matplotlib FigureCanvasQTAgg：QWidget 上的慢速旋转和网格重叠表面
Posted
技术标签:
【中文标题】matplotlib FigureCanvasQTAgg：QWidget 上的慢速旋转和网格重叠表面【英文标题】：matplotlib FigureCanvasQTAgg: slow rotation and grid overlaping surface on QWidget 【发布时间】：2012-11-23 07:24:58 【问题描述】：我有 qtdesigner 生成的小部件 - optim_plotting_frame.py（下面的代码）
from PyQt4 import QtCore, QtGui
try:
_fromUtf8 = QtCore.QString.fromUtf8
except AttributeError:
_fromUtf8 = lambda s: s
class Ui_optim_plotting_frame(object):
def setupUi(self, optim_plotting_frame):
optim_plotting_frame.setObjectName(_fromUtf8("optim_plotting_frame"))
optim_plotting_frame.setWindowModality(QtCore.Qt.ApplicationModal)
optim_plotting_frame.resize(700, 580)
optim_plotting_frame.setWindowTitle(QtGui.QApplication.translate("optim_plotting_frame", "Plotting", None, QtGui.QApplication.UnicodeUTF8))
self.verticalLayout_2 = QtGui.QVBoxLayout(optim_plotting_frame)
self.verticalLayout_2.setObjectName(_fromUtf8("verticalLayout_2"))
self.horizontalLayout = QtGui.QHBoxLayout()
self.horizontalLayout.setObjectName(_fromUtf8("horizontalLayout"))
self.verticalLayout = QtGui.QVBoxLayout()
self.verticalLayout.setSizeConstraint(QtGui.QLayout.SetMinimumSize)
self.verticalLayout.setObjectName(_fromUtf8("verticalLayout"))
self.label_function = QtGui.QLabel(optim_plotting_frame)
self.label_function.setMinimumSize(QtCore.QSize(111, 16))
self.label_function.setMaximumSize(QtCore.QSize(16777215, 16777215))
self.label_function.setToolTip(QtGui.QApplication.translate("optim_plotting_frame", "defines radial basis function for interpolation", None, QtGui.QApplication.UnicodeUTF8))
self.label_function.setText(QtGui.QApplication.translate("optim_plotting_frame", "\n"
"\n"
"p, li white-space: pre-wrap; \n"
"\n"
"Interpolation function:", None, QtGui.QApplication.UnicodeUTF8))
self.label_function.setObjectName(_fromUtf8("label_function"))
self.verticalLayout.addWidget(self.label_function)
self.cmb_function = QtGui.QComboBox(optim_plotting_frame)
self.cmb_function.setMinimumSize(QtCore.QSize(111, 22))
self.cmb_function.setMaximumSize(QtCore.QSize(16777215, 16777215))
self.cmb_function.setToolTip(QtGui.QApplication.translate("optim_plotting_frame", "defines radial basis function for interpolation", None, QtGui.QApplication.UnicodeUTF8))
self.cmb_function.setObjectName(_fromUtf8("cmb_function"))
self.cmb_function.addItem(_fromUtf8(""))
self.cmb_function.setItemText(0, QtGui.QApplication.translate("optim_plotting_frame", "multiquadric", None, QtGui.QApplication.UnicodeUTF8))
self.cmb_function.addItem(_fromUtf8(""))
self.cmb_function.setItemText(1, QtGui.QApplication.translate("optim_plotting_frame", "inverse", None, QtGui.QApplication.UnicodeUTF8))
self.cmb_function.addItem(_fromUtf8(""))
self.cmb_function.setItemText(2, QtGui.QApplication.translate("optim_plotting_frame", "gaussian", None, QtGui.QApplication.UnicodeUTF8))
self.cmb_function.addItem(_fromUtf8(""))
self.cmb_function.setItemText(3, QtGui.QApplication.translate("optim_plotting_frame", "linear", None, QtGui.QApplication.UnicodeUTF8))
