I am using QT and try to plot a graph with QGraphicsView and QGraphicsScene..i dont want any additional dependencies, thats why i dont use QWT. When i plot my data, at the moment i use
scene->drawLine(x1,y1,x2,y2,pen);
then this draws a line between the 2 points in the QGraphicScene. But this uses a top left x=0 y=0 system... i would like to use my own system like in the picture below. Another problem is that i have double values from -3 to +3..has anyone some experience with QGraphicScene and QGraphicsView and can tell me how i can accomplish this?
Try looking into QGraphicsView::setTransform. This matrix defines how "scene coordinates" are translated to "view coordinates", it's a common concept in graphics programming.
There are also convenience functions scale(), rotate(), translate() and shear() which modify the view's current transformation matrix.
So you could do something like this:
scale(1.0, -1.0); // invert Y axis
// translate to account for fact that Y coordinates start at -3
translate(0.0, 3.0);
// scale Y coordinates to height of widget
scale(1.0, (qreal)viewport()->size().height()/6.0);
And since this is dependent on the size of the widget, you'd also want to catch any resize events and reset the transformation matrix again there. Assuming you want "3" to represent the top of the viewport and "-3" the bottom.
Related
I'm using the Qt C++ ArcGIS Runtime SDK v100.9 and I have various shapes and labels being drawn on a map.
I want to be able to find out the area (bounding rectangle) of Graphic (which is a text label (TextSymbol) at a given point (SpatialReference::wgs84) on the map) so I can determine if the width of the label is more or less than another Graphic (lets say it has a Polygon for its Geometry which is being used to draw a circle) in order to decide if the label should be set to visible or not.
Within a class derived from Esri::ArcGISRuntime::MapGraphicsView the circle and the text label are created along the lines of:
Point centerWgs84(0.0, 0.0, SpatialReference::wgs84());
Graphic* circleGraphic_p = new Graphic(GeometryEngine::bufferGeodetic(centerWgs84, 1000.0, LinearUnit::meters(), 0.5, GeodeticCurveType::Geodesic));
this->graphicsOverlays()->at(0)->graphics()->append(circleGraphic_p);
circleGraphic_p->setSymbol(new SimpleLineSymbol(SimpleLineSymbolStyle::Solid, QColor(Qt::blue), 1.0));
circleGraphic_p->setVisible(true);
TextSymbol* textMarker_p = new TextSymbol("Some Label", Qt::black, 12.0, HorizontalAlignment::Center, VerticalAlignment::Bottom);
Graphic* labelGraphic_p = new Graphic(centerWgs84, textMarker_p);
this->graphicsOverlays()->at(0)->graphics()->append(labelGraphic_p);
labelGraphic_p->setVisible(true);
Rather than always setting the label visibility to true, I thought I would be able to take the Geometry of each Graphic and use it to construct an Envelope which would allow me to then get the width of each envelope that could then be compared:
Envelope circleEnvelope(circleGraphic_p->geometry());
Envelope labelEnvelope(labelGraphic_p->geometry());
labelGraphic_p->setVisible(circleEnvelope.width() >= labelEnvelope.width());
but when I try and do this, the width of each envelope is always a very small negative value (such as -2.25017... e-98)
Any suggestions on what I am doing wrong or if there is a better way to get the size on the map (or in device independent units) of the text label and a Graphic described by the Geometry of a Polyline or Polygon?
EDIT: I've discovered that the Geometry object has an extent() method from which I can get the width of the circle but the Geometry of the Graphic being used for the text label results in a width of zero from its extent() method. I expect this is because the Geometry is just a Point which has no width or height. So the question still stands of how to get the bounding rectangle of a TextSymbol?
You are correct, the extent is being returned of the underlying geometry, not the TextSymbol. I don't think you will be able to achieve what you are wanting in the way you are going about it, as there isn't a way to get the bounding box or screen coordinates of the symbol itself. Instead, have you considered using LabelDefinitions and setting the various deconfliction options? This sample shows labels on layers, but can be applied to graphics as well. There are many labeling options you can apply, and this would allow the internal labeling engine to deconflict for you.
