3.0 upgrade

srinivasrao.poladi · October 23, 2018, 2:39pm

Hi,

I was using Matplotlib ver 2.2.2 on windows 10 using anaconda distribution.
Now I have upgraded anaconda distribution with Matplotlib ver 3.0.0. Facing
few issues with the code that was working fine with 2.2.2.

1. In 3D plots using mplot3d, for multiple subplots,
plt.tight_layout(w_pad=5) worked fine with 2.2.2, but same code is giving
the following error
<http://matplotlib.1069221.n5.nabble.com/file/t5188/tight_layout_err.png> ,

and output display is one long vertical line with some jumbled letters.
When I remove titght_layout() statement, I get adequate space in between the
plots, but legend overalps with title. How can I fix this overlap issue?

2. For all 3D plots figure size gets too small, compared to the size I used
to get with 2.2.2 version. It is the same for single plot or multiple
subplots in the figure. Even figsize=(15,10) produces very small figure as
attached here
<http://matplotlib.1069221.n5.nabble.com/file/t5188/3Dfigure_Sizeissue.png>

whereas 2.2.2 produced full screen figure as shown below.
<http://matplotlib.1069221.n5.nabble.com/file/t5188/sizewith222.png>

Appreciate any help in resolving these two issues.

···

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html

jklymak1 · October 23, 2018, 3:45pm

Hi,

I was using Matplotlib ver 2.2.2 on windows 10 using anaconda distribution.
Now I have upgraded anaconda distribution with Matplotlib ver 3.0.0. Facing
few issues with the code that was working fine with 2.2.2.

1. In 3D plots using mplot3d, for multiple subplots,
plt.tight_layout(w_pad=5) worked fine with 2.2.2, but same code is giving
the following error
<http://matplotlib.1069221.n5.nabble.com/file/t5188/tight_layout_err.png> ,

and output display is one long vertical line with some jumbled letters.
When I remove titght_layout() statement, I get adequate space in between the
plots, but legend overalps with title. How can I fix this overlap issue?

This will be better in 3.0.x: FIX: make unused spines invisible by jklymak · Pull Request #12241 · matplotlib/matplotlib · GitHub

Take the legend out of the layout: `leg = plt.legend(); leg.set_in_layout(False)` may help, but for 3-d plots there is still the spine issue. Sorry for the bother.

2. For all 3D plots figure size gets too small, compared to the size I used
to get with 2.2.2 version. It is the same for single plot or multiple
subplots in the figure. Even figsize=(15,10) produces very small figure as
attached here
<http://matplotlib.1069221.n5.nabble.com/file/t5188/3Dfigure_Sizeissue.png>

whereas 2.2.2 produced full screen figure as shown below.
<http://matplotlib.1069221.n5.nabble.com/file/t5188/sizewith222.png>

Appreciate any help in resolving these two issues.

These look like dpi issues. Not sure what is causing that. How are you producing the pngs?

Cheers, Jody

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20181023/9ea92d7e/attachment.html>

···

On 23 Oct 2018, at 07:39, Srinivasa Rao <srinivasrao.poladi at gmail.com> wrote:

srinivasrao.poladi · October 23, 2018, 4:15pm

Thanks a lot, Judy for quick response.

1. Tried your solution 'leg = plt.legend(); leg.set_in_layout(False)`,
unfortunately that has not helped

2. I am running the code in Jupyter Notebook, in the output cell I see
figure size much smaller with 3.0.0 compared to 2.2.2. For this post, I
saved these output figures from both 3.0.0 and 2.2.2 using mouse right click
and save, which by default saves in png format. Hope this helps

···

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html

jklymak1 · October 23, 2018, 4:19pm

OK, this is all likely FIX: make unused spines invisible by jklymak · Pull Request #12241 · matplotlib/matplotlib · GitHub Either download and install from master or revert to 2.2.3 until 3.0.1 is out

Cheers, Jody

···

On 23 Oct 2018, at 09:15, Srinivasa Rao <srinivasrao.poladi at gmail.com> wrote:

1. Tried your solution 'leg = plt.legend(); leg.set_in_layout(False)`,
unfortunately that has not helped

2. I am running the code in Jupyter Notebook, in the output cell I see
figure size much smaller with 3.0.0 compared to 2.2.2. For this post, I
saved these output figures from both 3.0.0 and 2.2.2 using mouse right click
and save, which by default saves in png format. Hope this helps

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

--
Jody Klymak

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20181023/b03e8ec7/attachment.html>

srinivasrao.poladi · October 24, 2018, 3:05am

Thanks, Jody

I am not customizing spines or any other figure/axes/axis attributes. I am
plotting 3D data, for various combination of elevation and rotation
parameters, and adding title and legend for each of the axes(each
combination of elevation and rotation). Idea is to view the data from
different directions.

