a QGraphicsPixmapItem. The QGraphicsPixmapItem containts the pixmap to display.
The relevant part of the code is the following:
void MyGraphicsView::pinchTriggered(QPinchGesture *gesture)
{
QPinchGesture::ChangeFlags changeFlags = gesture->changeFlags();
if (changeFlags & QPinchGesture::ScaleFactorChanged) {
currentStepScaleFactor = gesture->totalScaleFactor();
}
if (gesture->state() == Qt::GestureFinished) {
scaleFactor *= currentStepScaleFactor;
currentStepScaleFactor = 1;
return;
}
// Compute the scale factor based on the current pinch level
qreal sxy = scaleFactor * currentStepScaleFactor;
// Get the pointer to the currently displayed picture
QList<QGraphicsItem *> listOfItems = items();
QGraphicsItem* item = listOfItems.at(0);
// Scale the picture
item.setScale(sxy);
// Adapt the scene to the scaled picture
setSceneRect(scene()->itemsBoundingRect());
}
As result of the pinch, the pixmap is scaled starting from the top-left corner of the view.
How to scale the pixmap respect to the center of the QPinchGesture?
From The Docs
The item is scaled around its transform origin point, which by default is (0, 0). You can select a different transformation origin by calling setTransformOriginPoint().
That function takes in a QPoint so you would need to find out your centre point first then set the origin point.
void QGraphicsItem::setTransformOriginPoint(const QPointF & origin)
My LineItem inheriting from QGraphicsLineItem can change its pen width.
I have created a boundingRect that uses the QGraphicsLineItem::boundingRect adjusted by pads that get calculated based on pen width and arrows. It works.
void LineItem::calculateStuff() // called on any change including pen width
{
qreal padLeft, padRight, padT;
padLeft = 0.5 * m_pen.width(); // if no arrows
padT = padLeft;
padRight = padLeft;
m_boundingRect = QGraphicsLineItem::boundingRect().adjusted(-padLeft, -padT, padRight, padT);
update();
}
QRectF LineItem::boundingRect() const
{
return m_boundingRect;
}
QPainterPath LineItem::shape() const
{
QPainterPath p;
p.addRect(m_boundingRect);
return p;
}
There is only one artifact that I get:
if I increase the pen width, then decrease it, I get traces:
these of course disappear as soon as i move mouse or any action (I had a hard time getting the screen shots)
As pretty as they are (seriously I consider them a "feature :-) ) - I am trying to eliminate them. I tried to remember previous bounding rectangle, and update the item with the previous bounding rectangle - i thought that was what the option was for - but it didn't work.
QRectF oldRect = selectedItem->boundingRect();
item->setItemPenWidth(p);
selectedItem->update(oldRect);
selectedItem->update();
My viewport has
setViewportUpdateMode(BoundingRectViewportUpdate);
If I change to
setViewportUpdateMode(FullViewportUpdate);
I don't get artifacts - but I think this will impact performance which is a major constraint.
How can I fix these artifacts - that only occur in that specific situation, decreasing pen width / decreasing bounding rect of line, without impacting performance ?
Simple fix... I had to add
prepareGeometryChange();
in my calculateStuff() function.
I have not seen any changes from this before, it is the first time I change my boundingRect that it does not update seamlessly.
I have a number of classes that inherit from QGraphicsItem, that get to be arranged in a certain way. For simplicity of calculations, I made the scenes, and items, centered in (0, 0) (with the boundingRect() having +/- coordinates).
QGraphicsTextItem subclass defies me, its pos() is relative to top left point.
I have tried a number of things to shift it so it centers in the text center (for example, the suggested solution here - the code referenced actually cuts my text and only shows the bottom left quarter).
I imagined that the solution should be something simple, like
void TextItem::paint(QPainter* painter, const QStyleOptionGraphicsItem* option, QWidget* widget)
{
painter->translate( -boundingRect().width()/2.0, -boundingRect().height()/2.0 );
QGraphicsTextItem::paint(painter, option, widget );
}
the above "sort of" works - but as I increase the item scale -> increase the font, the displayed item is cut off...
I tried to set the pos() - but the problem is, I still need to track the actual position on the scene, so I cannot just replace it.
A slightly unpleasant side effect - centering the QGraphicsView on the element does not work either.
How can I make my QGraphicsTextItem show its position relative to the center of the text ?
