Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jun 27 21:16
    matthewfeickert synchronize #23349
  • Jun 27 21:07
    timhoffm milestoned #23354
  • Jun 27 20:44
    QuLogic commented #23355
  • Jun 27 20:42
    QuLogic opened #23355
  • Jun 27 20:42
    QuLogic labeled #23355
  • Jun 27 20:40

    QuLogic on revert-23095-osx-tk-xfail

    Revert "Try to unbreak CI by xf… (compare)

  • Jun 27 20:35
    QuLogic commented #13131
  • Jun 27 20:15
    tacaswell labeled #23334
  • Jun 27 20:15
    tacaswell milestoned #23334
  • Jun 27 20:15
    tacaswell commented #23334
  • Jun 27 20:14
    QuLogic labeled #23353
  • Jun 27 20:08
    QuLogic labeled #23351
  • Jun 27 20:05
    QuLogic unlabeled #23060
  • Jun 27 18:36
    DWesl synchronize #22999
  • Jun 27 18:33
    DWesl commented #22999
  • Jun 27 16:25
    DWesl synchronize #22999
  • Jun 27 15:03
    oscargus synchronize #23354
  • Jun 27 15:02
    oscargus synchronize #23354
  • Jun 27 13:43
    oscargus labeled #23354
  • Jun 27 13:43
    oscargus opened #23354
Thomas A Caswell
@tacaswell
We should get cpython to only post the source for the first 2 weeks so that the only people who have it also definitely have a compiler !
Ryan May
@dopplershift
Both cynically and fair, I don't think the web people care that much about compiled extensions.
Thomas A Caswell
@tacaswell
we have started using fastAPI at work and there is a bunch of compiled stuff there
but, point taken
Ryan May
@dopplershift
I'm more baffled than annoyed right now. It takes 10 years to even start really getting rid of 2.7, but now we have a sizable population willing to run 3.9 on day 0?
We're just finally getting to the point where we have a reasonable set of conda-forge packages for 3.9.
Thomas A Caswell
@tacaswell
maybe "wait, we are free from the system packages!!!" that kicked in?
Ryan May
@dopplershift
Perhaps
hannah
@story645
also there's probably some sampling bias in folks who want to run on 3.9 and are also comfortable asking the devs for that
Ryan May
@dopplershift
But these aren't even just requests, they're "it no worky" reports.
hannah
@story645
maybe they just don't know it's not supposed to work?
Ryan May
@dopplershift
maybe. Regardless, none of that is criticism towards anyone, I'm just surprised we have that much demand for 3.9.
hannah
@story645
shrugs there are some fun features in newer python ?
Rohit Rawat
@xordux

Hi,
My PR is failed Codecov check however I am unable find why.
Previously Codecov was fine but then All I did was removing 2 assert statements from a Pytest and replaced them with match argument in the pytest.warns call.

matplotlib/matplotlib#18493 and click here to see that exact different

I am new to this Codecov, can anyone please help.

Rohit Rawat
@xordux
Perhaps this happened because I had not synced my forked repo with Matplotlib's repo. So I followed steps as mentioned here, will see if it works.
Grigory Starkov
@Gregstrq
Hi, guys. I am trying to draw a 3d contour over a 3d surface plot. However, the contour lines appear as if they are behind the surface. How do I make them appear in front of the surface?
Jody Klymak
@jklymak
In general Matplotlib is not very good at 3-D rendering of multiple objects; in particular, they do not know anything about one another, so just are drawn, with a somewhat arbitrary 2-D zorder.
Grigory Starkov
@Gregstrq
@jklymak Can you tweak this 2-D zorder?
Jody Klymak
@jklymak
I think zorder works, but I'm not a 3-D expert.
Thomas A Caswell
@tacaswell
I thought we did a best-effort at sorting the artstists, but we do not
Grigory Starkov
@Gregstrq
@jklymak In the case of 3D, are there alternatives to matplolib which provide the same level of control over all the parts of the figure? (and have nice integration with LaTeX)
Thomas A Caswell
@tacaswell
"normal" zorder should control the draw order, but if the artists intersect, then you will only change the escher effocts
Jody Klymak
@jklymak
Ive used mayavi. I don't know what "nice integration with latex" means - certainly you can produce figures that can be included in Latex. If you mean matching fonts, I doubt it, but I've always rolled my eyes at the community that wants their figure fonts to match their text fonts :wink:
Grigory Starkov
@Gregstrq
@jklymak Matching of the fonts is nice, but I would not insist on it. However, can you use math symbols for legends, labels and etc. in mayavi?
In reality, I only need one contour. Is it possible to create color map with an abrupt change in colors which happens at some level?
Like, take a continuos gradient map, cut a piece of it in the middle, and then adjust everything so that the cut corresponds to a particular z level I want?
Jody Klymak
@jklymak
Yes you’d just need to do a bit of surgery on a colormap. There is a tutorial on manipulating colormaps
My memory of mayavi is that font handling is very rudimentary
Thomas A Caswell
@tacaswell
most things that push to openGL have not-great font handling
Grigory Starkov
@Gregstrq
What about PlotLy?
Does any of you have experience with it?
Thomas A Caswell
@tacaswell
apparently paraview (at least at one point) supported formulas in text by calling us to get the rendered buffer
Probably with plotly
Rohit Rawat
@xordux

Hi,
A Question about PR https://github.com/matplotlib/matplotlib/pull/18493/

Azure test is failing for Windows_py37, macOS_py37, macOS_py38 . But I can't understand if there is any relation to the code changes I have done.

I can see that other PR are also failing Azure test(like this merged recently but had all Azure tests failing: matplotlib/matplotlib#18696 )

So, is this like some other issue with master branch or I need to check it in my code?

Thomas A Caswell
@tacaswell
I was just looking at this
it looks like something changed under us in the OSX on azure envirmont and imagemagik is fighting back...
Rohit Rawat
@xordux
Previously (few weeks ago) every check passed smoothly, but today I change assert statement to py.warn and this started failing(I also synced my repo today).
Thomas A Caswell
@tacaswell
between the last CI that passed on azure OSX and the first that failed imagemagick changed from 7.0.10-31 to 7.0.10-34
Rohit Rawat
@xordux
I can't find any release notes of imagemagick in web. Should comparing two releases' code help ?
Thomas A Caswell
@tacaswell
https://github.com/ImageMagick/ImageMagick/releases <- looks like they have a very rapid super-micro release cadence
Rohit Rawat
@xordux
And comparing releases got me more confused :D .. still checking if I get any hint
Jody Klymak
@jklymak
The tk error is more problematic.
Rohit Rawat
@xordux
@jklymak strange thing is that it is passing in Windows_Py38 but not in Windows_Py37

Regarding tk error: Will increasing _test_timeout (which in 10 right now) help ?

Edit: 7th line in test_backend_tk.py
_test_timeout = 10 # Empirically, 1s is not enough on Travis.

Jody Klymak
@jklymak
Yeah, it could just be a transitory failure.
Rohit Rawat
@xordux
So should I test it again or increase timeout to some value?
Jody Klymak
@jklymak
I'd just try running again. I don't understand the test well enough to say. @dopplershift is the resident animation guru
Thomas A Caswell
@tacaswell
Do you use brew locally?