Making space for a long legend outside of a barchart

Hi,
  I am having trouble to get space allocated for a long legend text,
lets say spanning 2/3 - 3/4 of the whole output. I would like to have
stacked barchart as 1st subplot and the place of remaining 3 subplots
to be actually allocated by the legend. Alternatively, could I get the
legend saved into a separate figure?

Or could the space for legend text be allocated automatically minimizing
output figure size? For example, the width would be 1120px while height
be multiples of 840px (840 for each subplot)?

  Attached is a quick example. It shows also that I tried tight_layout()
but wasn't successful with this either. I would be glad for some help,
ideally converting the whole thing into an object-oriented approach.
I am generating several figures in a row and would like to clear()/del()
any previously used data ASAP.

Thank you,
Martin
Am using mpl-1.2.2

stacked_barchart_with_long_legend.png

stacked_barchart_with_long_legend.py (1.11 KB)

Try "fig.savefig('foobar.png', bbox_inches='tight')" when saving the
image. It will make the figure size such that all the visible elements of
the figure will fit into the saved output. tight_layout() is meant to make
sure the elements don't overlap each other, but does nothing about making
sure nothing gets clipped.

Cheers!
Ben Root

···

On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs < mmokrejs@...3951...> wrote:

Hi,
  I am having trouble to get space allocated for a long legend text,
lets say spanning 2/3 - 3/4 of the whole output. I would like to have
stacked barchart as 1st subplot and the place of remaining 3 subplots
to be actually allocated by the legend. Alternatively, could I get the
legend saved into a separate figure?

Or could the space for legend text be allocated automatically minimizing
output figure size? For example, the width would be 1120px while height
be multiples of 840px (840 for each subplot)?

  Attached is a quick example. It shows also that I tried tight_layout()
but wasn't successful with this either. I would be glad for some help,
ideally converting the whole thing into an object-oriented approach.
I am generating several figures in a row and would like to clear()/del()
any previously used data ASAP.

Thank you,
Martin
Am using mpl-1.2.2

Hi Ben,

Benjamin Root wrote:

    Hi,
      I am having trouble to get space allocated for a long legend text,
    lets say spanning 2/3 - 3/4 of the whole output. I would like to have
    stacked barchart as 1st subplot and the place of remaining 3 subplots
    to be actually allocated by the legend. Alternatively, could I get the
    legend saved into a separate figure?

    Or could the space for legend text be allocated automatically minimizing
    output figure size? For example, the width would be 1120px while height
    be multiples of 840px (840 for each subplot)?

      Attached is a quick example. It shows also that I tried tight_layout()
    but wasn't successful with this either. I would be glad for some help,
    ideally converting the whole thing into an object-oriented approach.
    I am generating several figures in a row and would like to clear()/del()
    any previously used data ASAP.

    Thank you,
    Martin
    Am using mpl-1.2.2

Try "fig.savefig('foobar.png', bbox_inches='tight')" when saving the
image. It will make the figure size such that all the visible
elements of the figure will fit into the saved output. tight_layout()
is meant to make sure the elements don't overlap each other, but does
nothing about making sure nothing gets clipped.

Ah, would be nice to make this clear in the docs. So far was doing

import pylab
F = pylab.gcf()
F.set_tight_layout(True)

which as you say does not help the way I thought.

Unfortunately, while

fig.savefig('foobar.png', bbox_inches='tight')

helped to get everything into the .png file (attached), the barchart itself
should span according to the code I posted just 1/2 of the figure. But somehow
it is enlarged and rescaled so that it occupies *more than* 1/2 of the figure.
What in pylab is resizing my image? Note: the final image is 625x1075.

Martin

foobar.png

···

On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs <mmokrejs@…3951… <mailto:mmokrejs@…3951…>> wrote:

Dear Martin,

I worked out a similar example for your reference as I don’t catch your example very well.

fig = plt.figure()
ax1 = fig.add_subplot(211)
ax2 = fig.add_subplot(212)
arrlist = [np.random.normal(size=100) for i in range(50)]
ret = ax1.hist(arrlist,histtype=‘barstacked’)
reclist = [patchlist[0] for patchlist in ret[2]]
labellist = [‘data’+str(i) for i in range(50)]
ax2.legend(reclist,labellist,loc=‘upper left’,bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode=‘expand’)
ax2.set_frame_on(False)
ax2.tick_params(bottom=‘off’,left=‘off’,right=‘off’,top=‘off’)
plt.setp(ax2.get_yticklabels(),visible=False)
plt.setp(ax2.get_xticklabels(),visible=False)

you’re asking some object-oriented way, I personally don’t think using pylab and set_tight_layout are the good way
to be “object-oriented” as pylab is only a bounding wrapper by my understanding (maybe I am wrong!). legend and
hist are all matplotlib.axes.Axes method.