I have a window with a width of 260 pixels. By using the DrawSurface function I can put an image on the position which is not visible on the screen, for example (500, 10). Now I want to move the screen (by pressing the button) to the point where is the image. Is it possible?
I'm not sure how accurate or up-to-date this article is but it gives a lot of starting code for implementing a makeshift camera using an SDL_Rect variable. In your case, you would modify the x and y variables of the camera object and use the apply_surface() method to show textures relative to the camera's position.
I want to draw a triangle and text using C++ and Cairo like this:
|\
| \
|PP\
|___\
If I add the triangle and the text using Cairo I get:
___
| /
|PP/
| /
|/
So the y-axis is from top to bottom, but I want it from bottom to top. So I tried to changed the viewpoint matrix (cairo_transform(p, &mat);) or scale the data (cairo_scale(p, 1.0, -1.0);). I get:
|\
| \
|bb\
|___\
Now the triangle is the way I want it BUT the TEXT is MIRRORED, which I do not want to be mirrored.
Any idea how to handle this problem?
I was in a similar situation as the OP that required me to change a variety of coordinates in the cartesian coordinate system with the origin at the bottom left. (I had to port an old video game that was developed with a coordinate system different from Cairo's, and because of time constraints/possible calculation mistakes/port precision I decided it was better to not rewrite the whole bunch) Luckily, I found an okay approach to change Cairo's coordinate system. The approach is based around Cairo's internal transformation matrix, that transforms Cairo's input to the user device. The solution was to change this matrix to a reflection matrix, a matrix that mirrors it's input through the x-axis, like so:
cairo_t *cr;
cairo_matrix_t x_reflection_matrix;
cairo_matrix_init_identity(&x_reflection_matrix); // could not find a oneliner
/* reflection through the x axis equals the identity matrix with the bottom
left value negated */
x_reflection_matrix.yy = -1.0;
cairo_set_matrix(cr, &x_reflection_matrix);
// This would result in your drawing being done on top of the destination
// surface, so we translate the surface down the full height
cairo_translate(cr, 0, SURFACE_HEIGHT); // replace SURFACE_HEIGHT
// ... do your drawing
There is one catch however: text will also get mirrored. To solve this, one could alter the font transformation matrix. The required code for this would be:
cairo_matrix_t font_reflection_matrix;
// We first set the size, and then change it to a reflection matrix
cairo_set_font_size(cr, YOUR_SIZE);
cairo_get_font_matrix(cr, &font_reflection_matrix);
// reverse mirror the font drawing matrix
font_reflection_matrix.yy = font_reflection_matrix.yy * -1;
cairo_set_font_matrix(cr, &font_reflection_matrix);
Answer:
Rethink your coordinates and pass them correctly to cairo. If your coordinates source has an inverted axis, preprocess them to flip the geometry. That would be called glue code, and it is ofter neccessary.
Stuff:
It is a very common thing with 2D computer graphics to have the origin (0,0) in the top left corner and the y-axis heading downwards (see gimp/photoshop, positioning in html, webgl canvas). As allways there are other examples too (PDFs).
I'm not sure what the reason is, but I would assume the reading direction on paper (from top to bottom) and/or the process of rendering/drawing an image on a screen.
To me, it seems to be the easiest way to procedurally draw an image at some position from the first to the last pixel (you don't need to precalculate it's size).
I don't think that you are alone with your oppinion. But I don't think that there is a standard math coordinate system. Even the very common carthesian coordinate system is incomplete when the arrows that indicate axis direction are missing.
Summary: From the discussion I assume that there is only one coordinate system used by Cairo: x-axis to the right, y-axis down. If one needs a standard math coordinate system (x-axis to the right, y-axis up) one has to preprocess the data.
