Merge two pieces of code

john_hobbyist

What modifications should I do in the command?

john_hobbyist

I am trying to modify/incorporate from the correct running code from the stackoverflow that @JonB suggested. With this:

screen_pixmap = QScreen.grabWindow(app.primaryScreen(),QApplication.desktop().winId()).save(filename, 'png')

I get this:

Must construct a QApplication first.
Traceback (most recent call last):
  File "code_2.py", line 340, in <module>
    screen_pixmap = QScreen.grabWindow(app.primaryScreen(),QApplication.desktop().winId()).save(filename, 'png')
AttributeError: 'NoneType' object has no attribute 'winId'
Segmentation fault (core dumped)

And with this

screen_pixmap = QScreen.grabWindow(app.primaryScreen(),QApplication.desktop()).save(filename, 'png')

I get again this:

 construct a QApplication first.
Traceback (most recent call last):
  File "code_2.py", line 341, in <module>
    screen_pixmap = QScreen.grabWindow(app.primaryScreen(),QApplication.desktop()).save(filename, 'png')
TypeError: grabWindow(self, PyQt5.sip.voidptr, x: int = 0, y: int = 0, width: int = -1, height: int = -1): first argument of unbound method must have type 'QScreen'
Segmentation fault (core dumped)

in which I cannot find a reliable solution from google/stackoverflow...

SGaist

@john_hobbyist said in Merge two pieces of code:

grabWindow

This is not a class method, you need an instance of QScreen.

In what way is that related to your original QRubberBand question ?

john_hobbyist

@SGaist Yes, I try to use stackoverflow code that has an approach for what I am trying so many days to do. Please look previous posts. And these are errors...

SGaist

I already answered to the errors you are getting.

@SGaist said in Merge two pieces of code:

This is not a class method, you need an instance of QScreen.

You really should read the documentation of the classes you are using rather than just blindingly copy paste.

john_hobbyist

I have been making many changes all day. The rest of the application runs, but nowhere I can see a rubberband rectangle in order to select/printscreen an area. I get these warnings though...I am searching it, but any idea would help...

QWidget::paintEngine: Should no longer be called
QPainter::begin: Paint device returned engine == 0, type: 1
QPainter::end: Painter not active, aborted

SGaist

Where exactly are you doing custom painting ?

john_hobbyist

I don't know! This is the modified last part of my code:

if __name__ == "__main__":
    app = QApplication(sys.argv)
    icon = QIcon(os.path.join(selfDir, '_static', 'orthoview.ico'))

    app.setWindowIcon(icon)
    from PyQt5.QtWidgets import *
    data = take_screenshot()
    window = QWidget()
    layout = QtWidgets.QVBoxLayout(window) 

    pixmap = QPixmap()
    pixmap.loadFromData(data)
    label = QtWidgets.QLabel()
    label.setPixmap(pixmap)
    
    layout.addWidget(label)         
    geometry = app.desktop().availableGeometry()        
    window.setFixedSize(geometry.width(), geometry.height())        

    window = OrthoView()
    window.show()
    sys.exit(app.exec_())

I used/modified this:

import sys

from PySide2 import QtCore, QtGui, QtWidgets


def take_screenshot():
    from PyQt5 import QtCore as pyqt5c
    from PyQt5 import QtWidgets as pyqt5w

    screen = pyqt5w.QApplication.primaryScreen()
    winid = pyqt5w.QApplication.desktop().winId()
    pixmap = screen.grabWindow(winid)

    ba = pyqt5c.QByteArray()
    buff = pyqt5c.QBuffer(ba)
    pixmap.save(buff, "PNG")
    return ba.data()


if __name__ == "__main__":
    app = QtWidgets.QApplication(sys.argv)

    data = take_screenshot()

    pixmap = QtGui.QPixmap()
    pixmap.loadFromData(data)

    label = QtWidgets.QLabel()
    label.setPixmap(pixmap)
    label.show()

    sys.exit(app.exec_())

(source: https://stackoverflow.com/questions/59118938/type-error-when-calling-qscreen-grabwindow)

JonB

@john_hobbyist
Maybe I misunderstand, but you call take_screenshot()as the very first thing you do, before you have created or shown anything at all? What would you expect from that, logically?

EDIT
OK, you're trying to screenshot what's already on the screen? Sorry, I misunderstood. Still seems very early but may be correct. Ignore this post if it is :)

I'm not sure I see anything here to do with custom painting. Do you use a debugger? Do you step through your code to find out where your problems come from? We need you to (learn to) do that, people here don't know where the errors come in your code.

john_hobbyist

@JonB Yes, I am trying to find the errors!

JonB

@john_hobbyist
So, nicely, do you use some Python debugger to execute your code?

john_hobbyist

@JonB I write the code on a simple text editor and run it on Ubuntu's command line..

JonB

@john_hobbyist
Yeah, that's not a great starting point :) That means you have to put print() statements everywhere if you want to know what's going on.

Unfortunately so far as I know Qt Creator still offers nothing fur running/debugging Python. (Has this changed?) I used PyCharm as editor and debugger for my Python coding, it was great, but not integrated with Qt Creator if you are designing UI files there. Else, I don't know but running under pdb might be better than nothing at all.

john_hobbyist

Yes, I work with "prints" inside code.... So here how things are: The program goes to the below methods (source on my previous posts) , but it does not displays the rectangle so it does not call screenshot...Any ideas?