Also, I think it’s unrealistic to ask the figure do a nice job for you if there are 50 legned handlers and you want to show
them in 2 columns with a very high width/height ratio of the figure…

hope it could be of a bit help,

cheers,

Chao

···

On Mon, May 20, 2013 at 6:43 PM, Martin Mokrejs [via matplotlib] <[hidden email]> wrote:

Hi Ben,

Benjamin Root wrote:

On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs <[hidden email] <mailto:[hidden email]>> wrote:

Hi,
  I am having trouble to get space allocated for a long legend text,
lets say spanning 2/3 - 3/4 of the whole output. I would like to have
stacked barchart as 1st subplot and the place of remaining 3 subplots
to be actually allocated by the legend. Alternatively, could I get the
legend saved into a separate figure?
Or could the space for legend text be allocated automatically minimizing
output figure size? For example, the width would be 1120px while height
be multiples of 840px (840 for each subplot)?
  Attached is a quick example. It shows also that I tried tight_layout()
but wasn't successful with this either. I would be glad for some help,
ideally converting the whole thing into an object-oriented approach.
I am generating several figures in a row and would like to clear()/del()
any previously used data ASAP.
Thank you,
Martin
Am using mpl-1.2.2

Try “fig.savefig(‘foobar.png’, bbox_inches=‘tight’)” when saving the

image. It will make the figure size such that all the visible

elements of the figure will fit into the saved output. tight_layout()

is meant to make sure the elements don’t overlap each other, but does

nothing about making sure nothing gets clipped.

Ah, would be nice to make this clear in the docs. So far was doing

import pylab

F = pylab.gcf()

F.set_tight_layout(True)

which as you say does not help the way I thought.

Unfortunately, while

fig.savefig(‘foobar.png’, bbox_inches=‘tight’)

helped to get everything into the .png file (attached), the barchart itself

should span according to the code I posted just 1/2 of the figure. But somehow

it is enlarged and rescaled so that it occupies more than 1/2 of the figure.

What in pylab is resizing my image? Note: the final image is 625x1075.

Martin


AlienVault Unified Security Management (USM) platform delivers complete

security visibility with the essential security capabilities. Easily and

efficiently configure, manage, and operate all of your security controls

from a single console and one unified framework. Download a free trial.

http://p.sf.net/sfu/alienvault_d2d


Matplotlib-users mailing list

[hidden email]

https://lists.sourceforge.net/lists/listinfo/matplotlib-users

foobar.png (132K) Download Attachment


If you reply to this email, your message will be added to the discussion below:

http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41090.html

To start a new topic under matplotlib - users, email [hidden email]

  To unsubscribe from matplotlib, click here.


  [NAML](http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml)


Chao YUE
Laboratoire des Sciences du Climat et de l’Environnement (LSCE-IPSL)
UMR 1572 CEA-CNRS-UVSQ
Batiment 712 - Pe 119
91191 GIF Sur YVETTE Cedex

Tel: (33) 01 69 08 29 02; Fax:01.69.08.77.16



View this message in context: Re: Making space for a long legend outside of a barchart

Sent from the matplotlib - users mailing list archive at Nabble.com.

Hi Chao,

ChaoYue wrote:

Dear Martin,

I worked out a similar example for your reference as I don't catch your example very well.

I think you got the idea quite well.

fig = plt.figure()
ax1 = fig.add_subplot(211)
ax2 = fig.add_subplot(212)
arrlist = [np.random.normal(size=100) for i in range(50)]
ret = ax1.hist(arrlist,histtype='barstacked')
reclist = [patchlist[0] for patchlist in ret[2]]
labellist = ['data'+str(i) for i in range(50)]
ax2.legend(reclist,labellist,loc='upper left',bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode='expand')
ax2.set_frame_on(False)
ax2.tick_params(bottom='off',left='off',right='off',top='off')
plt.setp(ax2.get_yticklabels(),visible=False)
plt.setp(ax2.get_xticklabels(),visible=False)

I added plt.show() and it demonstrates my problem: the legend is not complete in the
figure. That is why I think I could instead use:

import pylab as plt
import numpy as np

fig = plt.figure()
DefaultSize = tuple(fig.get_size_inches())
fig.set_size_inches(DefaultSize[0], 4*DefaultSize[1])
ax1 = fig.add_subplot(411)
ax2 = fig.add_subplot(412)
arrlist = [np.random.normal(size=100) for i in range(50)]
ret = ax1.hist(arrlist,histtype='barstacked')
reclist = [patchlist[0] for patchlist in ret[2]]
labellist = ['data'+str(i) for i in range(50)]
ax2.legend(reclist,labellist,loc='upper left',bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode='expand')
ax2.set_frame_on(False)
ax2.tick_params(bottom='off',left='off',right='off',top='off')
plt.setp(ax2.get_yticklabels(),visible=False)
plt.setp(ax2.get_xticklabels(),visible=False)
plt.show()

But, this does not make the image 4* taller than I thought. But thank you
for the example how to extract the legend of ax1 and place it under ax2.

you're asking some object-oriented way, I personally don't think
using pylab and set_tight_layout are the good way to be
"object-oriented" as pylab is only a bounding wrapper by my
understanding (maybe I am wrong!). legend and hist are all
matplotlib.axes.Axes method.

