Troubleshooting and FAQ

Note

Depending on which backends you want to use, there may be additional steps required for installation; consult the advice here.

Jupyter Notebook Issues

When starting a jupyter notebook, I get a “Permission denied” error for a linking operation.

This may be related to jupyter upgrades. Manually remove the symlink and the notebook should be able to proceed once more.

When running in a jupyter notebook, nothing is displayed.

The solution to this problem depends on more details.

  • The canvas is displayed entirely black with “Uncaught TypeError: Cannot read property ‘handle’ of undefined” (or similar language): After the canvas.show() command in the cell, add a line import time;time.sleep(.1). You may need to increase the argument of time.sleep(). This is due to a race condition in vispy.
  • I get an error 404 in the browser console for vispy.min.js - Make sure that jupyter, ipywidgets, and all of the jupyter components are up to date (and have compatible versions, see https://bitbucket.org/snippets/glotzer/nMg8Gr/plato-dependency-installation-tips ).
  • I get an error 404 in the browser console for webgl-backend.js - Try removing your jupyter notebook cache (~/.jupyter and ~/Library/Jupyter on OSX) and restarting jupyter
  • Make sure the jupyter executable you are using is in the same virtualenv or conda environment as plato and its dependencies

Things aren’t displayed and I get a message “zmq message arrived on closed channel” in the console.

Try running your jupyter notebook command with an increased data rate limit:

jupyter notebook --NotebookApp.iopub_data_rate_limit=1000000000