Develop Reference

How to dynamically draw a recangle on the screen without glitching - python

https://media.giphy.com/media/3og0IDneIpXgdMDOzm/source.gif
I am working on a code that draws a transparent rectangle on screen and saves the section as an image, yet excessive rectangles are being left behind.
import wx
import win32gui
app = wx.App(False)
s = wx.ScreenDC()
s.Pen = wx.Pen("#FF0000")
int_x = win32gui.GetCursorPos()[0]
int_y = win32gui.GetCursorPos()[1]
while 1:
s.DrawRectangle(int_x,int_y,win32gui.GetCursorPos()[0]-int_x,win32gui.GetCursorPos()[1]-int_y)

I think you would need to do some kind of clear or refresh after the draw
s.Clear()
s.Refresh()
I haven't used this so I'm not sure if it will just work in your while loop but I'd imagine it won't and you would need something to handle the start (mousedown) and end (mouseup) of the drawing and then on mouseup do the clear and refresh that i mention.

Turtle graphics window's canvas color *not* showing up in Postscript (.ps) file [duplicate]

I am new to Python and have been working with the turtle module as a way of learning the language.
Thanks to stackoverflow, I researched and learned how to copy the image into an encapsulated postscript file and it works great. There is one problem, however. The turtle module allows background color which shows on the screen but does not show in the .eps file. All other colors, i.e. pen color and turtle color, make it through but not the background color.
As a matter of interest, I do not believe the import of Tkinter is necessary since I do not believe I am using any of the Tkinter module here. I included it as a part of trying to diagnose the problem. I had also used bgcolor=Orange rather than the s.bgcolor="orange".
No Joy.
I am including a simple code example:
# Python 2.7.3 on a Mac
import turtle
from Tkinter import *
s=turtle.Screen()
s.bgcolor("orange")
bob = turtle.Turtle()
bob.circle(250)
ts=bob.getscreen()
ts.getcanvas().postscript(file = "turtle.eps")
I tried to post the images of the screen and the .eps file but stackoverflow will not allow me to do so as a new user. Some sort of spam prevention. Simple enough to visualize though, screen has background color of orange and the eps file is white.
I would appreciate any ideas.

Postscript was designed for making marks on some medium like paper or film, not raster graphics. As such it doesn't have a background color per se that can be set to given color because that would normally be the color of the paper or unexposed film being used.
In order to simulate this you need to draw a rectangle the size of the canvas and fill it with the color you want as the background. I didn't see anything in the turtle module to query the canvas object returned by getcanvas() and the only alternative I can think of is to read the turtle.cfg file if there is one, or just hardcode the default 300x400 size. You might be able to look at the source and figure out where the dimensions of the current canvas are stored and access them directly.
Update:
I was just playing around in the Python console with the turtle module and discovered that what the canvas getcanvas() returns has a private attribute called _canvas which is a <Tkinter.Canvas instance>. This object has winfo_width() and winfo_height() methods which seem to contain the dimensions of the current turtle graphics window. So I would try drawing a filled rectangle of that size and see if that gives you what you want.
Update 2:
Here's code showing how to do what I suggested. Note: The background must be drawn before any other graphics are because otherwise the solid filled background rectangle created will cover up everything else on the screen.
Also, the added draw_background() function makes an effort to save and later restore the graphics state to what it was. This may not be necessary depending on your exact usage case.
import turtle
def draw_background(a_turtle):
""" Draw a background rectangle. """
ts = a_turtle.getscreen()
canvas = ts.getcanvas()
height = ts.getcanvas()._canvas.winfo_height()
width = ts.getcanvas()._canvas.winfo_width()
turtleheading = a_turtle.heading()
turtlespeed = a_turtle.speed()
penposn = a_turtle.position()
penstate = a_turtle.pen()
a_turtle.penup()
a_turtle.speed(0) # fastest
a_turtle.goto(-width/2-2, -height/2+3)
a_turtle.fillcolor(turtle.Screen().bgcolor())
a_turtle.begin_fill()
a_turtle.setheading(0)
a_turtle.forward(width)
a_turtle.setheading(90)
a_turtle.forward(height)
a_turtle.setheading(180)
a_turtle.forward(width)
a_turtle.setheading(270)
a_turtle.forward(height)
a_turtle.end_fill()
a_turtle.penup()
a_turtle.setposition(*penposn)
a_turtle.pen(penstate)
a_turtle.setheading(turtleheading)
a_turtle.speed(turtlespeed)
s = turtle.Screen()
s.bgcolor("orange")
bob = turtle.Turtle()
draw_background(bob)
ts = bob.getscreen()
canvas = ts.getcanvas()
bob.circle(250)
canvas.postscript(file="turtle.eps")
s.exitonclick() # optional
And here's the actual output produced (rendered onscreen via Photoshop):