Also, I think it's unrealistic to ask the figure do a nice job for
you if there are 50 legend handlers and you want to show them in 2
columns with a very high width/height ratio of the figure....

The problem is that the data are calculated dynamically and sometimes
I need to display data for 20 data types while sometimes for 200 data
types (and for each I need a legend).

I did not show that but I do calculate how many columns I could use
legend display and pass that via pylab.legend(..., ncol= ). Of course
at the same time I could calculate whether I will need 2 or 3 or 4
subplots on the page (the first will be the barchart itself), the
remaining space will be used by the long legend of subplot(211).
I would hope that matplotlib does not mind that I actually issue any
fig.add_subplot() foe the third or even fourth subplot at all. That
would be just a trick to get more space for the legend. If I can live
with just with subplot(211) and subplot(212)

The fig.savefig('foobar.png', bbox_inches='tight') which Ben mentioned
yesterday is nice but I want it to crop the image only vertically.
An optional argument like:
fig.savefig('foobar.png', bbox_inches='tight', keep_fig_width=True)
would maybe do the job for me.

What I still don't understand what is resizing the image in tight_layout.
It doesn't seem to me that just the unused border space is chopped away.
Fonts look different, ratio between x and y axes lengths seems different.
Certainly not what I want.

hope it could be of a bit help,

Sure, I am still learning to use matplotlib.

Martin

···

cheers,

Chao

On Mon, May 20, 2013 at 6:43 PM, Martin Mokrejs [via matplotlib] <[hidden email] </user/SendEmail.jtp?type=node&node=41102&i=0>> wrote:

    Hi Ben,

    Benjamin Root wrote:

    >
    >
    >
    > On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs <[hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=0> <mailto:[hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=1>>> wrote:
    >
    > Hi,
    > I am having trouble to get space allocated for a long legend text,
    > lets say spanning 2/3 - 3/4 of the whole output. I would like to have
    > stacked barchart as 1st subplot and the place of remaining 3 subplots
    > to be actually allocated by the legend. Alternatively, could I get the
    > legend saved into a separate figure?
    >
    > Or could the space for legend text be allocated automatically minimizing
    > output figure size? For example, the width would be 1120px while height
    > be multiples of 840px (840 for each subplot)?
    >
    > Attached is a quick example. It shows also that I tried tight_layout()
    > but wasn't successful with this either. I would be glad for some help,
    > ideally converting the whole thing into an object-oriented approach.
    > I am generating several figures in a row and would like to clear()/del()
    > any previously used data ASAP.
    >
    >
    > Thank you,
    > Martin
    > Am using mpl-1.2.2
    >
    >
    > Try "fig.savefig('foobar.png', bbox_inches='tight')" when saving the
    > image. It will make the figure size such that all the visible
    > elements of the figure will fit into the saved output. tight_layout()
    > is meant to make sure the elements don't overlap each other, but does
    > nothing about making sure nothing gets clipped.
    Ah, would be nice to make this clear in the docs. So far was doing

    import pylab
    F = pylab.gcf()
    F.set_tight_layout(True)

    which as you say does not help the way I thought.

    Unfortunately, while

    fig.savefig('foobar.png', bbox_inches='tight')

    helped to get everything into the .png file (attached), the barchart itself
    should span according to the code I posted just 1/2 of the figure. But somehow
    it is enlarged and rescaled so that it occupies *more than* 1/2 of the figure.
    What in pylab is resizing my image? Note: the final image is 625x1075.

    Martin

    ------------------------------------------------------------------------------
    AlienVault Unified Security Management (USM) platform delivers complete
    security visibility with the essential security capabilities. Easily and
    efficiently configure, manage, and operate all of your security controls
    from a single console and one unified framework. Download a free trial.
    http://p.sf.net/sfu/alienvault_d2d
    _______________________________________________
    Matplotlib-users mailing list
    [hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=2>
    https://lists.sourceforge.net/lists/listinfo/matplotlib-users

    *foobar.png* (132K) Download Attachment <http://matplotlib.1069221.n5.nabble.com/attachment/41090/0/foobar.png>

    ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
    If you reply to this email, your message will be added to the discussion below:
    http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41090.html
    To start a new topic under matplotlib - users, email [hidden email] </user/SendEmail.jtp?type=node&node=41102&i=1>
    To unsubscribe from matplotlib, click here.
    NAML <http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html!nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers!nabble%3Aemail.naml-instant_emails!nabble%3Aemail.naml-send_instant_email!nabble%3Aemail.naml>

--
***********************************************************************************
Chao YUE
Laboratoire des Sciences du Climat et de l'Environnement (LSCE-IPSL)
UMR 1572 CEA-CNRS-UVSQ
Batiment 712 - Pe 119
91191 GIF Sur YVETTE Cedex
Tel: (33) 01 69 08 29 02; Fax:01.69.08.77.16
************************************************************************************

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
View this message in context: Re: Making space for a long legend outside of a barchart <http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41102.html>
Sent from the matplotlib - users mailing list archive <http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html> at Nabble.com.