self.cmb_function.addItem(_fromUtf8(""))
self.cmb_function.setItemText(4, QtGui.QApplication.translate("optim_plotting_frame", "cubic", None, QtGui.QApplication.UnicodeUTF8))
self.cmb_function.addItem(_fromUtf8(""))
self.cmb_function.setItemText(5, QtGui.QApplication.translate("optim_plotting_frame", "quintic", None, QtGui.QApplication.UnicodeUTF8))
self.cmb_function.addItem(_fromUtf8(""))
self.cmb_function.setItemText(6, QtGui.QApplication.translate("optim_plotting_frame", "thin_plate", None, QtGui.QApplication.UnicodeUTF8))
self.verticalLayout.addWidget(self.cmb_function)
self.label_alpha = QtGui.QLabel(optim_plotting_frame)
self.label_alpha.setMinimumSize(QtCore.QSize(111, 16))
self.label_alpha.setMaximumSize(QtCore.QSize(16777215, 16777215))
self.label_alpha.setToolTip(QtGui.QApplication.translate("optim_plotting_frame", "Defines transparency: 0 - transparent, 1 - not transparent", None, QtGui.QApplication.UnicodeUTF8))
self.label_alpha.setText(QtGui.QApplication.translate("optim_plotting_frame", "Alpha:", None, QtGui.QApplication.UnicodeUTF8))
self.label_alpha.setObjectName(_fromUtf8("label_alpha"))
self.verticalLayout.addWidget(self.label_alpha)
self.dspb_alpha = QtGui.QDoubleSpinBox(optim_plotting_frame)
self.dspb_alpha.setMinimumSize(QtCore.QSize(111, 0))
self.dspb_alpha.setMaximumSize(QtCore.QSize(16777215, 16777215))
self.dspb_alpha.setToolTip(QtGui.QApplication.translate("optim_plotting_frame", "Defines transparency: 0 - transparent, 1 - not transparent", None, QtGui.QApplication.UnicodeUTF8))
self.dspb_alpha.setMaximum(1.0)
self.dspb_alpha.setSingleStep(0.1)
self.dspb_alpha.setProperty("value", 0.7)
self.dspb_alpha.setObjectName(_fromUtf8("dspb_alpha"))
self.verticalLayout.addWidget(self.dspb_alpha)
self.label_smooth = QtGui.QLabel(optim_plotting_frame)
self.label_smooth.setMinimumSize(QtCore.QSize(111, 16))
self.label_smooth.setMaximumSize(QtCore.QSize(16777215, 16777215))
self.label_smooth.setToolTip(QtGui.QApplication.translate("optim_plotting_frame", "Smoothness of the approximation", None, QtGui.QApplication.UnicodeUTF8))
self.label_smooth.setText(QtGui.QApplication.translate("optim_plotting_frame", "Smoothness:", None, QtGui.QApplication.UnicodeUTF8))
self.label_smooth.setObjectName(_fromUtf8("label_smooth"))
self.verticalLayout.addWidget(self.label_smooth)
self.dspb_smooth = QtGui.QDoubleSpinBox(optim_plotting_frame)
self.dspb_smooth.setMinimumSize(QtCore.QSize(111, 0))
self.dspb_smooth.setMaximumSize(QtCore.QSize(16777215, 16777215))
self.dspb_smooth.setToolTip(QtGui.QApplication.translate("optim_plotting_frame", "Smoothness of the approximation", None, QtGui.QApplication.UnicodeUTF8))
self.dspb_smooth.setSingleStep(0.1)
self.dspb_smooth.setObjectName(_fromUtf8("dspb_smooth"))
self.verticalLayout.addWidget(self.dspb_smooth)
self.chb_normxy = QtGui.QCheckBox(optim_plotting_frame)
self.chb_normxy.setText(QtGui.QApplication.translate("optim_plotting_frame", "normalized x,y ticks", None, QtGui.QApplication.UnicodeUTF8))
self.chb_normxy.setChecked(True)
