Memory_profiler While Using Lambda Expression To Connect Slots

I am experimenting memory_profiler in Python3 following https://medium.com/zendesk-engineering/hunting-for-memory-leaks-in-python-applications-6824d0518774 thanks to @eyllanesc he

Solution 1:

The problem pointed out in this post does not have to do with PyQt5 but with the lambda functions since, like any function, it creates a scope and stores memory, to test what the OP points out I have created the following example:

from PyQt5 import QtCore, QtWidgets


class MainWindow(QtWidgets.QMainWindow):
    def __init__(self, parent=None):
        super().__init__(parent)

        self.counter = 0
        self.timer = QtCore.QTimer()
        self.timer.timeout.connect(self.on_timeout)
        self.timer.start(1000)

        self.button = QtWidgets.QPushButton("Press me")
        self.setCentralWidget(self.button)

    @QtCore.pyqtSlot()
    def on_timeout(self):
        self.counter += 1
        if self.counter % 2 == 0:
            self.connect()
        else:
            self.disconnect()

    def connect(self):
        self.button.setText("Connected")
        self.button.clicked.connect(lambda checked: None)
        # or
        # self.button.clicked.connect(lambda checked, v=list(range(100000)): None)

    def disconnect(self):
        self.button.setText("Disconnected")
        self.button.disconnect()


def main():
    import sys

    app = QtWidgets.QApplication(sys.argv)
    w = MainWindow()
    w.show()
    sys.exit(app.exec_())


if __name__ == "__main__":
    main()

• self.button.clicked.connect(lambda checked: None)

• self.button.clicked.connect(lambda checked, v=list(range(100000)): None)

As observed at the time of connection and disconnection, the memory consumed increases since the lambda maintains the information of v=list(range(100000)).

---

But in your code the closure stores only the variable "j" which is minimum:

getattr(self, i).clicked.connect(lambda pippo, j=i: self.printo(j))

To see how this variable affects I am going to eliminate the unnecessary code for the test (hangman005.py, etc) in addition to offering alternatives:

• without connect:

class MainMenu(QtWidgets.QMainWindow):
    def __init__(self):
        super(MainMenu, self).__init__()
        uic.loadUi("main_window2.ui", self)
        self.hangman()

    def hangman(self):
        letters = "ABCDEFGHIJKLMNOPQRSTYVWXXYZ"
        for i in letters:
            pass

    def printo(self, i):
        print("oooooooooooooooooooooo :", i)

• with lambda:

class MainMenu(QtWidgets.QMainWindow):
    def __init__(self):
        super(MainMenu, self).__init__()
        uic.loadUi("main_window2.ui", self)
        self.hangman()

    def hangman(self):
        letters = "ABCDEFGHIJKLMNOPQRSTYVWXXYZ"
        for i in letters:
            getattr(self, i).clicked.connect(lambda pippo, j=i: self.printo(j))

    def printo(self, i):
        print("oooooooooooooooooooooo :", i)

• with functools.partial

class MainMenu(QtWidgets.QMainWindow):
    def __init__(self):
        super(MainMenu, self).__init__()
        uic.loadUi("main_window2.ui", self)
        self.hangman()

    def hangman(self):
        letters = "ABCDEFGHIJKLMNOPQRSTYVWXXYZ"
        for i in letters:
            getattr(self, i).clicked.connect(partial(self.printo, i))

    def printo(self, i):
        print("oooooooooooooooooooooo :", i)

• once connect

class MainMenu(QtWidgets.QMainWindow):
    def __init__(self):
        super(MainMenu, self).__init__()
        uic.loadUi("main_window2.ui", self)
        self.hangman()

    def hangman(self):
        letters = "ABCDEFGHIJKLMNOPQRSTYVWXXYZ"
        for i in letters:
            getattr(self, i).setProperty("i", i)
            getattr(self, i).clicked.connect(self.printo)

    def printo(self):
        i = self.sender().property("i")
        print("oooooooooooooooooooooo :", i)

As we can see there is no significant difference between all the methods so in your case there is no memory leak or rather it is very small.

In conclusion: Every time you create a lambda method it has a closure (j = i in your case) so the OP recommends taking that into account, for example in your case len(letters) * size_of(i) are consumed which at Being small letters and i makes it negligible, but if you otherwise store a heavier object unnecessarily then it will cause problems