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi Martin,

I don’t know tight_layout quite well. Probably you could also split the handlers of the barplot into
and 2 or 3 or 4 parts depending on the number, and then show them in sperate axes?

then you create n+1 subplots for the whole figure?

probably this is quite stupid.

cheers,

Chao

···

On Wed, May 22, 2013 at 1:03 PM, Martin Mokrejs [via matplotlib] <[hidden email]> wrote:

Hi Chao,

ChaoYue wrote:

Dear Martin,

I worked out a similar example for your reference as I don’t catch your example very well.

I think you got the idea quite well.

fig = plt.figure()
ax1 = fig.add_subplot(211)
ax2 = fig.add_subplot(212)
arrlist = [np.random.normal(size=100) for i in range(50)]

ret = ax1.hist(arrlist,histtype=‘barstacked’)

reclist = [patchlist[0] for patchlist in ret[2]]
labellist = [‘data’+str(i) for i in range(50)]
ax2.legend(reclist,labellist,loc=‘upper left’,bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode=‘expand’)

ax2.set_frame_on(False)

ax2.tick_params(bottom=‘off’,left=‘off’,right=‘off’,top=‘off’)
plt.setp(ax2.get_yticklabels(),visible=False)

plt.setp(ax2.get_xticklabels(),visible=False)

I added plt.show() and it demonstrates my problem: the legend is not complete in the

figure. That is why I think I could instead use:

import pylab as plt

import numpy as np

fig = plt.figure()

DefaultSize = tuple(fig.get_size_inches())

fig.set_size_inches(DefaultSize[0], 4*DefaultSize[1])

ax1 = fig.add_subplot(411)

ax2 = fig.add_subplot(412)

arrlist = [np.random.normal(size=100) for i in range(50)]

ret = ax1.hist(arrlist,histtype=‘barstacked’)

reclist = [patchlist[0] for patchlist in ret[2]]

labellist = [‘data’+str(i) for i in range(50)]

ax2.legend(reclist,labellist,loc=‘upper left’,bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode=‘expand’)

ax2.set_frame_on(False)

ax2.tick_params(bottom=‘off’,left=‘off’,right=‘off’,top=‘off’)

plt.setp(ax2.get_yticklabels(),visible=False)

plt.setp(ax2.get_xticklabels(),visible=False)

plt.show()

But, this does not make the image 4* taller than I thought. But thank you

for the example how to extract the legend of ax1 and place it under ax2.

you’re asking some object-oriented way, I personally don’t think

using pylab and set_tight_layout are the good way to be

“object-oriented” as pylab is only a bounding wrapper by my

understanding (maybe I am wrong!). legend and hist are all

matplotlib.axes.Axes method.

Also, I think it’s unrealistic to ask the figure do a nice job for

you if there are 50 legend handlers and you want to show them in 2

columns with a very high width/height ratio of the figure…

The problem is that the data are calculated dynamically and sometimes

I need to display data for 20 data types while sometimes for 200 data

types (and for each I need a legend).

I did not show that but I do calculate how many columns I could use

legend display and pass that via pylab.legend(…, ncol= ). Of course

at the same time I could calculate whether I will need 2 or 3 or 4

subplots on the page (the first will be the barchart itself), the

remaining space will be used by the long legend of subplot(211).

I would hope that matplotlib does not mind that I actually issue any

fig.add_subplot() foe the third or even fourth subplot at all. That

would be just a trick to get more space for the legend. If I can live

with just with subplot(211) and subplot(212)

The fig.savefig(‘foobar.png’, bbox_inches=‘tight’) which Ben mentioned

yesterday is nice but I want it to crop the image only vertically.

An optional argument like:

fig.savefig(‘foobar.png’, bbox_inches=‘tight’, keep_fig_width=True)

would maybe do the job for me.

What I still don’t understand what is resizing the image in tight_layout.

It doesn’t seem to me that just the unused border space is chopped away.

Fonts look different, ratio between x and y axes lengths seems different.

Certainly not what I want.

hope it could be of a bit help,

Sure, I am still learning to use matplotlib.