I haven't found a way to get the canvas background colour on the generated (Encapsulated) PostScript file (I suspect it isn't possible). You can however fill your circle with a colour, and then use Canvas.postscript(colormode='color') as suggested by #mgilson:
import turtle
bob = turtle.Turtle()
bob.fillcolor('orange')
bob.begin_fill()
bob.circle(250)
bob.begin_fill()
ts = bob.getscreen()
ts.getcanvas().postscript(file='turtle.eps', colormode='color')

Improving #martineau's code after a decade
import turtle as t
Screen=t.Screen()
Canvas=Screen.getcanvas()
Width, Height = Canvas.winfo_width(), Canvas.winfo_height()
HalfWidth, HalfHeight = Width//2, Height//2
Background = t.Turtle()
Background.ht()
Background.speed(0)
def BackgroundColour(Colour:str="white"):
Background.clear() # Prevents accumulation of layers
Background.penup()
Background.goto(-HalfWidth,-HalfHeight)
Background.color(Colour)
Background.begin_fill()
Background.goto(HalfWidth,-HalfHeight)
Background.goto(HalfWidth,HalfHeight)
Background.goto(-HalfWidth,HalfHeight)
Background.goto(-HalfWidth,-HalfHeight)
Background.end_fill()
Background.penup()
Background.home()
BackgroundColour("orange")
Bob=t.Turtle()
Bob.circle(250)
Canvas.postscript(file="turtle.eps")
This depends on what a person is trying to accomplish but generally, having the option to select which turtle to use to draw your background to me is unnecessary and can overcomplicate things so what one can do instead is have one specific turtle (which I named Background) to just update the background when desired.
Plus, rather than directing the turtle object via magnitude and direction with setheading() and forward(), its cleaner (and maybe faster) to simply give the direct coordinates of where the turtle should go.
Also for any newcomers: Keeping all of the constants like Canvas, Width, and Height outside the BackgroundColour() function speeds up your code since your computer doesn't have to recalculate or refetch any values every time the function is called.

Can you display a Splash Screen post launch?

I'm trying to use QSplashScreen as a 'simple' notification window. I build the actual notification in a dialog and am then trying to use QPixmap.grabWidget() on that and pass the pixmap to the SplashScreen but its not showing. So I'm wondering if the problem is due to me trying to use it post-start.
Simplified version of what I'm doing.
class NotifyHandler:
def showNotify(self, event):
widget = NotifyWidget.createInstance(event) # Build the QDialog
#Do I need to do widget.show()?
pixmap = QtGui.QPixmap.grabWidget(widget) # Convert to pixmap
splash = QtGui.QSplashScreen(pixmap, QtCore.Qt.WindowStayOnTopHint)
(x,y) = getDisplayLocation(splash) # Get coords to put it in bottom-right corner
splash.setGeometry(splash.width(), splash.height(), x, y)
splash.show()
QtCore.QTimer.singleShot(3000, splash.close)
I try this however nothing shows up. If I try it as a Dialog though:
widget.show()
widget.raise_()
widget.activateWindow()
It works fine. So I figure either I can't do this post-launch or there is something going on when I try to feed it a pixmap of a widget. Any ideas?

It seems my dialog issue and splash screen were having the same sort of issue.
I forgot that if you don't assign the dialog (or Splash screen in this case) to a scope outside the method then the GC will just eat the object after the method returns.
So by simply adding something like:
self.splash = splash
self.splash.show()
QtCore.QTimer.singleShot(3000, self.splash.close)
The splash screen wouldn't be garbage collected and I get my splash screen.

