Figure linewidth default in rcparams

[NGJROMO]

Hello,

I am using:
Matplotlib 1.5.1
Python 3.5
Windows 7 Professional 64-bit (6.1, Build 7601)

And I installed Python and Matplotlib newly through Anaconda.

I used to set my default plot settings with the following code:

import matplotlib as mpl
mpl.rcParams['savefig.bbox']='standard'
mpl.rcParams['axes.linewidth']=1
mpl.rcParams['axes.grid']=True
mpl.rcParams['grid.linestyle']='-'
mpl.rcParams['grid.linewidth']=1
mpl.rcParams["savefig.facecolor"]='white'
mpl.rcParams['savefig.edgecolor']='black'
mpl.rcParams['figure.edgecolor']='black'    
mpl.rcParams["figure.facecolor"]='white'
mpl.rcParams["axes.facecolor"]='white'
mpl.rcParams['axes.edgecolor']='black'

Which gave me a plot with a black grid, a white background and a black line around the plot. I initialized my figures with plt.figure():

After updating to Matplotlib 1.5.1 I run the same code, but the black box around the figure is gone:

I can manually fix the issue by initializing my figures with plt.figure(linewidth=1) (Which is how I produced above plots), but I would like to set the default figure linewidth in rcparams. To my knowledge this is currently not possible. e.g. an entry such as:

mpl.rcParams["figure.linewidth"]

would be perfect. It would be much appreciated if I could get some support on this topic.

I don't remember the exact version of Python and Matplotlib in which this used to work, is there a way to check my update history in Anaconda?

Kind regards,

Nick

[tacaswell]

Is the varying thickness of the box around the figure also a feature or a bug?

[NGJROMO]

That looks like it can be solved by using:

plt.figure(linewidth=1.5)

instead of

plt.figure(linewidth=1)

To initialize the figure:

[tacaswell]

What was the last version of mpl that those rcparams worked with?

Hopefully this level of customization will be enabled by traitlets.

[NGJROMO]

Let's see...

I initially downloaded matplotlib in august 2015 through the default anaconda installer and updated to 1.5.1 just before posting this issue (March 1st 2016).

I don't have a version timeline available but I suppose the version i downloaded in august 2015 would be 1.4.x ? Please verify

Kind regards,

Nick

[WeatherGod]

Yes, most likely, it would have been version 1.4.2 if it was August 2015.

On Tue, May 17, 2016 at 2:24 AM, NGJROMO notifications@github.com wrote:

Let's see...

I initially downloaded matplotlib in august 2015 through the default
anaconda installer and updated to 1.5.1 just before posting this issue
(March 1st 2016).

I don't have a version timeline available but I suppose the version i
downloaded in august 2015 would be 1.4.x ? Please verify

Kind regards,

Nick

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub
#6083 (comment)

[tacaswell]

Can you try down-grading mpl via anaconda (conda install matplotlib=1.4.2 etc) until you get the old behavior? This will greatly help us understand what caused this change and hence how to fix it.

[NGJROMO]

Hi,

I can confirm that downgrading to 1.4.3 does not resolve my problem. Attempting a downgrade to 1.4.2 gives me the following:

The following specifications were found to be in conflict:
  - conda -> conda-env
  - conda -> menuinst
  - conda -> pycosat
  - conda -> python 2.7*|3.3*|3.4*
  - conda -> pyyaml
  - conda -> requests
  - conda-env (target=conda-env-2.4.5-py35_0.tar.bz2) -> python 2.7*|3.3*|3.4*|3.5*
  - cycler (target=cycler-0.10.0-py35_0.tar.bz2) -> python 2.7*|3.4*|3.5*
  - cycler (target=cycler-0.10.0-py35_0.tar.bz2) -> six
  - matplotlib 1.4.2* -> numpy 1.9*
  - matplotlib 1.4.2* -> pyparsing 2.0.1
  - matplotlib 1.4.2* -> pyqt
  - matplotlib 1.4.2* -> python 2.6*|2.7*|3.3*|3.4*
  - matplotlib 1.4.2* -> pytz
  - menuinst (target=menuinst-1.3.2-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - mkl-service (target=mkl-service-1.1.2-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - nose (target=nose-1.3.7-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - numexpr (target=numexpr-2.5-np110py35_0.tar.bz2) -> numpy 1.10*|1.11*|1.6*|1.7*|1.8*|1.9*
  - numexpr (target=numexpr-2.5-np110py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - numpy (target=numpy-1.10.4-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*
  - pip (target=pip-8.1.1-py35_1.tar.bz2) -> python 2.7*|3.3*|3.4*
  - pycosat (target=pycosat-0.6.1-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - pyparsing (target=pyparsing-2.0.3-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*
  - pyqt (target=pyqt-4.11.4-py35_5.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*
  - pyqt (target=pyqt-4.11.4-py35_5.tar.bz2) -> qt >=4.8.6|>=4.8.7
  - pyqt (target=pyqt-4.11.4-py35_5.tar.bz2) -> sip >=4.16.4
  - python 3.5*
  - python-dateutil (target=python-dateutil-2.5.1-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - python-dateutil (target=python-dateutil-2.5.1-py35_0.tar.bz2) -> six
  - pytz (target=pytz-2016.2-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - pyyaml (target=pyyaml-3.11-py35_3.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*
  - qt (target=qt-4.8.7-vc14_7.tar.bz2) -> @vc10
  - qt (target=qt-4.8.7-vc14_7.tar.bz2) -> @vc14
  - qt (target=qt-4.8.7-vc14_7.tar.bz2) -> @vc9
  - requests (target=requests-2.9.1-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - scikit-learn (target=scikit-learn-0.17.1-np110py35_0.tar.bz2) -> numpy 1.10*|1.11*|1.6*|1.7*|1.8*
  - scikit-learn (target=scikit-learn-0.17.1-np110py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - scipy (target=scipy-0.17.0-np110py35_0.tar.bz2) -> numpy 1.10*|1.11*|1.6*|1.7*|1.8*|1.9*
  - scipy (target=scipy-0.17.0-np110py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - setuptools (target=setuptools-19.2-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - sip (target=sip-4.16.9-py35_2.tar.bz2) -> python 2.7*|3.4*|3.5*
  - six (target=six-1.10.0-py35_0.tar.bz2) -> python 2.6*|2.7*|3.3*|3.4*|3.5*
  - tk (target=tk-8.5.18-vc14_0.tar.bz2) -> @vc10
  - tk (target=tk-8.5.18-vc14_0.tar.bz2) -> @vc14
  - tk (target=tk-8.5.18-vc14_0.tar.bz2) -> @vc9
  - wheel (target=wheel-0.29.0-py35_0.tar.bz2) -> python 2.7*|3.3*|3.4*|3.5*
Use "conda info <package>" to see the dependencies for each package.

I've also changed my rcparams code since I first made it, so I am not sure if I will be able to reproduce the problem with a 100% accuracy

[tacaswell]

The easiest thing to do is just create a new enviroment with conda create -n testing matplotlib=1.4.2 instead of trying to down-grade inplace. Sorry gave bad advice before.

[NGJROMO]

Hi,

I can confirm that creating this testing environment and running my code again does not give me black lines around my plots.

[tacaswell]

Can you find a version of mpl where your code did work as expected?

[NGJROMO]

I cannot, so I assume it's not matplotlib that causes the problem?

It is now two months ago since I posted this issue, and in the meantime I've upgraded from windows 8 to 10 and reinstalled anaconda again so I assume a lot of things have changed along with that.

[tacaswell]

Without an example of when it worked it is hard for us to do much else to help. It seems you have found a suitable workaround so closing this due to not enough information.

If you can reproduce the old behavior please ping to have this re-opened!

[NGJROMO]

Right, so could you please show me how to set my rcParams to produce a frame around my plots by default?

[tacaswell]

I suggest sending an email to the matplotlib-user@python.org mailing list (you will need to subscribe to post un-moderated). We try to keep the issue tracking for bugs / feature requests and use the mailing list for this sort of user support.

[NGJROMO]

I'm not sure where to subscribe in order to mail to this address,

regardless I still think my request would classify as a feature request. As far as all I am asking is for a

mpl.rcParams["figure.linewidth"] (or similar)

entry to be added to the rcparams