Edit: one of the experiments of changing the boundingRect():
QRectF TextItem::boundingRect() const
{
QRectF rect = QGraphicsTextItem::boundingRect();
rect.translate(QPointF(-rect.width()/2.0, -rect.height()/2.0));
return rect;
}
I had to shift the initial position, as well as the resize, to trigger a new position - I was unable to do it in paint() because, as I thought from the start, any repaint would continuously recalculate the position.
Only the initial position needs to be adjusted - but as the font size (or style...) changes, its bounding rectangle also changes, so the position must be recalculated - based on previous position.
In the constructor,
setPos(- boundingRect().width()/2, - boundingRect().height()/2);
in the function that modifies item (font) size,
void TextItem::setSize(int s)
{
QRectF oldRect = boundingRect();
QFont f;
f.setPointSize(s);
setFont(f);
if(m_scale != s)
{
m_scale = s;
qreal x = pos().x() - boundingRect().width()/2.0 + oldRect.width()/2.0;
qreal y = pos().y() - boundingRect().height()/2.0 + oldRect.height()/2.0;
setPos(QPointF(x, y));
}
}
I have a QGraphicsTextItem parented to a QGraphicsItem. I want the QGraphicsTextItem to always reside directly above the QGraphicsItem, but I also want the text to remain the same size when the scale factor goes below 1, i.e. the text remains the size it is at a scale factor of 1 even when the parent graphics item is scaled smaller. I have found that setting the QGraphicsItem::ItemIgnoresTransformations flag to true when the scale factor is below 1 does the trick for retaining the size.
But I can’t seem to find a way to get the position of the text to always remain above the QGraphicsItem. Is there a way to do this? I tried using deviceTransform () function, but the text still moved off of the QGraphicsItem as I scrolled out. What was worse is that some of the text items started “jiggling”, i.e. they started continuously changing their position ever so slightly, so that it looked like they were shaking. If this is the function I need to use, I guess I don’t know how to use it properly.
In the constructor of my QGraphicsItem I’ve added a QGraphicsTextItem:
fTextItem = new QGraphicsTextItem(getName(), this);
fTextItem->setFlag(QGraphicsItem::ItemIgnoresTransformations);
Here is code snippet from paint function of QGraphicsItem
qreal lod = painter->worldTransform().m22();
if(lod <= 1.0) {
fTextItem-setFlag(QGraphicsItem::ItemIgnoresTransformations);
fTextItem->setPos(fTextItem->deviceTransform(view-viewportTransform()).inverted().map(view->mapFromScene(mapToScene(0,0))));
} else {
fTextItem->setFlag(QGraphicsItem::ItemIgnoresTransformations, false);
fTextItem->setPos(0, 0);
}
My suggestion is to subclass QGraphicsSimpleTextItem in this manner:
class TextItem
: public QGraphicsSimpleTextItem
{
public:
TextItem(const QString &text)
: QGraphicsSimpleTextItem(text)
{
}
void paint(QPainter *painter,
const QStyleOptionGraphicsItem *option, QWidget *widget)
{
painter->translate(boundingRect().topLeft());
QGraphicsSimpleTextItem::paint(painter, option, widget);
painter->translate(-boundingRect().topLeft());
}
QRectF boundingRect() const
{
QRectF b = QGraphicsSimpleTextItem::boundingRect();
return QRectF(b.x()-b.width()/2.0, b.y()-b.height()/2.0,
b.width(), b.height());
}
};
QGraphicsSimpleTextItem *mText = new TextItem("Item");
scene()->addItem(mText);
mText->setFlag(QGraphicsItem::ItemIgnoresTransformations, true);
mText->setPos(itemToFollow->pos());
Disclaimer: this may be overkill for what you are trying to do. We had some additional restrictions in our project that made this solution the easiest for us.
We had to do something similar in a project, and it ended up being easiest for us to not use ItemIgnoresTransformations and instead roll our own transform. Here is the main function we use to create a translation-only (no scaling) transform for drawing an item at a specific location. You might be able to modify it for your usage.
