Using patch & stream with Jupyter notebook

I’ve adapted this script;

http://pastebin.com/cU52J5Sw

from a simple Bokeh patch() and stream() demo kindly provided by Bryan at Bokeh. It shows a continuously updating plot. It was originally intended to run using the Bokeh server, I’m hacking it around to try and get it to run in a Jupyter notebook but an getting the error:

  ' warnings.warn(_NB_LOAD_WARNING)
C:\Users\GoFaster\Anaconda3\envs\py35\lib\site-packages\bokeh\io.py:582: UserWarning: Cannot find a last shown plot to update. Call output_notebook() and show() before push_notebook()
warnings.warn("Cannot find a last shown plot to update. Call output_notebook() and show() before push_notebook()")'

Is there a way to make it function as intended? I believe that show(), output_notebook() & push_notebook() are being called in the correct order.

Thanks

GoFaster

I am afraid for now this will not be possible. Notebook and Bokeh Server integration is not yet complete, there is an open issue you can track to follow progress:

  Notebook + Server Not Working · Issue #3461 · bokeh/bokeh · GitHub

Unfortunately integrating even static inline plots robustly with the notebook has been extremely challenging. In all likelihood we are going to have to switch to some genuine jupyter extension appraoch to make all this really robust, though I am afraid that will result in some lost convenience for users (Jupyter plugins have to be installed manually by users, AFAIK). In any event the point is that there is some serious discussion, planning, design, and testing that needs to happen first, before we can really devote effort the notebook and server combination.

Thanks,

Bryan