Martin

cheers,

Chao

On Mon, May 20, 2013 at 6:43 PM, Martin Mokrejs [via matplotlib] <[hidden email] </user/SendEmail.jtp?type=node&node=41102&i=0>> wrote:

Hi Ben,
Benjamin Root wrote:
>
>
>
> On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs <[hidden email] <[http://user/SendEmail.jtp?type=node&node=41090&i=0](http://user/SendEmail.jtp?type=node&node=41090&i=0)> <mailto:[hidden email] <[http://user/SendEmail.jtp?type=node&node=41090&i=1](http://user/SendEmail.jtp?type=node&node=41090&i=1)>>> wrote:
>
>     Hi,
>       I am having trouble to get space allocated for a long legend text,
>     lets say spanning 2/3 - 3/4 of the whole output. I would like to have
>     stacked barchart as 1st subplot and the place of remaining 3 subplots
>     to be actually allocated by the legend. Alternatively, could I get the
>     legend saved into a separate figure?
>
>     Or could the space for legend text be allocated automatically minimizing
>     output figure size? For example, the width would be 1120px while height
>     be multiples of 840px (840 for each subplot)?
>
>       Attached is a quick example. It shows also that I tried tight_layout()
>     but wasn't successful with this either. I would be glad for some help,
>     ideally converting the whole thing into an object-oriented approach.
>     I am generating several figures in a row and would like to clear()/del()
>     any previously used data ASAP.
>
>
>     Thank you,
>     Martin
>     Am using mpl-1.2.2
>
>
> Try "fig.savefig('foobar.png', bbox_inches='tight')" when saving the
> image. It will make the figure size such that all the visible
> elements of the figure will fit into the saved output. tight_layout()
> is meant to make sure the elements don't overlap each other, but does
> nothing about making sure nothing gets clipped.
Ah, would be nice to make this clear in the docs. So far was doing
import pylab
F = pylab.gcf()
F.set_tight_layout(True)
which as you say does not help the way I thought.
Unfortunately, while
fig.savefig('foobar.png', bbox_inches='tight')
helped to get everything into the .png file (attached), the barchart itself
should span according to the code I posted just 1/2 of the figure. But somehow
it is enlarged and rescaled so that it occupies *more than* 1/2 of the figure.
What in pylab is resizing my image? Note: the final image is 625x1075.
Martin
------------------------------------------------------------------------------
AlienVault Unified Security Management (USM) platform delivers complete
security visibility with the essential security capabilities. Easily and
efficiently configure, manage, and operate all of your security controls
from a single console and one unified framework. Download a free trial.
[http://p.sf.net/sfu/alienvault_d2d](http://p.sf.net/sfu/alienvault_d2d)
_______________________________________________
Matplotlib-users mailing list
[hidden email] <[http://user/SendEmail.jtp?type=node&node=41090&i=2](http://user/SendEmail.jtp?type=node&node=41090&i=2)>
[https://lists.sourceforge.net/lists/listinfo/matplotlib-users](https://lists.sourceforge.net/lists/listinfo/matplotlib-users)

*foobar.png* (132K) Download Attachment <[http://matplotlib.1069221.n5.nabble.com/attachment/41090/0/foobar.png](http://matplotlib.1069221.n5.nabble.com/attachment/41090/0/foobar.png)>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
If you reply to this email, your message will be added to the discussion below:
[http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41090.html](http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41090.html)
To start a new topic under matplotlib - users, email [hidden email] </user/SendEmail.jtp?type=node&node=41102&i=1>
To unsubscribe from matplotlib, click here.
NAML <[http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml](http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml)>


Chao YUE

Laboratoire des Sciences du Climat et de l’Environnement (LSCE-IPSL)

UMR 1572 CEA-CNRS-UVSQ

Batiment 712 - Pe 119

91191 GIF Sur YVETTE Cedex

Tel: (33) 01 69 08 29 02; Fax:01.69.08.77.16



View this message in context: Re: Making space for a long legend outside of a barchart <http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41102.html>

Sent from the matplotlib - users mailing list archive <http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html> at Nabble.com.


Try New Relic Now & We’ll Send You this Cool Shirt

New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your

browser, app, & servers with just a few lines of code. Try New Relic

and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may


Matplotlib-users mailing list

[hidden email]

https://lists.sourceforge.net/lists/listinfo/matplotlib-users


Try New Relic Now & We’ll Send You this Cool Shirt

New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your

browser, app, & servers with just a few lines of code. Try New Relic

and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may


Matplotlib-users mailing list

[hidden email]

https://lists.sourceforge.net/lists/listinfo/matplotlib-users


If you reply to this email, your message will be added to the discussion below:

http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41104.html

To start a new topic under matplotlib - users, email [hidden email]

  To unsubscribe from matplotlib, click here.