Sorry! Updated the correct code!

    def mousePressEvent(self, event):
        print("56")
        '''
            Mouse is pressed. If selection is visible either set dragging mode (if close to border) or hide selection.
            If selection is not visible make it visible and start at this point.
        '''

        if event.button() == QtCore.Qt.LeftButton:
            print("57")

            position = QtCore.QPoint(event.pos())
            print("58")
            if self.selection.isVisible():
                print("59")
                # visible selection
                if (self.upper_left - position).manhattanLength() < 20:
                    print("60")
                    # close to upper left corner, drag it
                    self.mode = "drag_upper_left"
                    print("61")
                elif (self.lower_right - position).manhattanLength() < 20:
                    print("62")
                    # close to lower right corner, drag it
                    self.mode = "drag_lower_right"
                    print("63")
                else:
                    print("64")
                    # clicked somewhere else, hide selection
                    self.selection.hide()
                    print("65")
            else:
                # no visible selection, start new selection
                print("66")
                self.upper_left = position
                print("67")
                self.lower_right = position
                print("68")
                self.mode = "drag_lower_right"
                print("69")
                self.selection.show()
                print("70")

    def mouseMoveEvent(self, event):
        print("71")
        '''
            Mouse moved. If selection is visible, drag it according to drag mode.
        '''
        if self.selection.isVisible():
            print("72")
            # visible selection
            if self.mode == "drag_lower_right":
                print("73")
                self.lower_right = QtCore.QPoint(event.pos())
                print("74")
            elif self.mode == "drag_upper_left":
                print("75")
                self.upper_left = QtCore.QPoint(event.pos())
                print("76")
            # update geometry
            self.selection.setGeometry(QtCore.QRect(self.upper_left, self.lower_right).normalized())
            print("77")

def take_screenshot():
    print("205")
    from PyQt5 import QtCore as pyqt5c
    print("206")
    from PyQt5 import QtWidgets as pyqt5w
    print("207")
    screen = pyqt5w.QApplication.primaryScreen()
    print("208")
    winid = pyqt5w.QApplication.desktop().winId()
    print("209")
    pixmap = screen.grabWindow(winid)
    print("210")

    ba = pyqt5c.QByteArray()
    print("211")
    buff = pyqt5c.QBuffer(ba)
    print("212")
    pixmap.save(buff, "PNG")
    print("213")
    return ba.data()

if __name__ == "__main__":
    print("214")
    app = QApplication(sys.argv)
    print("215")
    icon = QIcon(os.path.join(selfDir, '_static', 'orthoview.ico'))
    print("216")
    app.setWindowIcon(icon)
    print("217")
    from PyQt5.QtWidgets import *
    print("218")
    data = take_screenshot()
    print("219")
    window = QWidget()
    print("220")
    layout = QtWidgets.QVBoxLayout(window)
    print("221")
    pixmap = QPixmap()
    print("222")
    pixmap.loadFromData(data)
    print("223")
    label = QtWidgets.QLabel()
    print("224")
    label.setPixmap(pixmap)
    print("225")
    
    layout.addWidget(label)     
    print("226")
    geometry = app.desktop().availableGeometry()        
    print("227")
    window.setFixedSize(geometry.width(), geometry.height())       
    print("228")
    window = OrthoView()
    print("229")
    window.show()
    print("230")
    sys.exit(app.exec_())

JonB

@john_hobbyist
Seems to be much the same code as you started with, re-pasted. I don't understand how you said you had moved from Qt4 to Qt5 with QtGui changes, yet now you have QtGui.QLabel, which as far as I know should not work....

Anyway, I don't know what path your code is or is not following. It's time for you to put your own print() statements in if you want to understand what is going on....

john_hobbyist

@JonB Yes, I have put "prints" inside the code, that is how I figure out on which parts of code the execution goes...

JonB

@john_hobbyist
But you don't tell me, so you want me to guess?

Does it hit the self.selection.show()? Does it hit the move code? What is your usage of Python is instead of == in if self.mode is "drag_lower_right"?

I leave it to you now.

john_hobbyist

@JonB Sorry! I updated the right code! Please see above 2-3 posts..
When I press the mouse button it goes: 56, 57,58, 59, 62, 63, 56, 57,58, 59, 62, 63....
When I move the mouse, it goes 71, 72, 73, 74, 77, 71, 72, 73, 74, 77...
Either I have loaded an image or not!

JonB

@john_hobbyist
I don't know.

 self.lower_right = QtCore.QPoint(event.pos())
self.upper_left = QtCore.QPoint(event.pos())
self.selection.setGeometry(QtCore.QRect(self.upper_left, self.lower_right).normalized())

Are the coordinates of event.pos() correct for setGeometry()? Start by getting some visible rubber band, e.g. put suitable constants in there instead of relying on mouse positions?

SGaist

@john_hobbyist did you try to just get the rubber band stuff working on a minimal widget ?

You can then apply that knowledge to your more complex application.