best practice for dynamic cell output with ipython kernel in jupyter

[Denis Akhiyarov]

Hi Jupyter users/hackers and developers,

What is the best practice for dynamic cell output with ipython kernel in jupyter?

Is this the only approach right now:

https://github.com/jupyter-incubator/dashboards

I also noticed pyxley and bokeh.

Basically I need to keep a table with dynamic updates in its columns.

Another wild idea is using widgets, is this possible??

Thanks,

Denis

[MinRK]

widgets would be the logical starting place for something like that. qgrid has recently added an IPython widget representation.

-MinRK

​

--
You received this message because you are subscribed to the Google Groups "Project Jupyter" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jupyter+u...@googlegroups.com.
To post to this group, send email to jup...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jupyter/39695bbf-158b-4d27-9d2f-b7820dd171d2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Denis Akhiyarov]

qgrid does not support coloring cells in grid and is it even dynamic?

https://github.com/quantopian/qgrid/issues/28

[Denis Akhiyarov]

just to clarify: I need to avoid redrawing the table for every update in data, but only change specific columns, rows, or cells. Plus update dynamically, based on changes in data. Is this too much? :)

[Peter Parente]

Another option is to use the channel concept in https://github.com/jupyter-incubator/declarativewidgets. It builds on ipywidgets and gives you a pretty simple mechanism for publishing whatever data you want to the frontend and binding it to UI representations there. We just pushed out a tutorial of using it to build a streaming data dashboard. Look down in the "Publish Data to Widget Channels" section. (The table of contents links don't seem to work in the GitHub rendering.)

https://github.com/jupyter-incubator/dashboards/blob/master/etc/notebooks/stream_demo/meetup-streaming.ipynb

Note the tutorial is much bigger than your use case. You don't *have* to use dashboards, Spark, streaming, etc. I'm just pointing you to the one example use channel use that I know. There might be other examples in the declarative widgets repo.

We haven't gotten that tutorial it to the tmpnb site for the incubator assets we've setup (jupyter.cloudet.xyz). But you can get a temp notebook there and drag/drop upload it and experiment yourself.

[Gino Bustelo]

Just to build on what Pete said above. We just added capabilities to set and watch data from the kernel code. This means that you can build a template for the ui using the declarative approach and then modify the data it is bound to using Python code. See an example athttps://github.com/jupyter-incubator/declarativewidgets/blob/master/notebooks/demos/Walkthrough.ipynb.