matplotlib how to adjust navigation and cursor display

Question

When plotting two time series on two different y-scales using twinx:

ax1 = pp.subplot(111)
ax2 = ax1.twinx()

the default behavior for navigation seems to be:

1. zoom/pan will control both axes simultaneously
2. the lower-right corner displays the mouse cursor coordinates with y-value of ax2

I often find it useful to turn off zoom/pan for one of the axes with set_navigate (thanks to this tip). However when navigation is turned off for ax2, the coordinates display is also turned off.

Is there anyway to disable zoom/pan without turning off the coordinates display?

Even better, is there a way to specify which one of the two axes to take the y-value from when displaying the cursor coordinates?

[version info: python 3.4.3 + matplotlib 1.5.1]

Answer 1

It seems that there is no simple way (ala set_navigate) to do this. I managed to find this question and modified it a bit to accomplish what I need:

class Cursor(object):
    def __init__(self, fig, ax, ax2=None, xfmt=None, yfmt=None):
        self.fig = fig
        self.ax = ax
        self.ax2 = ax2
        self.xfmt = xfmt
        self.yfmt = yfmt
        plt.connect('button_press_event', self)

    def __call__(self, event):
         if event.button == 3 and self.ax2 is not None: # right click
            ax = self.ax2
        elif event.button:
            ax = self.ax
        inv = ax.transData.inverted()
        x, y = inv.transform((event.x, event.y))
        self.fig.canvas.toolbar.set_message('x={} y={}'.format(x if self.xfmt is None else self.xfmt(x), y if self.yfmt is None else self.yfmt(y)))

I can then instantiate a Cursor like this:

fig,ax = plt.subplots(1,1)
ax2 = ax.twinx()
Cursor(fig, ax, ax2)

and regardless of the navigate status of either axis, left/right click on the plot will display the coordinates of left/right axis in the lower right corner of the navigation toolbar.