  [NAML](http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml)


Chao YUE
Laboratoire des Sciences du Climat et de l’Environnement (LSCE-IPSL)
UMR 1572 CEA-CNRS-UVSQ
Batiment 712 - Pe 119
91191 GIF Sur YVETTE Cedex

Tel: (33) 01 69 08 29 02; Fax:01.69.08.77.16



View this message in context: Re: Making space for a long legend outside of a barchart

Sent from the matplotlib - users mailing list archive at Nabble.com.

ChaoYue wrote:

Hi Martin,

I don't know tight_layout quite well. Probably you could also split the handlers of the barplot into
and 2 or 3 or 4 parts depending on the number, and then show them in sperate axes?

then you create n+1 subplots for the whole figure?

No, the reason why I use stacked bar chart is that I can squeeze the data
into a single figure. It is only difficult for the reader to compare 20
separate barcharts between each other whereas seeing 20 stacked bars in a
single chart on top of each other is easy.

I think your approach combined with something like a future improvement:
fig.savefig('foobar.png', bbox_inches='tight', keep_fig_width=True)
will do the best for me (unused space below the legend will be chopped away
while figure width will be untouched).

Lets see what Ben or others say about the tight_layout rescaling issue and the
"feature request". :wink:
Maritn

···

probably this is quite stupid.

cheers,

Chao

On Wed, May 22, 2013 at 1:03 PM, Martin Mokrejs [via matplotlib] <[hidden email] </user/SendEmail.jtp?type=node&node=41105&i=0>> wrote:

    Hi Chao,

    ChaoYue wrote:
    > Dear Martin,
    >
    > I worked out a similar example for your reference as I don't catch your example very well.

    I think you got the idea quite well.

    >
    > fig = plt.figure()
    > ax1 = fig.add_subplot(211)
    > ax2 = fig.add_subplot(212)
    > arrlist = [np.random.normal(size=100) for i in range(50)]
    > ret = ax1.hist(arrlist,histtype='barstacked')
    > reclist = [patchlist[0] for patchlist in ret[2]]
    > labellist = ['data'+str(i) for i in range(50)]
    > ax2.legend(reclist,labellist,loc='upper left',bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode='expand')
    > ax2.set_frame_on(False)
    > ax2.tick_params(bottom='off',left='off',right='off',top='off')
    > plt.setp(ax2.get_yticklabels(),visible=False)
    > plt.setp(ax2.get_xticklabels(),visible=False)
    >
    I added plt.show() and it demonstrates my problem: the legend is not complete in the
    figure. That is why I think I could instead use:

    import pylab as plt
    import numpy as np

    fig = plt.figure()
    DefaultSize = tuple(fig.get_size_inches())
    fig.set_size_inches(DefaultSize[0], 4*DefaultSize[1])
    ax1 = fig.add_subplot(411)
    ax2 = fig.add_subplot(412)
    arrlist = [np.random.normal(size=100) for i in range(50)]
    ret = ax1.hist(arrlist,histtype='barstacked')
    reclist = [patchlist[0] for patchlist in ret[2]]
    labellist = ['data'+str(i) for i in range(50)]
    ax2.legend(reclist,labellist,loc='upper left',bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode='expand')
    ax2.set_frame_on(False)
    ax2.tick_params(bottom='off',left='off',right='off',top='off')
    plt.setp(ax2.get_yticklabels(),visible=False)
    plt.setp(ax2.get_xticklabels(),visible=False)
    plt.show()

    But, this does not make the image 4* taller than I thought. But thank you
    for the example how to extract the legend of ax1 and place it under ax2.

    >
    > you're asking some object-oriented way, I personally don't think
    > using pylab and set_tight_layout are the good way to be
    > "object-oriented" as pylab is only a bounding wrapper by my
    > understanding (maybe I am wrong!). legend and hist are all
    > matplotlib.axes.Axes method.
    >
    > Also, I think it's unrealistic to ask the figure do a nice job for
    > you if there are 50 legend handlers and you want to show them in 2
    > columns with a very high width/height ratio of the figure....

    The problem is that the data are calculated dynamically and sometimes
    I need to display data for 20 data types while sometimes for 200 data
    types (and for each I need a legend).

    I did not show that but I do calculate how many columns I could use
    legend display and pass that via pylab.legend(..., ncol= ). Of course
    at the same time I could calculate whether I will need 2 or 3 or 4
    subplots on the page (the first will be the barchart itself), the
    remaining space will be used by the long legend of subplot(211).
    I would hope that matplotlib does not mind that I actually issue any
    fig.add_subplot() foe the third or even fourth subplot at all. That
    would be just a trick to get more space for the legend. If I can live
    with just with subplot(211) and subplot(212)

    The fig.savefig('foobar.png', bbox_inches='tight') which Ben mentioned
    yesterday is nice but I want it to crop the image only vertically.
    An optional argument like:
    fig.savefig('foobar.png', bbox_inches='tight', keep_fig_width=True)
    would maybe do the job for me.