1. Goal
My colleague and I have been trying to render rotated ellipsoids in Qt. The typical solution approach, as we understand it, consists of shifting the center of the ellipsoids to the origin of the coordinate system, doing the rotation there, and shifting back:
http://qt-project.org/doc/qt-4.8/qml-rotation.html
2. Sample Code
Based on the solution outlined in the link above, we came up with the following sample code:
// Constructs and destructors
RIEllipse(QRect rect, RIShape* parent, bool isFilled = false)
: RIShape(parent, isFilled), _rect(rect), _angle(30)
{}
// Main functionality
virtual Status draw(QPainter& painter)
{
const QPen& prevPen = painter.pen();
painter.setPen(getContColor());
const QBrush& prevBrush = painter.brush();
painter.setBrush(getFillBrush(Qt::SolidPattern));
// Get rectangle center
QPoint center = _rect.center();
// Center the ellipse at the origin (0,0)
painter.translate(-center.x(), -center.y());
// Rotate the ellipse around its center
painter.rotate(_angle);
// Move the rotated ellipse back to its initial location
painter.translate(center.x(), center.y());
// Draw the ellipse rotated around its center
painter.drawEllipse(_rect);
painter.setBrush(prevBrush);
painter.setPen(prevPen);
return IL_SUCCESS;
}
As you can see, we have hard coded the rotation angle to 30 degrees in this test sample.
3. Observations
The ellipses come out at wrong positions, oftentimes outside the canvas area.
4. Question
What is wrong about the sample code above?
Best regards,
Baldur
P.S. Thanks in advance for any constructive response?
P.P.S. Prior to posting this message, we searched around quite a bit on stackoverflow.com.
Qt image move/rotation seemed to reflect a solution approach similar to the link above.
In painter.translate(center.x(), center.y()); you shift your object by the amount of current coordinate which makes (2*center.x(), 2*center.y()) as a result. You may need:
painter.translate(- center.x(), - center.y());
The theory of moving an object back to its origin, rotating and then replacing the object's position is correct. However, the code you've presented is not translating and rotating the object at all, but translating and rotating the painter. In the example question that you've referred to, they're wanting to rotate the whole image about an object, which is why they move the painter to the object's centre before rotating.
The easiest way to do rotations about a GraphicsItem is to initially define the item with its centre in the centre of the object, rather than in its top left corner. That way, any rotation will automatically be about the objects centre, without any need to translate the object.
To do this, you'd define the item with a bounding rect for x,y,width,height with (-width/2, -height/2, width, height).
Alternatively, assuming your item is inherited from QGraphicsItem or QGraphicsObject, you can use the function setTransformOriginPoint before any rotation.
I have a scene which contains objects located anywhere in space and I'm making a trackball-like interface.
I'd like to make it so that I can move 2 separate sliders to rotate it in x and y axes respectively:
glRotatef(drawRotateY,0.0,1.0f,0);
glRotatef(drawRotateX,1.0f,0.0,0.0);
//draw stuff in space
However, the above code won't work because X rotation will then be dependent on the Y rotation.
How can I achieve this without using gluLookAt()?
Edit:
I'd like to say that my implementation is even simpler than a trackball interface. Basically, if the x slider value is 80 and y slider is 60, rotate vertically 80 degrees and horizontally 60 degrees. I just need to make them independent of each other!
This code should get you started: http://www.cse.ohio-state.edu/~crawfis/Graphics/VirtualTrackball.html
It explains how to implement a virtual trackball using the current mouse position in the GL window.
You could probably use something like this:
Vector v = new Vector(drawRotateX, drawRotateY, 0);
float length = v.length();
v.normalize();
glRotatef(length, v.x, v.y, v.z);
When you say rotate vertically and horizontally, do you mean like an anti-aircraft gun - rotate around the vertical Z axis, to face in a particular compass heading (yaw) and then rotate to a particular elevation (pitch)?
If this is the case, then you just need to do your two rotations in the right order, and all will be well. In this example, you must do the 'pitch' rotation first, and then the 'yaw' rotation. All will work out fine.
If you mean something more complicated (eg. like the 'Cobra' spaceship in Elite) then you will need a more fiddly solution.