QuantStack/Lobby

Ask any question about QuantStack projects: xtensor, xeus, voila, Jupyter widgets, mamba

Dave Hirschfeld

@dhirschfeld

:+1: :eyes:

Jaime Rodríguez-Guerra

@jaimergp

I am using umamba in one of my systems for three months now and no big issues so far! I keep my environments simple and numerous (one per project, or even tool). The biggest problem is typing micromamba all the time, but nothing that an alias cannot change :D

Dave Hirschfeld

@dhirschfeld

:+1: Thanks - good to have some confirmation that it's not a crazy idea! :D

Bastian Zimmermann

@BastianZim

That's great to hear! Would I be breaking anything if I alias micromamba and mamba? Just for muscle memory.

Wolf Vollprecht

@wolfv

You'd break running mamba and micromamba side by side :)

I've been using micromamba quite a lot locally, as well, and @adriendelsalle does it, too :)

I can recommend it (and it's got nice autocompletion, too, on bash and zsh)

one missing thing is self-updating

Bastian Zimmermann

@BastianZim

Ahh yeah should clarify that I'd only use micromamba :smile:
So just substitution of the underlying software.

Adrien Delsalle

@adriendelsalle

that would work very smoothly. just take care that some additional flags are handled by micromamba, if you alias to mamba it will be hard to rollback ;)

1 reply

Wolf Vollprecht

@wolfv

Also we use the provision-with-micromamba github action quite a lot now, and it's also really nice (@beckermr started that one)

1 reply

Dave Hirschfeld

@dhirschfeld

Great! I'm looking to switch to umamba so don't need to run side-by-side. I'm also not that fussed on complete api compatibility (and would be happy for umamba to break compatibility to make improvements) so I'll give it a go as my primary package manager!

Wolf Vollprecht

@wolfv

yeah the biggest "CLI" breakage is that we don't have the same kind of split wiht the env subcommand

Dave Hirschfeld

@dhirschfeld

IMHO that's always been weirdly dichotomous so that might be an improvement

Wolf Vollprecht

@wolfv

hopefully!

in conda, it's also quite weird that some commands read text files of specs with -f or --file and the env command reads yaml files

so this weirdness is gone

Dave Hirschfeld

@dhirschfeld

As someone who has been using conda since 2013, I think it has evolved enormously, but has also accumulated a lot of cruft that umamba is ideally placed to just drop

Maybe umanba can be conda The Good Parts

(after a decades' experience)

Bastian Zimmermann

@BastianZim

Would it make sense to bring umamba to brew or is it still too unstable? Could maybe make updating easier (At least for Linux and Mac)

Taiguara Tupinambás

@tatatupi

Hi everyone! I am having an issue with xtensor-io usage of highfive. Just figured out that the probem is with the version. I installed it via vcpkg, which points to 0.12.8, a version with broken references to H5Easy::DumpMode.

So I just opened a PR to vcpkg, but not sure if I did it correctly

microsoft/vcpkg#21946

Taiguara Tupinambás

@tatatupi

(maybe I should change my project's package manager to mamba already!)

Thorsten Beier

@DerThorsten

Feel free to checkout my latest blog post =) Jupyter Games: A Python Box2D integration in Jupyter
https://twitter.com/ThorstenBeier/status/1470841293047836678

Francesco Biscani

@bluescarni

@DerThorsten looks great! I have been looking for a while for a 2D physics simulation library with jupyter support, looking forward to try pyb2d.

ndbecker

@nbecker

Does xtensor-python do automatic conversion of python list input argument to xt::pyarray? I have a function defined to accept xt::pyarray<double>const&, and when passed a list[float] from python, it (surprisingly) worked just fine.

ndbecker

@nbecker

Is it safe to assign to views that alias, e.g.: xt::view (hist, xt::range (1, size)) = xt::view (hist, xt::range (0, size-1));

like x[1:] = x[:-1]

