Description
Bug summary
This may not be a bug. It is my use case.
I have embedded matplotlib in PyQt5 and I am using Ctrl+C to copy the figure to clipboard (for pasting to ppt)
So in the minimum example attached, pressing Ctrl+C reduces figure size to specified size (6x4). This is kind of a print preview for me. So I adjust legend/annotations in the smaller figure.
Dragging legend is not working properly in the smaller figure. It used to work in older versions for me. So I went through the changes. Reverting the change done in the following line of blit function of /backends/backend_qt.py fixes the problem (Changed as part of #17478)
self.repaint(l, self.rect().height() - t, w, h)
to
self.repaint(l, int(self.renderer.height / self.device_pixel_ratio)- t, w, h)
For the normal figure self.rect().height() is equal to int(self.renderer.height / self.device_pixel_ratio)
But for the smaller figure self.rect().height() is different from int(self.renderer.height / self.device_pixel_ratio)
Is this ok?
Thanks
Code for reproduction
import io
import sys
from PyQt5.QtWidgets import QWidget, QApplication, QShortcut, QVBoxLayout
from PyQt5.QtCore import QMimeData
from PyQt5.QtGui import QKeySequence
from matplotlib.backends.backend_qt5agg import FigureCanvasQTAgg as FigureCanvas
from matplotlib.figure import Figure
import random
class Window(QWidget):
def __init__(self, parent=None):
super(Window, self).__init__(parent)
self.figure = Figure()
data = [random.random() for _ in range(10)]
ax = self.figure.add_subplot(111)
ax.plot(data, '*-',label='RandomData')
legend = ax.legend()
self.canvas = FigureCanvas(self.figure)
self.canvas.draw()
legend.set_draggable(True,use_blit=True) # Dragging is not ok in smaller figure/ ok in normal figure
# legend.set_draggable(True) # Dragging is ok in smaller figure when blitting is off
ctc = QShortcut(QKeySequence("Ctrl+C"), self)
ctc.activated.connect(self.copy_to_clipboard)
layout = QVBoxLayout()
layout.addWidget(self.canvas)
self.setLayout(layout)
def copy_to_clipboard(self):
fig = self.canvas.figure
fig.set_size_inches(6,4)
buf = io.BytesIO()
fig.savefig(buf,dpi=300,transparent=True,bbox_inches='tight')
mimeData = QMimeData()
mimeData.setData("PNG",buf.getvalue())
QApplication.clipboard().setMimeData(mimeData)
buf.close()
if __name__ == '__main__':
app = QApplication(sys.argv)
main = Window()
main.showMaximized()
sys.exit(app.exec_())Actual outcome
Expected outcome
Additional information
No response
Operating system
Windows
Matplotlib Version
3.6.2
Matplotlib Backend
PyQt5
Python version
3.8.10
Jupyter version
NA
Installation
pip