What is the correct sequence for bliting surfaces to the screen in pygame?

I am creating a simple mp3 player and my first task was to create a simple button that a user could press. I created a class called Button which handled this behavior and detects if a user has clicked it and then changes color. I am now trying to have a default text that the button displays and another string (pres_string) which will be displayed if the button is being pressed.
The only problem is my background surface seems to be in the wrong place and is drawing over any changes I have made.
Here is my code:
http://pastebin.com/Nh3yy01X
As you can see I've commented out the lines I described and tried it with basic variables in the main function just to test what was going wrong.
Thanks for any help.
(Feel free to change the title of the question, I wasn't sure what most accuratelydescribed my problem)

Clear the surface every loop
def draw(self):
# clear screen."
self.screen.fill( self.color_bg )
# first draw background
# Then buttons
# then any extra top level text
# update
pygame.display.flip()
tip: For colors, you can call pygame.Color() with human-names like red ( gray20 and gray80 have a nice contrast, to use for bg and text. )
from pygame import Color
text = Color('gray20')
Your button, psuedocode. Fix: moved color as an instance member.
class Button(object):
def __init__(self, text, rect=None):
self.color_bg = Color("gray20")
self.color_text = color("gray80")
if rect is None: rect = Rect(0,0,1,1)
self.rect = rect
self._render()
def _render(self):
# draw button and .render() font, cache to surface for later.
self.surface_cached = self.surface.copy()
# render text
#if size changes, save rect.size of cached surface , mantaining location
self.rect.size = cache.get_rect.size
def draw(self):
# draw cached surface
screen.blit( self.surface_cached, self.rect)
For testClick use Rect.collidepoint http://www.pygame.org/docs/ref/rect.html#Rect.collidepoint

2D bitmap-based computer graphics are like drawing or painting - you put the new ink on top of whatever was there already. So your background must be the first thing you draw each time.

wxPython: Exit Fullscreen

To display a wxPython window in full screen mode you use:
ShowFullScreen(True)
How do you get out of full screen though? I've tried the obvious way:
ShowFullScreen(True)
sleep(5)
ShowFullScreen(False)
This doesn't work though. When I run the script, nothing appears. After 5 seconds a window roughly 200x250 appears in the top-left corner of the screen, without anything inside of it. It doesn't appear to have any borders either.
If I change this to
showFullScreen(True)
then I get stuck with a full screen window that I have to use Alt + F2 -> xkill to get out of.

It looks like you need to Show() the window first. (According to the documentation, you shouldn't have to. Maybe this is a bug.) I tested on Mac OS X and Windows - they both exhibit issues if you don't call Show() first.
Also note that you shouldn't sleep in the main GUI thread. You'll hang the UI. Using CallLater is one potential solution, as shown in my example.
Working example:
import wx
def main():
app = wx.PySimpleApp()
frame = wx.Frame(None, -1, 'Full Screen Test')
frame.Show()
frame.ShowFullScreen(True)
wx.CallLater(5000, frame.ShowFullScreen, False)
app.MainLoop()
if __name__ == '__main__':
main()

The documentation for ShowFullScreen reads:
ShowFullScreen(show, style=wx.FULLSCREEN_ALL)
Depending on the value of show parameter the window is either shown full screen or restored to its normal state.
Parameters:
show (bool)
style (long): is a bit list containing some or all of the following values, which indicate what elements of the window to hide in full-screen mode:
wx.FULLSCREEN_NOMENUBAR
wx.FULLSCREEN_NOTOOLBAR
wx.FULLSCREEN_NOSTATUSBAR
wx.FULLSCREEN_NOBORDER
wx.FULLSCREEN_NOCAPTION
wx.FULLSCREEN_ALL (all of the above)
So put your Full Screen toggle event/s in a Menu and start full screen mode with:
self.window.ShowFullScreen(True, style=(wx.FULLSCREEN_NOTOOLBAR | wx.FULLSCREEN_NOSTATUSBAR |wx.FULLSCREEN_NOBORDER |wx.FULLSCREEN_NOCAPTION))
Note that I omitted wx.FULLSCREEN_NOMENUBAR, in this way you will still be able to access the menu to turn full screen mode off again.