Regarding the figure size issue also, I am not specifying and dpi parameter,
but using default setup.

Encountered few issues with axes_grid1 toolkit functionality also with
version 3.0.0 compared to 2.2.2. Posted them in developer group.

I just want to make sure that I am not making any mistake at my end, and if
they are real issues with version 3.0.0, let development team know about
them so that they can fix them at some point.

Thanks once again for your attention and prompt response.

Regards
PSR

···

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html

jklymak1 · October 24, 2018, 3:07am

This has all been fixed in FIX: make unused spines invisible by jklymak · Pull Request #12241 · matplotlib/matplotlib · GitHub. It was definitely a bug in 3.0.0, and will be fixed in 3.0.1 when it is released.

Thanks, Jody

···

On Oct 23, 2018, at 20:05 PM, Srinivasa Rao <srinivasrao.poladi at gmail.com> wrote:

Thanks, Jody

I am not customizing spines or any other figure/axes/axis attributes. I am
plotting 3D data, for various combination of elevation and rotation
parameters, and adding title and legend for each of the axes(each
combination of elevation and rotation). Idea is to view the data from
different directions.

Regarding the figure size issue also, I am not specifying and dpi parameter,
but using default setup.

Encountered few issues with axes_grid1 toolkit functionality also with
version 3.0.0 compared to 2.2.2. Posted them in developer group.

I just want to make sure that I am not making any mistake at my end, and if
they are real issues with version 3.0.0, let development team know about
them so that they can fix them at some point.

Thanks once again for your attention and prompt response.

Regards
PSR

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20181023/67d9526e/attachment.html>

srinivasrao.poladi · October 24, 2018, 3:19am

I am sorry, I am unable to relate my problems with the issue mentioned in
this thread!

Have you got a chance to look at my other post in developer group?

···

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html

jklymak1 · October 24, 2018, 3:47am

Did you click through to the issue report?

github.com/matplotlib/matplotlib

3d axes are collapsed by tight_layout

opened 08:14PM - 23 Sep 18 UTC

closed 11:58PM - 13 Oct 18 UTC

adeak

topic: mplot3d

Calling `fig.tight_layout()` on a figure with a 3d `Axes` inside will collapse t…he axes into a vertical line. Minimal example: ``` import matplotlib.pyplot as plt from mpl_toolkits.mplot3d import Axes3D for tighten in False,True: fig = plt.figure() ax = fig.add_subplot(111, projection='3d') ax.plot([1,2], [1,2], [1,2]) if tighten: fig.tight_layout() plt.show() ``` The following figures are produced with python 3.7, matplotlib 3.0.0 (also master): ![figure without tight_layout: all's fine](https://user-images.githubusercontent.com/17914410/45932469-be0a5c00-bf7c-11e8-9988-1bf55e10ced5.png?s=50) ![figure with tight_layout: axes is a single vertical line](https://user-images.githubusercontent.com/17914410/45932470-c19de300-bf7c-11e8-8368-52e12d0c48cc.png) There's also a warning that says `/home/user/matplotlib-env/lib/python3.7/site-packages/matplotlib/tight_layout.py:177: UserWarning: The left and right margins cannot be made large enough to accommodate all axes decorations.` but that's hardly surprising since there's no room for any decorations whatsoever. May or may not be related to the recent-ish issues fixed by https://github.com/matplotlib/matplotlib/pull/11739 and https://github.com/matplotlib/matplotlib/pull/11627.

Cheers, Jody

···

On Oct 23, 2018, at 20:19 PM, Srinivasa Rao <srinivasrao.poladi at gmail.com> wrote:

I am sorry, I am unable to relate my problems with the issue mentioned in
this thread!

Have you got a chance to look at my other post in developer group?

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20181023/4f68e442/attachment-0001.html>

srinivasrao.poladi · October 24, 2018, 5:44am

I am sorry, earlier I clicked thru 12241 thread, which was not clear. This
12239 thread is clear to me now.

Any thoughts on other issues that I posted to developer forum(should I also
post them on user forum as well?) and figure size issue in 3D plots?

Thanks
PSR

···

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html

tacaswell · October 28, 2018, 7:31pm

I don't see a difference in the figure size.

Could you post a minimal example that generates the issue you are seeing?
That makes it much easier for us to help debug the issue!

Tom

···

On Wed, Oct 24, 2018 at 1:44 AM Srinivasa Rao <srinivasrao.poladi at gmail.com> wrote:

I am sorry, earlier I clicked thru 12241 thread, which was not clear. This
12239 thread is clear to me now.

Any thoughts on other issues that I posted to developer forum(should I also
post them on user forum as well?) and figure size issue in 3D plots?

Thanks
PSR

--
Sent from:
http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

--
Thomas Caswell
tcaswell at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20181028/c00b1c25/attachment.html>