Johan Mabille

@JohanMabille

@nbecker yes it's safe, unless you explicitely use noalias, xtensor will create a temporary before assigning to avoid aliasing

ndbecker

@nbecker

So performance wise I'd be better off using std::copy_backward

Johan Mabille

@JohanMabille

in this case yes

Sylvain Corlay

@SylvainCorlay

@/all Let’s look back at what we have accomplished in 2021!

https://twitter.com/QuantStack/status/1472234833451663364

Patrick Bos

@egpbos

nice, i didn't know about voila-gridstack, that looks really useful

Patrick Bos

@egpbos

what's the status of quetz btw? I see there's now a front-end that you can download tarred packages from... is it already possible to use mamba.pm as my main package server?

we've recently been thinking about this (at the NL eScience Center), because of the changes in pricing policy at anaconda, which could impact some of our audience (we mostly target scientists at public institutions, but don't want to exclude scientists at companies just because our distribution channel now costs them)

Patrick Bos

@egpbos

@btel do i understand correctly that it is not yet possible to persist somehow the manually made grid layout in voila-gridstack?

Gayan Chinthaka

@ndgayan

VOILA notebooks giving "WARNING | Notebook Test.ipynb is not trusted." warning. Do you know how to fix this issue?

Robert Seidl

@seidlr

Hi, I have a question related to loading additional javascript in voila. I need to include a js library to draw sparklines in a panel tabulator widget in voila. I managed to get it to run in jupyter lab but it is not working with voila.
The library used is actually a jQuery extension and this seems to make it a bit tricky to load.
Here is a repo with binder links to lab and voila
https://github.com/seidlr/voila-panel-tabulator-sparklines
and the tipps i got from the panel guys: https://discourse.holoviz.org/t/add-sparkline-formatters-to-tabulator/3197/3
If you have an idea how to get this to run, please let me know. Thanks.

Benjamin Bertrand

@beenje

https://twitter.com/QuantStack/status/1472234833451663364

Nice article! Congrats for all the work.
I saw there is a new frontend for Quetz. I guess it's this repo https://github.com/mamba-org/quetz-frontend? Is it still WIP or would you recommend trying to use it?
The article also mentions a plugin to convert PyPI wheels to conda packages on the fly. I'd be interested in that. Where can I find that plugin?

cmartinotti

@cmartinotti

hello guys, I am baffled. I have coded a pretty complex jupyer notebook involving various interconnected widgets and I thought I well understood how ipywidget worked. I might have found a bug though as today I am trying to simply run the following code:

out=widgets.Output()
dfa=pd.DataFrame({'x':[1,2,3],'y':[1,2,3]})


with out:
    out.clear_output(wait=True)
    print('ciao')
    fig=px.scatter(dfa,x='x',y='y', width=800, height=400 )
    fig.show()

out

And the output cell is displaying nothing. I am pretty sure that it should display the plotly express plot in fig. Instead it only display the 'ciao' being printed. The really weird thing is that I am using the same exact kernel with very similar code for my other complex applications involving multiple boxes and observes codes and everything is working fine. I don't understand what's wrong.

martinRenou

@martinRenou

Hey @cmartinotti, I tried your example in JupyterLab, and it seems to work fine:

though plotly seems to fail in my classic Jupyter Notebook setup for some reason

2 replies

just

fig=px.scatter(dfa,x='x',y='y', width=800, height=400 )
fig.show()

shows an empty area

it seems to be due to a 404 error on the plotly JS code

require.js?v=d37b48bb2137faa0ab98157e240c084dd5b1b5e74911723aa1d1f04c928c2a03dedf922d049e4815f7e5a369faa2e6b6a1000aae958b7953b5cc60411154f593:1958 GET http://localhost:8888/static/plotly.js?v=20220106094745 net::ERR_ABORTED 404 (Not Found)

DougRzz

@DougRzz

Hi @cmartinotti . I tried your code and I had the same issue as you in Jupyter Classic. I fixed it with.

    fig.show(renderer='notebook')

Where communities thrive

People

Repo info

Activity