static QTransform GenerateTranslationOnlyTransform(
const QTransform &original_transform,
const QPointF &target_point) {
// To draw the unscaled icons, we desire a transform with scaling factors
// of 1 and shearing factors of 0 and the appropriate translation such that
// our icon center ends up at the same point. According to the
// documentation, QTransform transforms a point in the plane to another
// point using the following formulas:
// x' = m11*x + m21*y + dx
// y' = m22*y + m12*x + dy
//
// For our new transform, m11 and m22 (scaling) are 1, and m21 and m12
// (shearing) are 0. Since we want x' and y' to be the same, we have the
// following equations:
// m11*x + m21*y + dx = x + dx[new]
// m22*y + m12*x + dy = y + dy[new]
//
// Thus,
// dx[new] = m11*x - x + m21*y + dx
// dy[new] = m22*y - y + m12*x + dy
qreal dx = original_transform.m11() * target_point.x()
- target_point.x()
+ original_transform.m21() * target_point.y()
+ original_transform.m31();
qreal dy = original_transform.m22() * target_point.y()
- target_point.y()
+ original_transform.m12() * target_point.x()
+ original_transform.m32();
return QTransform::fromTranslate(dx, dy);
}
To use, take the QPainter transform that is passed to the paint method and do something like:
painter->save();
painter->setTransform(GenerateTranslationOnlyTransform(painter->transform(),
some_point));
// Draw your item.
painter->restore();
I've found another solution, which does not involve messing with any transformations or by hand scaling/positioning. There is a hint in QGraphicsItem::ItemIgnoresTransformations flag description:
QGraphicsItem::ItemIgnoresTransformations
The item ignores inherited transformations (i.e., its position is
still anchored to its parent, but the parent or view rotation, zoom or
shear transformations are ignored). [...]
And that's the key! We need two items: a parent that will keep the relative position (without any flags set) and a child item that will do the drawing at parent's (0,0) point (with QGraphicsItem::ItemIgnoresTransformations flag set). Simple as that!
I've encapsulated this functionality into a single class - here is some code:
#include <QGraphicsItem>
#include <QPainter>
class SampleShape : public QGraphicsItem
{
private:
/* This class implements shape drawing */
class SampleShapeImpl : public QGraphicsItem
{
public:
SampleShapeImpl (qreal len, QGraphicsItem *parent = nullptr)
: QGraphicsItem(parent), m_len(len)
{
/* ignore transformations (!) */
setFlag(QGraphicsItem::ItemIgnoresTransformations);
}
QRectF boundingRect (void) const override
{
/* sample bounding rectangle */
return QRectF(-m_len, -m_len, m_len*2, m_len*2);
}
void paint (QPainter *painter, const QStyleOptionGraphicsItem *, QWidget *) override
{
/* draw a shape, (0,0) is an anchor */
painter->drawLine(0, -m_len, 0, m_len);
painter->drawLine(-m_len, 0, m_len, 0);
// ...
}
private:
qreal m_len; // sample shape parameter
};
public:
/* This is actually almost an empty class, you only need to set
* a position and pass any parameters to a SampleShapeImpl class.
*/
SampleShape (qreal x, qreal y, qreal len, QGraphicsItem *parent = nullptr)
: QGraphicsItem(parent), m_impl(len, this) // <-- IMPORTANT!!!
{
/* set position at (x, y), view transformations will apply */
setPos(x, y);
}
QRectF boundingRect (void) const override
{
return QRectF(); // it's just a point, no size
}
void paint (QPainter *, const QStyleOptionGraphicsItem *, QWidget *) override
{
// empty, drawing is done in SampleShapeImpl
}
private:
SampleShapeImpl m_impl;
};
Great answer by Dave Mateer! I had the problem that I wanted to define a different scale factor at different zoom levels. This is how I did it:
void MyGraphicsItem::paint(QPainter * painter, const QStyleOptionGraphicsItem* option, QWidget* widget)
{
//save painter for later operations
painter->save();
QTransform originalTransform = painter->transform();
QPointF originalCenter = rect().center();
qreal dx = originalTransform.m11() * originalCenter.x() + originalTransform.m21() * originalCenter.y() + originalTransform.m31();
qreal dy = originalTransform.m22() * originalCenter.y() + originalTransform.m12() * originalCenter.x() + originalTransform.m32();
//normally our target scale factor is 1, meaning the item has keeps its size, regardless of zoom
//we adjust the scale factor though when the item is smaller than one pixel in comparison to the background image
qreal factor = 1.0;
//check if scale factor if bigger that the item size, and thus it occupies less that a pixel in comparision to the background image
if (rect().width() < originalTransform.m11()) {
//calculate adjusted scale factor
factor = originalTransform.m11() / rect().width();
}
//adjust position according to scale factor
dx -= factor * originalCenter.x();
dy -= factor * originalCenter.y();
//set the new transform for painting
painter->setTransform(QTransform::fromScale(factor, factor) * QTransform::fromTranslate(dx, dy));
//now paint...
