Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 05:24
    MarcSkovMadsen edited #2298
  • 05:24
    MarcSkovMadsen edited #2298
  • 05:24
    MarcSkovMadsen labeled #2298
  • 05:24
    MarcSkovMadsen review_requested #2298
  • 05:23
    MarcSkovMadsen edited #2298
  • 05:22
    MarcSkovMadsen synchronize #2298
  • 05:22

    MarcSkovMadsen on discrete-slider-numpy64

    use INT_FORMATTER in test (compare)

  • 05:20
    MarcSkovMadsen opened #2298
  • 05:16
    MarcSkovMadsen labeled #2297
  • 05:16
    MarcSkovMadsen opened #2297
  • 05:16

    MarcSkovMadsen on discrete-slider-numpy64

    fix discreteslider formatting add documentation (compare)

  • May 07 19:04
    nghenzi commented #2292
  • May 07 18:19
    codecov[bot] commented #2296
  • May 07 18:18
    codecov[bot] commented #2296
  • May 07 18:17
    codecov[bot] commented #2296
  • May 07 18:15
    codecov[bot] commented #2296
  • May 07 18:15
    codecov[bot] commented #2296
  • May 07 18:08
    codecov[bot] commented #2296
  • May 07 18:08
    codecov[bot] commented #2296
  • May 07 18:00
    syamajala commented #2293
Milind Sharma
@gewitterblitz
What's the easiest way to create a custom layout in holoviews while using datashader functionality?
Here's my code but it want to customize it by having only subplot in row1 covering two columns
import colorcet as cc

ltg_data = lma.get_active()  # returns a copy of the data, as a Pandas Dataframe
ltg_data = ltg_data[(ltg_data['alt'] <=20)]

points1 = hv.Points(ltg_data,kdims=['lon','alt'])
datashaded1 = hd.datashade(points1,cmap=cc.fire).redim.range(y=(0,20),x=(-87.6,-86.8))
plot1 = hd.dynspread(datashaded1, threshold=0.10, how='over').opts(height=300,width=600)

points2 = hv.Points(ltg_data,kdims=['lat','alt'])
datashaded2 = hd.datashade(points2,cmap=cc.fire).redim.range(y=(0,20),x=(-34,34.4))
plot2 = hd.dynspread(datashaded2, threshold=0.10, how='over').opts(height=300,width=600)

points3 = hv.Points(ltg_data,kdims=['time','alt'])
datashaded3 = hd.datashade(points3,cmap=cc.fire).redim.range(y=(0,20))
plot3 = hd.dynspread(datashaded3, threshold=0.10, how='over').opts(height=300,width=600)

points4 = hv.Points(ltg_data,kdims=['lon','lat'])
datashaded4 = hd.datashade(points4,cmap=cc.fire).redim.range(y=(-34,34.4),x=(-87.6,-86.8))
plot4 = hd.dynspread(datashaded4, threshold=0.10, how='over').opts(height=300,width=600)

hv.Layout([plot1,plot2,plot3,plot4]).cols(2)
Marc Skov Madsen
@MarcSkovMadsen
HI @gewitterblitz . Try posting the question on Discourse https://discourse.holoviz.org/. Looks like something that the community can help with and it will help build the knowledge base. Thanks.
Milind Sharma
@gewitterblitz
Thanks, @MarcSkovMadsen . I wasn't sure what's the right platform for asking questions.
James A. Bednar
@jbednar
See the top line of this page. :-)
Milind Sharma
@gewitterblitz
whoops!
James A. Bednar
@jbednar
:-)
Milind Sharma
@gewitterblitz
Thanks, @jbednar . I posted my question over there.
ea42gh
@ea42gh
Anybody else see a problem with the ytick powers:
hv.Curve(np.linspace(.1,10,20)).opts(width=420, logy=True)
kitaev-chen
@kitaev-chen
Hi
Does anyone know how to draw the line & labels across multiple indices?
I can draw this bars
but I don't know how to draw the lines and labels, since the indices are multiple columns
kitaev-chen
@kitaev-chen
demo = {
    'ds': [3,3,3,3,4,4,4,4], 
    'case': ['A', 'A', 'B', 'B', 'A', 'A', 'B', 'B'],
    'model': ['model1', 'model2', 'model1', 'model2', 'model1', 'model2', 'model1', 'model2'],
    'acc': [0.1, 0.5, 0.3, 0.7, 0.2, 0.4, 0.6, 0.8],
    'acc_pct': ['10.00%', '50.00%', '30.00%', '70.00%', '20.00%', '40.00%', '60.00%', '80.00%'],
    'acc_std': [0.01, 0.02, 0.03, 0.04, 0.02, 0.01, 0.04, 0.03]
    }

pd.DataFrame(demo)
image.png
demo=pd.DataFrame(demo)
demo_plot = demo.query('ds==3').filter(items=['case', 'model', 'acc'])
print(demo_plot)

# ------ bars ------ #
bars = hv.Bars(demo_plot, ['case', 'model'], ['acc']).opts(opts.Bars(
    ylim=(0, 1),
    bar_width=0.5, 
    # border=0,
    cmap='Category20',
    # show_frame=False,
    ))
bars
how about Labels & lines?
James A. Bednar
@jbednar
I'm not sure that Lines can be drawn on categorical axes like those.
(In Bokeh.)
kitaev-chen
@kitaev-chen
em... perhaps I should try a grid way...
ea42gh
@ea42gh

I was looking at http://holoviews.org/Examples/index.html
Does philippjfr.com still exist? The link appears dead.