self.chb_normxy.setObjectName(_fromUtf8("chb_normxy"))
self.verticalLayout.addWidget(self.chb_normxy)
spacerItem = QtGui.QSpacerItem(20, 40, QtGui.QSizePolicy.Minimum, QtGui.QSizePolicy.Expanding)
self.verticalLayout.addItem(spacerItem)
self.horizontalLayout.addLayout(self.verticalLayout)
self.widget = QtGui.QWidget(optim_plotting_frame)
self.widget.setFocusPolicy(QtCore.Qt.StrongFocus)
self.widget.setObjectName(_fromUtf8("widget"))
self.horizontalLayout.addWidget(self.widget)
self.horizontalLayout.setStretch(1, 10)
self.verticalLayout_2.addLayout(self.horizontalLayout)
self.retranslateUi(optim_plotting_frame)
QtCore.QMetaObject.connectSlotsByName(optim_plotting_frame)
def retranslateUi(self, optim_plotting_frame):
pass
和 main.py 中的类（下面的代码）继承生成的小部件并在其上绘制表面（将带有绘图表面的 FigureCanvasQTAgg 添加到小部件）
import sys
from PyQt4 import QtGui, QtCore
from PyQt4.QtGui import QApplication, QDialog
import numpy as np
from scipy.interpolate import Rbf
from optim_plotting_frame import Ui_optim_plotting_frame
from matplotlib.backends.backend_qt4agg import FigureCanvasQTAgg as FigureCanvas
from matplotlib.backends.backend_qt4agg import NavigationToolbar2QTAgg as NavigationToolbar
from mpl_toolkits.mplot3d import Axes3D
from matplotlib.figure import Figure
from matplotlib import cm
class optim_plotting_frame(QtGui.QWidget):
"Plotting settings widget"
def __init__(self, x, y, z, minx=None, maxx=None, miny=None, maxy=None, xname=None, yname=None, zname=None):
super(optim_plotting_frame, self).__init__()
self.ui = Ui_optim_plotting_frame()
self.ui.setupUi(self)
"""
Parameters:
x: numpy.ndarray
set of first parameter data (first coordinate)
y: numpy.ndarray
set of second parameter data (second coordinate)
z: numpy.ndarray
set of data corresponding to x and y (result)
alpha: float, optional
Number between 0 and 1 that defines transparency: 0 - transparent,
1 - not transparent.
function: string, optional
defines radial basis function for interpolation, by default - 'multiquadric'
'multiquadric': sqrt((r/self.epsilon)**2 + 1)
'inverse': 1.0/sqrt((r/self.epsilon)**2 + 1)
'gaussian': exp(-(r/self.epsilon)**2)
'linear': r
'cubic': r**3
'quintic': r**5
'thin_plate': r**2 * log(r)
smooth: float, optional
Values greater than zero increase the smoothness of the approximation.
0 is for interpolation (default), the function will always go through
the nodal points in this case.
minx: float, opyional
minimum possible value of x
maxx: float, opyional
maximum possible value of x
miny: float, opyional
minimum possible value of y
maxy: float, opyional
maximum possible value of y
xname: string, optional
name of x parameter
yname: string, optional
name of y parameter
zname: string, optional
name of z parameter (objective function)
"""
#==========================Data validation=============================
if any((not isinstance(x, np.ndarray),
not isinstance(y, np.ndarray),
not isinstance(z, np.ndarray))):
print "Error: x, y, z must be of numpy.ndarray type."
return None
if x.size != y.size != z.size:
print "Error: x, y, z must be of equal size."
