by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 15:23
    jonmmease commented #4532
  • 14:44
    jbednar commented #4532
  • 14:41
    MarcSkovMadsen commented #1252
  • 14:31
    jlstevens commented #4532
  • 13:04
    kcpevey commented #1499
  • 12:31
    nickderobertis commented #1252
  • 11:44

    philippjfr on table_auto_edit

    (compare)

  • 11:44

    philippjfr on master

    Expose auto_edit option on Data… (compare)

  • 11:44
    philippjfr closed #1520
  • 11:34
    philippjfr synchronize #1520
  • 11:34

    philippjfr on table_auto_edit

    Do not round-trip table edits (compare)

  • 11:20
    philippjfr labeled #1520
  • 11:20
    philippjfr opened #1520
  • 11:20

    philippjfr on table_auto_edit

    Expose auto_edit option Do not trigger update on edit (compare)

  • 11:10

    philippjfr on location_fix

    (compare)

  • 11:10

    philippjfr on master

    Fix adding and deleting Locatio… (compare)

  • 11:10
    philippjfr closed #1519
  • 11:10
    codecov[bot] commented #1519
  • 11:08
    philippjfr edited #1519
  • 11:07
    codecov[bot] commented #1519
Philipp Rudiger
@philippjfr
We should definitely add binder links to all of our docs.
Enis Arik
@earik87
Hi everyone! I am not sure if I should ask here or not. Discourse does not seem super active :)
I have an issue with axiswise=True command. It does not work with my figures. There is still y-axis sharing between them which I dont want. It is strange because I have 4 figures, and for first two this command does work, but not for the last 2 :)
Enis Arik
@earik87
sameerCoder
@sameerCoder

Hello Every ,
Resources for deploying dynamic interactive holoviews plot / panel plot i am not finding sufficient.
Can please some one suggest me some good resources / materials / tutorials for dynamic interactive holoviews plot / panel plot deploy with my own server ,
I have my own IBM WebSphere HTTP & Application Server 8.5.1 and i want my holoview / panel plot have dynamic nature with my own IBM WebSphere HTTP & Application Server 8.5.1.

below documents links are not so understable by me
http://holoviews.org/user_guide/Deploying_Bokeh_Apps.html
https://docs.bokeh.org/en/0.9.3/docs/user_guide/server.html

Kindly provide some resources where i can learn above mention requirement,
I will really appreciate it. thank you.

Jacob Barhak
@Jacob-Barhak
I wanted to thank the Holoviz team for their long term support. I just got a COVID-19 model published that heavily relies on dask for HPC and on Holoviz for Visualization. You will find the publication here
Barhak J , The Reference Model initial Use Case for COVID-19. Cureus. Online . Interactive Results are available directly using this link . You will find acknowledgments to dask and Holoviz teams in the paper - those technologies both helped. Holoviz allowed creating the nice interactive visualization. Please feel free to circulate. Hopefully more people will see the potential of python tools you develop.
Jean-Luc Stevens
@jlstevens
very nice and thanks for the acknowledgements!
James A. Bednar
@jbednar
Congratulations!
@sameerCoder , if you are running your own web server, can you not simply log into that server and run panel serve your_file.py? I may be misunderstanding your situation, though.
sameerCoder
@sameerCoder
@jbednar
As we have bokeh server , through which we can render / dynamic natue of plot/ current modify docs is generated same property i want through my own IBM WebSphere HTTP & Application Server 8.5.1 ,
In Sort i want to replace bokeh server with IBM WebSphere HTTP & Application Server 8.5.1.
For example- i have plot in which there is drop down option - as per selecting the drop down option plot change and zooming of plot open street map zoom in as well as plot zoom.- these zooming property and changing of plot is happeining perfect by using bokeh server (bokeh.server()) , now as i already have IBM WebSphere HTTP & Application Server 8.5.1 server so i am thinking insted of using bokeh server; can i do plots with my own IBM WebSphere HTTP & Application Server 8.5.1 server
thank you
Jacob Barhak
@Jacob-Barhak
@jlstevens and @jbednar - this time the acknowledgement was for the team and not per person. Yet you all deserve it in this channel. Without your help, I would have been stuck with much less convincing graphics that cannot show the rich information.
James A. Bednar
@jbednar
You did all the work! :-)
@sameerCoder, ah. We've done some work on making it easier to integrate Panel and Bokeh into Django as a web server, but I don't know of anyone who has done that work for WebSphere.
Jacob Barhak
@Jacob-Barhak
Well @jbednar , you forgot how about two years ago @philippjfr contributed code so I can use Holoviews. You also forgot how many times I was bothering your team while you had different pressing priorities and you still found time to fix things so I can use your latest versions. I may have been persistent, yet your team did help - a lot!
sameerCoder
@sameerCoder
@jbednar
can you please provide the links for integrate Panel and Bokeh into Django as a web server, hope these resources will help me
Acutually while using bokeh server its taking lots of time at intial to load the plot ,but once at intial plot loaded then its easy & fast to change the plot as per selecting the drop down option.
sameerCoder
@sameerCoder
@Jacob-Barhak
Congratulations for your recent published
It really a valuable & knowledgable share published paper.
i too want to show the plot which u r showing in app/website
can you please let me know how you are able to show the data visualization into standalone webpage link:- https://jacob-barhak.netlify.app/thereferencemodel/results_covid19_2020_06_27/combinedplot
I too have holoviews plot I want to publish the data visualization somewhat similar to you so that others can see.
Thank you.
James A. Bednar
@jbednar
Django integration is described at https://panel.holoviz.org/user_guide/Django_Apps.html . But you can export HoloViews and Panel objects to static HTML and then use it on any web site; the Django guide is about enabling live two-way Python<->JavaScript communication. See https://panel.holoviz.org/user_guide/Deploy_and_Export.html#Embedding and "jslink" in https://panel.holoviz.org/user_guide/Links.html .
sameerCoder
@sameerCoder
My holoview plot is not static , plot is dynamic in nature properties like zooming open street map , drop down option to select particular date time plot and states selection.
sameerCoder
@sameerCoder
u can find minimal example at below link
https://discourse.bokeh.org/t/good-resource-for-run-bokeh-app-from-own-server/6045/5?u=sameercoder
Thank you.
I will really appreciate ur help.
Harman Deep Singh
@hdsingh
Hello everyone, I am looking into 3d volume rendering of netcdf data (by converting to numpy arrays). Is there an efficient method available in HoloViews. I know it could be done using Plotly, but I wonder if it will be efficient for large size netcdf arrays (since datashading is not available in Plotly or is it)?
Philipp Rudiger
@philippjfr