    What I still don't understand what is resizing the image in tight_layout.
    It doesn't seem to me that just the unused border space is chopped away.
    Fonts look different, ratio between x and y axes lengths seems different.
    Certainly not what I want.

    > hope it could be of a bit help,

    Sure, I am still learning to use matplotlib.

    Martin

    >
    > cheers,
    >
    > Chao
    >

    >
    > On Mon, May 20, 2013 at 6:43 PM, Martin Mokrejs [via matplotlib] <[hidden email] </user/SendEmail.jtp?type=node&node=41102&i=0>> wrote:
    >
    > Hi Ben,
    >
    > Benjamin Root wrote:
    >
    > >
    > >
    > >
    > > On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs <[hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=0> <mailto:[hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=1>>> wrote:

    > >
    > > Hi,
    > > I am having trouble to get space allocated for a long legend text,
    > > lets say spanning 2/3 - 3/4 of the whole output. I would like to have
    > > stacked barchart as 1st subplot and the place of remaining 3 subplots
    > > to be actually allocated by the legend. Alternatively, could I get the
    > > legend saved into a separate figure?
    > >
    > > Or could the space for legend text be allocated automatically minimizing
    > > output figure size? For example, the width would be 1120px while height
    > > be multiples of 840px (840 for each subplot)?
    > >
    > > Attached is a quick example. It shows also that I tried tight_layout()
    > > but wasn't successful with this either. I would be glad for some help,
    > > ideally converting the whole thing into an object-oriented approach.
    > > I am generating several figures in a row and would like to clear()/del()
    > > any previously used data ASAP.
    > >
    > >
    > > Thank you,
    > > Martin
    > > Am using mpl-1.2.2
    > >
    > >
    > > Try "fig.savefig('foobar.png', bbox_inches='tight')" when saving the
    > > image. It will make the figure size such that all the visible
    > > elements of the figure will fit into the saved output. tight_layout()
    > > is meant to make sure the elements don't overlap each other, but does
    > > nothing about making sure nothing gets clipped.
    > Ah, would be nice to make this clear in the docs. So far was doing
    >
    >
    > import pylab
    > F = pylab.gcf()
    > F.set_tight_layout(True)
    >
    > which as you say does not help the way I thought.
    >
    >
    > Unfortunately, while
    >
    > fig.savefig('foobar.png', bbox_inches='tight')
    >
    > helped to get everything into the .png file (attached), the barchart itself
    > should span according to the code I posted just 1/2 of the figure. But somehow
    > it is enlarged and rescaled so that it occupies *more than* 1/2 of the figure.
    > What in pylab is resizing my image? Note: the final image is 625x1075.
    >
    > Martin

Hi Chao,
  I spent some time to figure out why I cannot replace ax1.hist() with ax1.scatter().
It seems hist() returns list of 'Rectangle' (sadly if there is just one, it does return
just the 'Rectangle' (not wrapped in a list) ... somewhere a trick

a = [a, ]

is likely needed.

Anyway, my problem is that scatter() returns 'PathCollection' object, whatever that is.
How can I grab handles to individual legend items to move them under ax2 like in your
hist-plot example?

Thank you for your help,
Martin

ChaoYue wrote:

···

Dear Martin,

I worked out a similar example for your reference as I don't catch your example very well.

fig = plt.figure()
ax1 = fig.add_subplot(211)
ax2 = fig.add_subplot(212)
arrlist = [np.random.normal(size=100) for i in range(50)]
ret = ax1.hist(arrlist,histtype='barstacked')
reclist = [patchlist[0] for patchlist in ret[2]]
labellist = ['data'+str(i) for i in range(50)]
ax2.legend(reclist,labellist,loc='upper left',bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode='expand')
ax2.set_frame_on(False)
ax2.tick_params(bottom='off',left='off',right='off',top='off')
plt.setp(ax2.get_yticklabels(),visible=False)
plt.setp(ax2.get_xticklabels(),visible=False)

you're asking some object-oriented way, I personally don't think using pylab and set_tight_layout are the good way
to be "object-oriented" as pylab is only a bounding wrapper by my understanding (maybe I am wrong!). legend and
hist are all matplotlib.axes.Axes method.

