non-png output #4
Comments
|
@tobyhodges PNG definitely isn't a hard requirement. Currently this works by extracting the PNG from the notebook and saving it to disk. The filename gets passed down into the Jinja template. I can easily modify this. Can you submit a PR with a single Bokeh example that either renders a SVG or a PNG? I can figure out how to integrate it and we can go from there. |
|
The most important thing to me right now is that the plots are rendered in a way that I can display responsively using Bootstrap. PNG is nice because I can just use the img-fluid class. That should work with SVGs too. |
|
@tobyhodges: I created a way to display Bokeh plots: http://pythonplot.com/index.html#simple-histogram The notebook is just running |
|
For some reason, I'm able to render these locally (on my Mac), but it's failing on travis https://travis-ci.org/tdhopper/pythonplot.com/builds/247215019 |
|
Given this, I think I'm going to give up on using Bokeh charts for the time being. |
|
Ok, thanks for investigating. I'll keep an eye on the issue discussion in the |
This is a really nice idea, and a good start. After reading the note about Bokeh in the Readme, my immediate thought was that I could contribute some examples using
bokeh.chartsorbokeh.plotting.Unfortunately, this is limited by the requirement for plots to be output as
pngfrom the code cells in the notebook. Bokeh output is typically HTML, which can be rendered inline in a Jupyter notebook withbokeh.io.output_notebook()(docs). With the latest release, it became possible to output aspngandsvgbut these options either require additional dependencies, or are incompatible with output into a notebook. For more, see here and here.If you are happy to accept examples that either don't output
pnginto the notebook, or that do outputpngbut not inline and with additional dependencies onselenium, I would like to contribute.The text was updated successfully, but these errors were encountered: