Matplotlib pick event from multiple series

Question

I am doing a scatterplot where I want to click on the individual points to do something. This is like the existing example code.

https://matplotlib.org/examples/event_handling/pick_event_demo.html

I have implemented an on_pick method

def on_pick(event):
    ind = event.ind
    for i in ind:
        ...do something here with blue or red data...

However, I am stuck because I putting multiple series (red and blue) in the same plot

fig, ax = plt.subplots()
ax.set_title('click on a point...')
line, = ax.plot(red_xs, red_ys, 'o', picker=5, color='red') 
line, = ax.plot(blue_xs, blue_ys, 'o', picker=5, color='blue')

The event.ind is a collection of integers. They are indexes into a series. However, there seems to be no way to determine which series they are an index into.

There must be a way to do this. Does anyone know the trick?

Thank you Peter

opus111 · Accepted Answer · 2017-08-11 16:15:35Z

Here is the solution I found. It seems clunky, but maybe it is the best one.

By writing a "custom picker" one can add meta data to the event. Here I add "ind" and "series"

def picker(self,line,mouseevent,series):
    if mouseevent.xdata is None:
        return False, dict()
    xdata = line.get_xdata()
    ydata = line.get_ydata()
    maxd = 0.05
    d = np.sqrt((xdata - mouseevent.xdata) ** 2. + (ydata - mouseevent.ydata) ** 2.)

    ind = np.nonzero(np.less_equal(d, maxd))
    if len(ind):
        props = dict(ind=ind, series=series)
        return True, props
    else:
        return False, dict()

One one can then attach a different "custom picker" to each scatter plot

    line, = ax.plot(self.red['False'], self.red['True'], 'o', picker=lambda line,mouseevent: self.picker(line,mouseevent,self.red), color='red')
    line, = ax.plot(self.blue['False'], self.blue['True'], 'o', picker=lambda line,mouseevent: self.picker(line,mouseevent,self.blue), color='blue')

And then pull the metadata off in the on_pick() function

fig.canvas.mpl_connect('pick_event', lambda e: self.on_pick(e))

...

def on_pick(self,event):
    for i in event.ind:
        for j in i:
            series = event.series
            ...do something with item j of series...

ImportanceOfBeingErnest · Accepted Answer · 2017-08-12 00:16:41Z

2

The pick_event_demo you link to acutally tells you how to know which line is which. It says

thisline = event.artist
xdata = thisline.get_xdata()
ydata = thisline.get_ydata()
ind = event.ind

So ind will index whatever thisline is.

To give a more thorough example

line1, = ax.plot(self.red['False'], self.red['True'], 'o', picker=5, color='red') 
line2, = ax.plot(self.blue['False'], self.blue['True'], 'o', picker=5, color='blue') 

dic = {line1 : self.red, line2 : self.blue}

def on_pick(self,event):
    series = dic[event.artist]
    # do something with series
    for i in event.ind:
         print(series[i])


fig.canvas.mpl_connect('pick_event', on_pick)

edited Aug 12, 2017 at 0:16

answered Aug 10, 2017 at 20:15

ImportanceOfBeingErnest

342k61 gold badges739 silver badges771 bronze badges

3 Comments

opus111 Over a year ago

Thank you IOBE. I think I wasn't clear in my question. I need to attach additional meta-data to the event, which in my case if the Pandas dataframe that generated the x's and y's. I found one way to do it. Maybe you know of a better way.

ImportanceOfBeingErnest Over a year ago

If your solution works, that's fine. My point was that you already have the data available as xdata and ydata inside the on_pick function. If instead you need the original data, you may use a simple mapping of the artist to the original data. I updated the answer to show how that could be done.

MLNN Over a year ago

This is a good answer, event.artist actually shows which series the event is called on.

Collectives™ on Stack Overflow

Matplotlib pick event from multiple series

2 Answers 2

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related