compilation failed - Attribute error in example I of continuos time convolution

[TheFabulousDeath]

I tried running the example in Binder and locally and both versions had the same (lengthy) compilation error. I cut out the parts of the message that just referenced elements in conda, so here is the shortened one:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[5], line 6
      2 xB = h
      3 y = sp.integrate(xA.subs(t, t-tau) * xB.subs(t, tau),
      4                  (tau, 0, t))  # for causal signals
----> 6 anim = animate_convolution(
      7     x=xA,
      8     h=xB,
      9     y=y,
     10     td=np.arange(-2, 10, .1),
     11     taud=np.arange(-10, 10, 0.01),
     12     xlim=(-2, 5))
     13 anim

Cell In[2], line 63, in animate_convolution(x, h, y, td, taud, xlim)
     59     lines[3].set_data(ti, y_eval(ti))
     61     return lines
---> 63 return animation.FuncAnimation(fig, animate, td, interval=50, blit=True)

Cell In[2], line 51, in animate_convolution.<locals>.animate(ti)
     50 def animate(ti):
---> 51     ax[0].collections.clear()
     52     ax[0].fill_between(taud, 0, h_eval(ti, taud) *
     53                        x_eval(taud), facecolor='gray', alpha=0.3)
     55     lines[0].set_data(taud, h_eval(ti, taud))

AttributeError: 'ArtistList' object has no attribute 'clear'

Changing line 51 from ax[0].collections.clear() to ax[0].clear() fixed the compilation error, however I don't know if the result is the intended one.

https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.subplots.html - the documentation for the command used to create the subplot of which the collection was tried to be deleted.

Thank you so much for providing this platform, it's much appreciated!

[fs446]

Thanks for this hint, highly appreciated!

I assume that 11ecd7a fixes this issue, which was due to a matplotlib API change. The new handling might not be the most elegant, but we then can still use fill_between. New handling is: we give each fill_between one and the same label, and this can be found in a collection list, and once we found it, we can remove it. Currently, jupyter notebook seems to have problems with tornado, I had some flickering issues therefore in the animations. Downgrading to tornado 6.1 solved this.