I'm currently working on a volume regridding implementation in Datashader: https://github.com/holoviz/datashader/pull/946/files

That combined with the VTKVolume pane in Panel should work pretty well.

sameerCoder
@sameerCoder
@jbednar
Thanks for providing those deployments links but i gone through these deployment will provide & will work as url localhost:8000 but
i want my panel plot dynamic in nature properties like zooming open street map , drop down option to select particular date time plot and states selection and my web browser url to be link mysystemipaddress:owndefinedportnumber (for ex: https://172.16.4.4:50007)
please correct me or guide me to get above mention requirement.
thank you.
Marc Skov Madsen
@MarcSkovMadsen
image.png
image.png
FYI. Yesterday on the Triage meeting there was again a mention of Dark mode for Panel. And in general there is an issue styling things nicely in Panel. I've been trying different approaches 1) custom css to override Bokeh defaults 2) using (Material MWC) web components with a design system. Approach 1. is difficult (I believe) because there is not really a general design system. As far as I can see you have to update in many places instead of one and you (and I) have to be the designer (!!!). The MWC Web Components have design systems supported by css so changing a few css parameters the rest is dynamically calculated and can be used to change from light to dark and back in a live app. Now I've seen something at least as good. Microsoft have released their web components and it comes with a design system. But also json definitions and schemas that might be useful to auto generate a lot of widgets and layout and get a general design system in place. If you would like to get inspiration take a look at https://fast.design and https://fast.design/docs/tools/component-explorer . I believe the definitions and schemas could help speed up the process of wrapping a set of nice looking widgets. I'm currently working on the Material MWC web components using the WebComponent extension. What takes time is to make their interface consistent with the existing widgets such that they support for example usage via pn.Param.
Marc Skov Madsen
@MarcSkovMadsen
One temporary solution without implementing a lot of widgets would also be to support some functionality which is similar to the WebComponent and the Dash callback philosophy. My hypothesis is that in the same way as the WebComponent (or using the WebComponent as a stepping stone) we could actually provide functionality to add widgets directly in the template with an id identifier and then just support bidirectional syncing between that element and its attributes/ properties/ events to a python Parameterized object. In that way there would be much better support for any design system or widget system on the fly. People would not have to wait for improvements in Panel.
Any comments or ideas?
Marc Skov Madsen
@MarcSkovMadsen
The api could be something like
image.png
Marc Skov Madsen
@MarcSkovMadsen
image.png
and the WebComponent api could be changed to something almost identical (above)
and you would get a nice design system that Microsoft and others have spend a lot of resources on
image.png
You would just have to spend a little bit of time creating a light and dark bokeh theme in accordance with this to be used by hvplot, holoviews etc.
Marc Skov Madsen
@MarcSkovMadsen
This would make Panel so much more flexible, powerful and extensible I believe.
Marc Skov Madsen
@MarcSkovMadsen
And this video explains why web components, why design systems, why fast and why Panel needs something like this https://youtu.be/sYTH_xYH3iA?t=862
Marc Skov Madsen
@MarcSkovMadsen
Marc Skov Madsen
@MarcSkovMadsen
I have been trying to get Panel Dashboards running on Kubeflows JupyterHub via https://github.com/holoviz/jupyter-panel-proxy. But without luck. If you have any ideas or experience to share please do so here kubeflow/kubeflow#5194. It would be very helpful for me and my use of Panel to get this working. Thanks.
Steven C. Howell
@StevenCHowell
When I run it using python crossfilter.py I get a traceback
ValueError: No plotting extension is currently loaded. Ensure you load an plotting extension with hv.extension or import it explicitly from holoviews.plotting before applying any options.
I added hv.extension('bokeh') close to the top of the script (just after importing holoviews), and now it runs but I don't get a plot.
Steven C. Howell
@StevenCHowell
I put the code in a Jupyter notebook and now I see the plot.
Marc Skov Madsen
@MarcSkovMadsen
If you would like to checkout an alternative data exploration and development environment for Python DataViz and HoloViz in particular. Checkout the Awesome Panel Designer. I would really like some feedback and ideas for improvements. Checkout the guide here https://awesome-panel.readthedocs.io/en/latest/packages/awesome-panel-extensions/designer.html
awesome-panel-designer-guide.gif
image.png
Kim Pevey
@kcpevey
Is there any reasoning behind the panel.pane.Dataframe and panel.widgets.Dataframe having different options? I want max_rows and index=False from the pane and widths from the widget in order to make my dataframe look decent. Seems like these options could be available on both?