return None
if not isinstance(minx, (float, int)):
minx = x.min()
if not isinstance(maxx, (float, int)):
maxx = x.max()
if not isinstance(miny, (float, int)):
miny = y.min()
if not isinstance(maxy, (float, int)):
maxy = y.max()
if minx > maxx:
tmp = minx
minx = maxx
maxx = tmp
if miny > maxy:
tmp = miny
miny = maxy
maxy = tmp
if not isinstance(xname, str):
xname = ""
if not isinstance(yname, str):
yname = ""
if not isinstance(zname, str):
zname = ""
#======================================================================
self.initialized = False
self.x = x
self.y = y
self.z = z
self.minx = minx
self.maxx = maxx
self.miny = miny
self.maxy = maxy
self.xname = xname
self.yname = yname
self.zname = zname
# map to [0,1] range
self.x = (x - minx) / maxx
self.y = (y - miny) / maxy
# update alpha, function, smooth values from widget
self.alpha = self.ui.dspb_alpha.value()
self.function = self.ui.cmb_function.currentText()
self.smooth = self.ui.dspb_smooth.value()
self.create_main_frame()
self.initialized = True
self.plot()
def updateVals(self):
"""
update alpha, function, smooth values from widget and parameters
"""
self.alpha = self.ui.dspb_alpha.value()
self.function = self.ui.cmb_function.currentText()
self.smooth = self.ui.dspb_smooth.value()
def plot(self):
self.updateVals()
self.axes.clear()
try:
# getting coordinate matrices from two coordinate vectors.
tx = np.linspace(self.x.min(), self.x.max(), 100)
ty = np.linspace(self.y.min(), self.y.max(), 100)
XI, YI = np.meshgrid(tx, ty)
# interpolating by radial basis function
rbf = Rbf(self.x, self.y, self.z, function=str(self.function), smooth=self.smooth)
# getting interpolation function results corresponding to (XI, YI)
ZI = rbf(XI, YI)
# plotting interpolaed surface
self.axes.plot_surface(XI, YI, ZI, cmap=cm.jet, alpha=self.alpha)
except Exception as e:
print "Error occured! original message: " + e.message
# plotting initial points
self.axes.scatter(self.x, self.y, self.z)
self.axes.set_xlim(self.x.min(), self.x.max())
self.axes.set_ylim(self.y.min(), self.y.max())
self.axes.set_title('RBF interpolation ' + self.function)
if not self.ui.chb_normxy.isChecked():
# setting ticks labels on the x line
self.axes.set_xticklabels((self.axes.get_xticks() * self.maxx + self.minx).round(1))
# setting ticks labels on the y line
self.axes.set_yticklabels((self.axes.get_yticks() * self.maxy + self.miny).round(1))
self.axes.set_xlabel(self.xname)
self.axes.set_ylabel(self.yname)
self.axes.set_zlabel(self.zname)
## adding colorbar
#m = cm.ScalarMappable(cmap=cm.jet)
#m.set_array(ZI)
#self.axes.figure.colorbar(m)
self.canvas.draw()
def create_main_frame(self):
# Create the mpl Figure and FigCanvas objects. 5x4 inches, 100 dots-per-inch
self.dpi = 100
self.fig = Figure((8.0, 4.0), dpi=self.dpi)
# setting diagram background
self.fig.patch.set_facecolor('white')
self.canvas = FigureCanvas(self.fig)
self.canvas.setParent(self.ui.widget)
self.axes = Axes3D(self.fig)
# Create the navigation toolbar, tied to the canvas
self.mpl_toolbar = NavigationToolbar(self.canvas, self.ui.widget, coordinates=False)
# connecting signals
self.ui.cmb_function.currentIndexChanged.connect(self.plot)
self.ui.dspb_alpha.valueChanged.connect(self.plot)
self.ui.dspb_smooth.valueChanged.connect(self.plot)
self.ui.chb_normxy.stateChanged.connect(self.plot)
# Vertical layout for canvas and toolbar
vbox = QtGui.QVBoxLayout()