QGraphicsXYZItem::paint(painter, option, widget);
//restore original painter
painter->restore();
}
You do need to adjust the bounding rectangle too in that case:
QRectF MyGraphicsItem::boundingRect() const
{
QRectF rect = QGraphicsEllipseItem::boundingRect();
//this is a bit hackish, let me know if you know another way...
if (scene() != NULL && scene()->views().at(0) != NULL)
{
//get viewport transform
QTransform itemTransform = scene()->views().at(0)->transform();
QPointF originalCenter = rect.center();
//calculate back-projected original size of item
qreal realSizeX = rect.width() / itemTransform.m11();
qreal realSizeY = rect.height() / itemTransform.m11();
//check if scale factor is bigger that the item size, and thus it occupies less that a pixel in comparison
//to the background image and adjust size back to equivalent of 1 pixel
realSizeX = realSizeX < 1.0 ? 1.0 : realSizeX;
realSizeY = realSizeY < 1.0 ? 1.0 : realSizeY;
//set adjusted position and size according to scale factor
rect = QRectF(rect.center().x() - realSizeX / 2.0, rect.center().y() - realSizeY / 2.0, realSizeX, realSizeY);
}
return rect;
}
With this solution the item work very well in my case.
Adding to Dave Mateer's answer, I think it'd be helpful to add that in some scenario, you should also maintain proper bounding rectangle (as well as shape) of the object. For me, I need to modify boundingRect() a little too for proper object selection behavior. Remember that the bounding rect of the object will be scaled and transformed as usual if we do NOT use ItemIgnoresTransformations flag. So we also need to rescale the boundingRect to maintain the view independence effect.
To maintain the view-independent bounding rectangle turns out to be quite easy: just grab the scaling factor from deviceTransform(m_view->viewportTransform()).inverted().m11() and multiply this constant to your local coordinate bounding rectangle. For example:
qreal m = this->deviceTransform(m_view->viewportTransform()).inverted().m11();
return QRectF(m*(m_shapeX), m*(m_shapeY),
m*(m_shapeR), m*(m_shapeR));
here is a solution I devised of very moderate complexity :
1) Get the boundingRect() of the parent and map it to scene
2) take the minimum X and Y of this list of points, this is the real origin of your item, in scene coordinates
3) set the position of the child
In Pyside :
br = parent.mapToScene(parent.boundingRect())
realX = min([item.x() for item in br])
realY = min([item.y() for item in br])
child.setPos(parent.mapFromScene(realX, realY)) #modify according to need
As seen in the pictures.
I have QWidget inside a QScrollArea.
QWidget act as a render widget for cell image and some vector based contour data.
User can performe zoom in/out and what simply happens is, it changes the QPainters scale and change the size of QWidget size accordinly.
Now I want to perform the zooming in/out on the point under the mouse. (like zooming action in GIMP).
How to calculate the new positions of the scrollbars according to the zoom level ?
Is it better to implement this using transformations without using a scrollarea?
One solution could be to derive a new class from QScrollArea and reimplementing wheelEvent for example so that zooming is performed with the mouse wheel and at the current mouse cursor position.
This method works by adjusting scroll bar positions accordingly to reflect the new zoom level. This means as long as there is no visible scroll bar, zooming does not take place under mouse cursor position. This is the behavior of most image viewer applications.
void wheelEvent(QWheelEvent* e) {
double OldScale = ... // Get old scale factor
double NewScale = ... // Set new scale, use QWheelEvent...
QPointF ScrollbarPos = QPointF(horizontalScrollBar()->value(), verticalScrollBar()->value());
QPointF DeltaToPos = e->posF() / OldScale - widget()->pos() / OldScale;
QPointF Delta = DeltaToPos * NewScale - DeltaToPos * OldScale;
widget()->resize(/* Resize according to new scale factor */);
horizontalScrollBar()->setValue(ScrollbarPos.x() + Delta.x());
verticalScrollBar()->setValue(ScrollbarPos.y() + Delta.y());
}
Will void QScrollArea::ensureVisible(int x, int y, int xmargin = 50, int ymargin = 50) do what you need?
You need to pick up the wheelEvent() on the QWidget, get the event.pos() and pass it into the QscrollArea.ensureVisible(), right after scaling your QWidget.
def wheelEvent(self, event):
self.setFixedSize(newWidth, newHeight)
self.parent().ensureVisible(event.pos())
That should more or less produce what you want.