Specifically, I was looking for the Julia notebook.
I'd really like to be able to combine holoviews plots and Julia functions....

James A. Bednar
@jbednar
@philippjfr ?
Marc Skov Madsen
@MarcSkovMadsen
Seems like I'm not the only one experimenting with Pyodide and data apps https://www.linkedin.com/posts/itaydafna_dashboards-plotly-dataviz-activity-6792319260841656320-Ad4S
Seems like I am not the only experimenting with Transcrypt and data apps https://community.plotly.com/t/transcrypt-transpiling-python-to-javascript/44106/8
brianmingus2
@brianmingus2
What is the benfit of Lumen vs Panel?
Lumen is more declarative? Does it really scale out to billions of rows?
ie will it appropriately pull in Datashader for the raster, and is it configurable enough, or is it too early.
Marc Skov Madsen
@MarcSkovMadsen
I'm trying to get Panel running on containDS which is a way for users of a Jupyter Hub to launch and share apps across frameworks. In the list of frameworks I can select Bokeh and I can get it working for a very simple hello world Panel app.
I.e. using the command app = build_single_handler_application(app_py_path, ['/']).
Will that work for Panel apps in general? Or is Panel doing something today so that it is no longer compatible with Bokeh?
James A. Bednar
@jbednar
I think that's a question for @philippjfr .
Hugues-Yanis Amanieu
@hyamanieu
Thanks for porting panel's doc to Sphinx, it got easier for me to move around the site!

Seems like I am not the only experimenting with Transcrypt and data apps https://community.plotly.com/t/transcrypt-transpiling-python-to-javascript/44106/8

Isn't it when you should be coming in and mention panel :p ?

Marc Skov Madsen
@MarcSkovMadsen
Maybe @hyamanieu. I don't know if is good behavior to open source colleagues. For example on Streamlit github and discourse there are lots of requests from users that are easy to solve with Panel. But going in and commenting you can try out Panel is probably not good behavior.
EfforiaKnight
@EfforiaKnight
Hi,
I'm new to Panel, started one week ago, and I can't figure out why the app (python file) I build and serve using panel serve app.py is mirrored when multi users connecting.
I'm using the last version 0.12.0a8
Is there anything that need to be configured in order to deploy server for multi user?
I'm using Windows and Pycharm.
Thanks in advance.
James A. Bednar
@jbednar
I'm not sure what you mean, but this sounds like a discussion topic for the discourse (see first line on this page).
Hugues-Yanis Amanieu
@hyamanieu

Maybe @hyamanieu. I don't know if is good behavior to open source colleagues. For example on Streamlit github and discourse there are lots of requests from users that are easy to solve with Panel. But going in and commenting you can try out Panel is probably not good behavior.

yes of course, I was just teasing.

Marc Skov Madsen
@MarcSkovMadsen
@EfforiaKnight . Please ask you question on Discourse. In order for the community to be able to help you would need to include a minimum, reproducible example. If you can do that my bet is your problem is pretty easy to solve.
EfforiaKnight
@EfforiaKnight
@MarcSkovMadsen . Thanks. After couple of hours of trial and error I found my mistake in the code.
Taras Novak
@RandomFractals
Hi guys and gals. Impressed by the stack of demos @MarcSkovMadsen built with panel, streamlit, and some data analytics packages I enjoy working with too, I've decided to join your discourse to learn more about the cool things we can do with your stack. My handle there: https://discourse.holoviz.org/u/datapixy You might know me as 'dataPixy' on twitter or in vscode dev community slack workspace where I've been an active contributor for some time. Some might even use my data viz extensions in vscode: https://marketplace.visualstudio.com/publishers/RandomFractalsInc I am looking to dive deeper into data analytics and maybe draft some cool demos with your libs in the months to come. Just wanted to say Hi for now.
Marc Skov Madsen
@MarcSkovMadsen
Great to hear @EfforiaKnight . See you on Discourse :-)
Hi @RandomFractals . Welcome. I use some of the VS Code extensions and they are great and of high quality. Thanks.
1 reply
I know you are very capable with DataViz in JS also. And a user of ObservableHQ. I have sometimes been thinking that when you have all that skill in JS what draws you to Python still? And when would you use Python over JS?
6 replies
Hugues-Yanis Amanieu
@hyamanieu
I've had the feeling developers don't really like javascript as a programming language, I believe this is also why typescript was created. I would be interested to hear as well his response.
1 reply
ea42gh
@ea42gh
hv.Path label parameter has no effect. If this is a bug and somebody points me to what needs to be changed, I'd be happy to submit a PR.
ea42gh
@ea42gh
Found the issue on github/holoviews. My own, at that!! I'll add a discourse entry for future reference
ea42gh
@ea42gh
@philippjfr Are the PathPlots defaults for show_legend in ./plotting/bokeh/path.py and ./plotting/mpl/path.py theonly two changes required? If so, do you want a PR?
Marc Skov Madsen
@MarcSkovMadsen
FYI. Philip Vollet was so kind to share panel-highcharts on Linked in and Twitter yesterday. On LinkedIn the video had +6000 views. And today he shared awesome-panel.org on LinkedIn and Twitter. I've tried to add information about and links to Panel.
James A. Bednar
@jbednar
Nice.