vbox.addWidget(self.canvas)
vbox.addWidget(self.mpl_toolbar)
self.ui.widget.setLayout(vbox)
self.ui.widget.setFocus(QtCore.Qt.MouseFocusReason)
# Create a Qt application
app = QApplication(sys.argv)
window = QDialog()
opf = optim_plotting_frame(x = np.array([ 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000.]),
y = np.array([ 100., 100., 100., 100., 100., 100., 300., 300., 300., 300., 300., 300., 500., 500., 500., 500., 500., 500., 700., 700., 700., 700., 700., 700., 900., 900., 900., 900., 900., 900., 1000., 1000., 1000., 1000., 1000., 1000.]),
z = np.array([374712.60107421875, 526249.09765625, 500842.119140625, 391724.2041015625, 329192.123046875, 298277.92041015625, 526249.259765625, 601555.873046875, 598078.173828125, 529956.01953125, 502884.986328125, 485526.5244140625, 500841.181640625, 598078.400390625, 587555.86328125, 530815.837890625, 495623.544921875, 474902.572265625, 391725.0869140625, 529956.8408203125, 530815.6259765625, 447601.33081054688, 402540.9443359375, 385187.92944335938, 329192.2392578125, 502885.27734375, 495623.6396484375, 402541.17431640625, 365774.16870117188, 343962.6298828125, 298277.88305664062, 485526.775390625, 474903.0673828125, 385187.75439453125, 343962.728515625, 326735.05065917969]),
minx = 100,
maxx = 1000,
miny = 100,
maxy = 1000,
xname = 'width',
yname = 'height',
zname = 'WOPT')
opf.show()
sys.exit(app.exec_())
还有两个问题： 1）轴网格与我的表面重叠。 2）sirface的旋转和缩放速度很慢。
当我不使用 QWidget 时，一切正常（下面的示例），但我应该在我的应用程序中绘制这个表面并且我需要在 QWidget 上进行。因此，任何解决此问题的建议将不胜感激。
import numpy as np
from scipy.interpolate import Rbf
import matplotlib.pyplot as plt
import matplotlib.colors as colors
from matplotlib import cm
from mpl_toolkits.mplot3d import Axes3D
def plotSurface(x, y, z, alpha=None, function=None, smooth=None,
minx=None, maxx=None, miny=None, maxy=None, xname=None, yname=None, zname=None):
"""
This function will create new window with plotted surface on it.
Parameters:
x: numpy.ndarray
set of first parameter data (first coordinate)
y: numpy.ndarray
set of second parameter data (second coordinate)
z: numpy.ndarray
set of data corresponding to x and y (result)
alpha: float, optional
Number between 0 and 1 that defines transparency: 0 - transparent,
1 - not transparent.
function: string, optional
defines radial basis function for interpolation, by default - 'multiquadric'
'multiquadric': sqrt((r/self.epsilon)**2 + 1)
'inverse': 1.0/sqrt((r/self.epsilon)**2 + 1)
'gaussian': exp(-(r/self.epsilon)**2)
'linear': r
'cubic': r**3
'quintic': r**5
'thin_plate': r**2 * log(r)
smooth: float, optional
Values greater than zero increase the smoothness of the approximation.
0 is for interpolation (default), the function will always go through
the nodal points in this case.
minx: float, opyional
minimum possible value of x
maxx: float, opyional
maximum possible value of x
miny: float, opyional
minimum possible value of y
maxy: float, opyional
maximum possible value of y
xname: string, optional
name of x parameter
yname: string, optional
name of y parameter
zname: string, optional
name of z parameter (objective function)
"""
#==========================Data validation=================================
if any((not isinstance(x, np.ndarray),
not isinstance(y, np.ndarray),
not isinstance(z, np.ndarray))):
print "Error: x, y, z must be of numpy.ndarray type."
return None
if x.size != y.size != z.size:
print "Error: x, y, z must be of equal size."
