首页
学习
活动
专区
圈层
工具
发布
社区首页 >问答首页 >组合包含matplotlib图形的多个TraitsUi视图

组合包含matplotlib图形的多个TraitsUi视图
EN

Stack Overflow用户
提问于 2018-03-19 19:55:59
回答 1查看 358关注 0票数 1

我正在尝试使用TraitsUI构建一个交互式图像查看器。这可以使用皮埃尔·海西格共享的实现来完成。

接下来,我想同时提出其中的许多观点。这里也概述了类似的方法。然而,我无法用这种方法来解决我手头的问题。构建这样一个容器会抛出一个AttributeError: 'Test' object has no attribute 'dpi'。我想我误解了编辑在这种情况下的工作方式,但我想不出如何使其工作。

可运行的示例代码:

代码语言:javascript
复制
# By Pierre Haessig, https://gist.github.com/pierre-haessig/9838326
from pyface.qt import QtGui, QtCore

import matplotlib
# We want matplotlib to use a QT backend
matplotlib.use('Qt4Agg')
from matplotlib.backends.backend_qt4agg import FigureCanvasQTAgg as FigureCanvas
from matplotlib.figure import Figure

from traits.api import Any, Instance
from traitsui.qt4.editor import Editor
from traitsui.qt4.basic_editor_factory import BasicEditorFactory

class _MPLFigureEditor(Editor):

    scrollable  = True

    def init(self, parent):
        self.control = self._create_canvas(parent)
        self.set_tooltip()

    def update_editor(self):
        pass

    def _create_canvas(self, parent):
        """ Create the MPL canvas. """
        # matplotlib commands to create a canvas
        mpl_canvas = FigureCanvas(self.value)
        return mpl_canvas

class MPLFigureEditor(BasicEditorFactory):

    klass = _MPLFigureEditor


if __name__ == "__main__":
    # Create a window to demo the editor
    from traits.api import HasTraits, Int, Float, on_trait_change
    from traitsui.api import View, Item
    from numpy import sin, cos, linspace, pi

    class Test(HasTraits):

        figure = Instance(Figure, ())
        n = Int(11)
        a = Float(0.5)

        view = View(Item('figure', editor=MPLFigureEditor(), show_label=False),
                    Item('n'),
                    Item('a'),
                    width=400,
                    height=300,
                    resizable=True)

        def __init__(self):
                super(Test, self).__init__()
                axes = self.figure.add_subplot(111)
                self._t = linspace(0, 2*pi, 200)
                self.plot()

        @on_trait_change('n,a')
        def plot(self):
            t = self._t
            a = self.a
            n = self.n
            axes = self.figure.axes[0]
            if not axes.lines:
                axes.plot(sin(t)*(1+a*cos(n*t)), cos(t)*(1+a*cos(n*t)))
            else:
                l = axes.lines[0]
                l.set_xdata(sin(t)*(1+a*cos(n*t)))
                l.set_ydata(cos(t)*(1+a*cos(n*t)))
            canvas = self.figure.canvas
            if canvas is not None:
                canvas.draw()

    ##############
    # This works #
    ##############
    t = Test()
    t.configure_traits()

    class Container(HasTraits):
        p1 = Instance(Test)
        p2 = Instance(Test)

        view = View(
            Item("p1", editor=MPLFigureEditor()),
            Item("p2", editor=MPLFigureEditor())
            )

    ##############
    # This fails #
    ##############
    c = Container(p1=Test(), p2=Test())
    c.configure_traits()
EN

回答 1

Stack Overflow用户

回答已采纳

发布于 2018-08-02 03:44:53

Test类本身不是matplotlib图,这就是为什么为p1p2指定editor=MPLFigureEditor()会导致错误。如果您相反地设置了style='custom',这两个数字都会出现。

代码语言:javascript
复制
class Container(HasTraits):
    p1 = Instance(Test)
    p2 = Instance(Test)

    view = View(
        Item("p1", style='custom'),
        Item("p2", style='custom')
        )
票数 1
EN
页面原文内容由Stack Overflow提供。腾讯云小微IT领域专用引擎提供翻译支持
原文链接:

https://stackoverflow.com/questions/49371365

复制
相关文章

相似问题

领券
问题归档专栏文章快讯文章归档关键词归档开发者手册归档开发者手册 Section 归档