Python - pass arguments when adding a button to plot's toolbar (matplotlib)

Question

I am trying to add a button to toolbar. The button is a custom class that inherits ToolToggleBase. In my custom SelectButton class, I want to pass few arguments, but I getting an error:

import numpy as np
import matplotlib.pyplot as plt
plt.rcParams["toolbar"] = "toolmanager"
from matplotlib.backend_tools import ToolToggleBase


def simple_plot(mat):
    fig = plt.figure()
    ax = fig.add_subplot()
    ax.plot(np.reshape(mat, [-1, 1]))
    ax.grid(True)
    ax.legend()
    tm = fig.canvas.manager.toolmanager
    tm.add_tool("CustomButton", SelectButton(fig, ax))
    fig.canvas.manager.toolbar.add_tool(tm.get_tool("CustomButton"), "toolgroup")
    return fig, ax


class SelectButton(ToolToggleBase):
    default_toggled = False

    def __init__(self, fig1, ax1, *args, **kwarg):
        super().__init__(*args, **kwarg)
        print("fig: ", fig1)
        print("ax: ", ax1)


x = [1, 2, 3]
fig, ax = simple_plot(x)
plt.show()

Results in:

super().init(*args, **kwargs) TypeError: init() missing 2 required positional arguments: 'toolmanager' and 'name'

However, if I do not pass any arguments and pass only the class name, everything works as expected:

import numpy as np
import matplotlib.pyplot as plt
plt.rcParams["toolbar"] = "toolmanager"
from matplotlib.backend_tools import ToolToggleBase


def simple_plot(mat):
    fig = plt.figure()
    ax = fig.add_subplot()
    ax.plot(np.reshape(mat, [-1, 1]))
    ax.grid(True)
    ax.legend()
    tm = fig.canvas.manager.toolmanager
    tm.add_tool("CustomButton", SelectButton)
    fig.canvas.manager.toolbar.add_tool(tm.get_tool("CustomButton"), "toolgroup")
    return fig, ax


class SelectButton(ToolToggleBase):
    default_toggled = False

    def __init__(self, *args, **kwarg):
        super().__init__(*args, **kwarg)

x = [1, 2, 3]
fig, ax = simple_plot(x)
plt.show()

I understand I must call add_tool wrong when use an instance of a class rather then the class name, but I fail to understand what am I supposed to do if I want to pass arguments to SelectButton class.

Stef · Accepted Answer · 2023-06-28 10:23:36Z

You need to pass the class SelectButton itself, not an instance SelectButton() of it to add_tool. You should pass the additional parameters as kwargs:

def simple_plot(mat):
    fig = plt.figure()
    ax = fig.add_subplot()
    # ...
    tm = fig.canvas.manager.toolmanager
    tm.add_tool("CustomButton", SelectButton, fig=fig, ax=ax)
    fig.canvas.manager.toolbar.add_tool(tm.get_tool("CustomButton"), "toolgroup")
    return fig, ax


class SelectButton(ToolToggleBase):
    default_toggled = False

    def __init__(self, *args, fig, ax, **kwarg):
        super().__init__(*args, **kwarg)
        print("fig: ", fig)
        print("ax: ", ax)

This prints:

fig:  Figure(640x480)
ax:  Axes(0.125,0.11;0.775x0.77)

If you for some reason really want to use args instead of kwargs you should use a partial as shown in the other answer. Passing only the first 2 args to super and using the rest by yourself is more a hack than a solution as it depends on matplotlib's implementation details:

tm.add_tool("CustomButton", SelectButton, fig, ax)

def __init__(self, *args, **kwarg):
    super().__init__(*args[:2], **kwarg)
    print("fig: ", args[2])
    print("ax: ", args[3])

Karo · Accepted Answer · 2023-06-28 09:54:51Z

When utilizing an instance of the SelectButton class, you are using add_tool incorrectly, which results in the problem you are experiencing. A class instance, not a string representation of the class name, is what the add_tool method anticipates.

You can use a lambda function or partial function to build a callable object that accepts the necessary arguments to pass arguments to the SelectButton class. Here is an illustration of how you can change your code to accomplish this:


import numpy as np
import matplotlib.pyplot as plt
plt.rcParams["toolbar"] = "toolmanager"
from matplotlib.backend_tools import ToolToggleBase
from functools import partial


def simple_plot(mat):
    fig = plt.figure()
    ax = fig.add_subplot()
    ax.plot(np.reshape(mat, [-1, 1]))
    ax.grid(True)
    ax.legend()
    tm = fig.canvas.manager.toolmanager
    create_button = partial(SelectButton, fig, ax)
    tm.add_tool("CustomButton", create_button)
    fig.canvas.manager.toolbar.add_tool("CustomButton", "toolgroup")
    return fig, ax


class SelectButton(ToolToggleBase):
    default_toggled = False

    def __init__(self, fig1, ax1, *args, **kwargs):
        super().__init__(*args, **kwargs)
        print("fig:", fig1)
        print("ax:", ax1)


x = [1, 2, 3]
fig, ax = simple_plot(x)
plt.show()

The create_button variable in this code is given a partial function that accepts the arguments fig and ax and returns a brand-new instance of the SelectButton class. The tool name and this incomplete function are then supplied to add_tool.

The error should now be fixed when you execute the changed code since the SelectButton class will now correctly get the fig and ax arguments.

Collectives™ on Stack Overflow

Python - pass arguments when adding a button to plot's toolbar (matplotlib)

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related