return None
if not isinstance(alpha, (float, int)):
alpha = 0.7
if not isinstance(function, str):
try:
function = str(function)
except:
function = 'multiquadric'
if not any((function == 'multiquadric',
function == 'inverse',
function == 'gaussian',
function == 'linear',
function == 'cubic',
function == 'quintic',
function == 'thin_plate')):
function = 'multiquadric'
if smooth is None:
smooth = 0
if not isinstance(minx, (float, int)):
minx = x.min()
if not isinstance(maxx, (float, int)):
maxx = x.max()
if not isinstance(miny, (float, int)):
miny = y.min()
if not isinstance(maxy, (float, int)):
maxy = y.max()
if minx > maxx:
tmp = minx
minx = maxx
maxx = tmp
if miny > maxy:
tmp = miny
miny = maxy
maxy = tmp
if not isinstance(xname, str):
xname = ""
if not isinstance(yname, str):
yname = ""
if not isinstance(zname, str):
zname = ""
#==========================================================================
fig = plt.figure()
# setting diagram background
fig.patch.set_facecolor('white')
ax = Axes3D(fig)
# map to [0,1] range
x = (x - minx) / maxx
y = (y - miny) / maxy
try:
# getting coordinate matrices from two coordinate vectors.
tx = np.linspace(x.min(), x.max(), 100)
ty = np.linspace(y.min(), y.max(), 100)
XI, YI = np.meshgrid(tx, ty)
# interpolating by radial basis function
rbf = Rbf(x, y, z, function=function, smooth=smooth)
# getting interpolation function results corresponding to (XI, YI)
ZI = rbf(XI, YI)
# plotting interpolaed surface
ax.plot_surface(XI, YI, ZI, cmap=cm.jet, alpha=alpha)
except Exception as e:
print "Error occured! original message: " + e.message
# plotting initial points
ax.scatter(x, y, z)
ax.set_xlim(x.min(), x.max())
ax.set_ylim(y.min(), y.max())
ax.set_title('RBF interpolation ' + function)
# setting ticks positions on the x line
ax.set_xticklabels((ax.get_xticks() * maxx + minx).round(1))
# setting ticks labelson the y line
ax.set_yticklabels((ax.get_yticks() * maxy + miny).round(1))
ax.set_xlabel(xname)
ax.set_ylabel(yname)
ax.set_zlabel(zname)
# adding colorbar
m = cm.ScalarMappable(cmap=cm.jet)
m.set_array(ZI)
ax.figure.colorbar(m)
# showing window with diagram
plt.show()
plotSurface(x = np.array([ 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000., 100., 300., 500., 700., 900., 1000.]),
y = np.array([ 100., 100., 100., 100., 100., 100., 300., 300., 300., 300., 300., 300., 500., 500., 500., 500., 500., 500., 700., 700., 700., 700., 700., 700., 900., 900., 900., 900., 900., 900., 1000., 1000., 1000., 1000., 1000., 1000.]),
z = np.array([374712.60107421875, 526249.09765625, 500842.119140625, 391724.2041015625, 329192.123046875, 298277.92041015625, 526249.259765625, 601555.873046875, 598078.173828125, 529956.01953125, 502884.986328125, 485526.5244140625, 500841.181640625, 598078.400390625, 587555.86328125, 530815.837890625, 495623.544921875, 474902.572265625, 391725.0869140625, 529956.8408203125, 530815.6259765625, 447601.33081054688, 402540.9443359375, 385187.92944335938, 329192.2392578125, 502885.27734375, 495623.6396484375, 402541.17431640625, 365774.16870117188, 343962.6298828125, 298277.88305664062, 485526.775390625, 474903.0673828125, 385187.75439453125, 343962.728515625, 326735.05065917969]),
alpha = 0.7,
function = 'multiquadric',
smooth = 0.0,
minx = 100,
maxx = 1000,
miny = 100,
maxy = 1000,
xname = 'width',
yname = 'height',
zname = 'WOPT')
【问题讨论】：
【参考方案1】：使用 matplotlib 旋转和缩放总是一个相当慢的操作，因为它需要再次渲染整个图形。
【讨论】：
以上是关于matplotlib FigureCanvasQTAgg：QWidget 上的慢速旋转和网格重叠表面的主要内容，如果未能解决你的问题，请参考以下文章