first off: great library! I have been working with bokeh for a while now and it is a great piece of software. Please keep up the good work.
But now my actual problem: I am trying to use bokeh in combination with iPython notebooks and it simply does not work. The iPython notebook keeps telling me it is busy and does not allow any further interaction (cells, iPython notebook buttons…). Below is a minimal code example:
It reports that BokehJS has been successfully loaded but the status remains busy. Below you will find the Chome console output:
Default extension for cell metadata editing loaded.
rawcell.js:87 Raw Cell Format toolbar preset loaded.
slideshow.js:48 Slideshow extension for metadata editing loaded.
main.js:29 patching CM for undefined indent
http://localhost:8888/static/fonts/fontawesome-webfont.woff?v=4.1.0 Failed to load resource: the server responded with a status of 404 (Not Found)
http://localhost:8888/static/fonts/fontawesome-webfont.ttf?v=4.1.0 Failed to load resource: the server responded with a status of 404 (Not Found)
kernel.js:124 Kernel started: 0ab24091-b7f5-4d1a-86ea-7bec5d104856
kernel.js:151 Starting WebSockets: ws://localhost:8888/api/kernels/0ab24091-b7f5-4d1a-86ea-7bec5d104856
VM234:22 Bokeh: setting prefix to /
VM234:25 Bokeh: setting log level to: ‘info’
I tried variations of this minimal example with url=‘default’ as explicit parameter to output_notebook() as well as separating them to different iPython notebook cells. The code is not executed double and always at the start of the script just like the documentation requests. Here is some information about my setup:
Fedora Linux 21 4.0.4-202.fc21.x86_64
bokeh 0.9.0 (pip)
I tested it in two browsers, Chrome as well as Firefox, version 43.0.2357.125 (64-bit) or 38.0.5 respectively. Both show the very same behavior. Rebooting machine/iPython notebook server did not change anything. There is some strange thing that I observed, though. If I try to run other code lines directly after output_notebook() in the same cell as that call, they will actually be executed and display a plot. Since I would like to use the iPython notebook in an interactive fashion, just running the notebook in whole is not acceptable.
Any clues what I am doing wrong and/or needs to be fixed in bokeh/iPython/…?