VPython7 issue with Jupyter Notebook

[Michael Bentley]

Hello

I am teaching an undergrad course where I extensively use VPython. Last year I used VPython 6 (Classic) in Python 2.7 with VIDLE and have now upgraded to using VPython 7 with Python 3.6. I want to use the Jupyter notebook to teach the course, since it has many advantages, but I have hit a problem I cant solve. 

When running VPython7 from within Jupyter, the graphics window is displayed properly when the program/cell is executed. However, if I try to run the program/cell (containing VPython code) again, the graphics window does not appear. I believe this is because the graphics environment is still running from the last execute (I have seen posts about this). I presume this is so that the user can still interact with the window, change views etc. The only way to stop this completely, to enable a re-run of the code, seems to be to kill the kernel and clear the output. But then, if using multiple cells, all cells would need to be re-run since the kernel has been cleared. This seems to defeat the point of the notebook, and this is very painful way develop some code using VPython.

Is there a way to stop/kill/end the VPython display cleanly, without clearing the kernel? I had thought that VPython7 was designed with Jupyter in mind, so I have probably missed something crucial here. 

Thanks

Mike

[Aaron Titus]

I’ve used VPython in Jupyter for a while, and I’ve found that creating a scene explicitly by calling canvas() before creating 3D objects is essential. Try using:

scene=canvas(title="My 3D scene”)

sphere()

This also allows me to embed multiple scenes in a single notebook. Each scene will appear below the cell where it is created. Also, when I edit the cell and re-execute the cell, the scene will be updated properly and will not disappear.

I occasionally have to restart the kernel. But I think calling canvas() solves most of the problems you described.

Aaron

--
You received this message because you are subscribed to the Google Groups "VPython-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vpython-user...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Michael Bentley]

Aaron - thank you. This has indeed solved some of the problems, but not all.

I also heavily use the gcurve() graphing function. For example, I might use the following to draw two graphs:

win1 = graph(xtitle="Distance [m]", ytitle="Height [m]")

win2 = graph(xtitle="Distance [m]", ytitle="Height [m]")

c1 = gcurve(graph=win1, color=color.green)

c2 = gcurve(graph=win2, color=color.red)

I dont see how I can use the scene() function here, so I end up the same problem once both graphs are drawn... 

Any further help you could give would be much appreciated. 

Thanks

Mike

[Michael Bentley]

Apologies - I have it now:

scene1 = canvas()

scene2 = canvas()

g1 = graph(canvas=scene1, xtitle="Distance [m]", ytitle="Height [m]")

g2 = graph(canvas=scene2, xtitle="Distance [m]", ytitle="Height [m]")

c1 = gcurve(graph=g1, color=color.green)

c2 = gcurve(graph=g2, color=color.red)

Thank you for the help

Mike

[Aaron Titus]

Super. Thanks for letting us know.

Aaron

[Bruce Sherwood]

Two comments:

Although it's not caught as an error in VPython 7, "canvas" is not an attribute of a graph. It should give an error (it does in GlowScript VPython) but because there isn't an explicit check for it it's treated as a "user attribute" (that is, an attribute that VPython doesn't do anything with).

Also, note that if you aren't actually using the Jupyter notebook as a presentation and/or documentation device, but you're just writing and testing VPython programs, you might prefer to launch programs from IDLE or other IDE. See the first page of vpython.org for details.

Bruce

[Todd Zimmerman]

Thanks Aaron, I've been having this problem too and was just about to post to this forum.

-Todd