Also, I think it's unrealistic to ask the figure do a nice job for you if there are 50 legned handlers and you want to show
them in 2 columns with a very high width/height ratio of the figure....

hope it could be of a bit help,

cheers,

Chao

On Mon, May 20, 2013 at 6:43 PM, Martin Mokrejs [via matplotlib] <[hidden email] </user/SendEmail.jtp?type=node&node=41102&i=0>> wrote:

    Hi Ben,

    Benjamin Root wrote:

    >
    >
    >
    > On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs <[hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=0> <mailto:[hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=1>>> wrote:
    >
    > Hi,
    > I am having trouble to get space allocated for a long legend text,
    > lets say spanning 2/3 - 3/4 of the whole output. I would like to have
    > stacked barchart as 1st subplot and the place of remaining 3 subplots
    > to be actually allocated by the legend. Alternatively, could I get the
    > legend saved into a separate figure?
    >
    > Or could the space for legend text be allocated automatically minimizing
    > output figure size? For example, the width would be 1120px while height
    > be multiples of 840px (840 for each subplot)?
    >
    > Attached is a quick example. It shows also that I tried tight_layout()
    > but wasn't successful with this either. I would be glad for some help,
    > ideally converting the whole thing into an object-oriented approach.
    > I am generating several figures in a row and would like to clear()/del()
    > any previously used data ASAP.
    >
    >
    > Thank you,
    > Martin
    > Am using mpl-1.2.2
    >
    >
    > Try "fig.savefig('foobar.png', bbox_inches='tight')" when saving the
    > image. It will make the figure size such that all the visible
    > elements of the figure will fit into the saved output. tight_layout()
    > is meant to make sure the elements don't overlap each other, but does
    > nothing about making sure nothing gets clipped.
    Ah, would be nice to make this clear in the docs. So far was doing

    import pylab
    F = pylab.gcf()
    F.set_tight_layout(True)

    which as you say does not help the way I thought.

    Unfortunately, while

    fig.savefig('foobar.png', bbox_inches='tight')

    helped to get everything into the .png file (attached), the barchart itself
    should span according to the code I posted just 1/2 of the figure. But somehow
    it is enlarged and rescaled so that it occupies *more than* 1/2 of the figure.
    What in pylab is resizing my image? Note: the final image is 625x1075.

    Martin

    ------------------------------------------------------------------------------
    AlienVault Unified Security Management (USM) platform delivers complete
    security visibility with the essential security capabilities. Easily and
    efficiently configure, manage, and operate all of your security controls
    from a single console and one unified framework. Download a free trial.
    http://p.sf.net/sfu/alienvault_d2d
    _______________________________________________
    Matplotlib-users mailing list
    [hidden email] <http://user/SendEmail.jtp?type=node&node=41090&i=2>
    https://lists.sourceforge.net/lists/listinfo/matplotlib-users

    *foobar.png* (132K) Download Attachment <http://matplotlib.1069221.n5.nabble.com/attachment/41090/0/foobar.png>

    ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
    If you reply to this email, your message will be added to the discussion below:
    http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41090.html
    To start a new topic under matplotlib - users, email [hidden email] </user/SendEmail.jtp?type=node&node=41102&i=1>
    To unsubscribe from matplotlib, click here.
    NAML <http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html!nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers!nabble%3Aemail.naml-instant_emails!nabble%3Aemail.naml-send_instant_email!nabble%3Aemail.naml>

--
***********************************************************************************
Chao YUE
Laboratoire des Sciences du Climat et de l'Environnement (LSCE-IPSL)
UMR 1572 CEA-CNRS-UVSQ
Batiment 712 - Pe 119
91191 GIF Sur YVETTE Cedex
Tel: (33) 01 69 08 29 02; Fax:01.69.08.77.16
************************************************************************************

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
View this message in context: Re: Making space for a long legend outside of a barchart <http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41102.html>
Sent from the matplotlib - users mailing list archive <http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html> at Nabble.com.

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi Martin,

I am not sure that I understand your question very well.

For a single scatter() plot, I guess I agree with you, you need to put it in [] because
legend() function must receive iterable as far as I understand.

I don’t think scatter() allows you to pass a series of group of (x,y) data as plot().

So probably if you want to scatter more than one groups of data, you need to:

handle_list = []
label_list = []
for (x,y) in zip(xdata_list, ydata_list):
d = ax1.scatter(x,y)
handle_list.append(x)
label_list.append(…)

ax2.legend(handle_list,label_list,…)

This is what I could think of, perhaps others have better ways.

cheers,

Chao

···

On Thu, May 23, 2013 at 4:57 PM, Martin Mokrejs [via matplotlib] <[hidden email]> wrote:

Hi Chao,

I spent some time to figure out why I cannot replace ax1.hist() with ax1.scatter().

It seems hist() returns list of ‘Rectangle’ (sadly if there is just one, it does return

just the ‘Rectangle’ (not wrapped in a list) … somewhere a trick

a = [a, ]

is likely needed.

Anyway, my problem is that scatter() returns ‘PathCollection’ object, whatever that is.

How can I grab handles to individual legend items to move them under ax2 like in your

hist-plot example?

Thank you for your help,

Martin

ChaoYue wrote:

Dear Martin,

I worked out a similar example for your reference as I don’t catch your example very well.

fig = plt.figure()
ax1 = fig.add_subplot(211)
ax2 = fig.add_subplot(212)
arrlist = [np.random.normal(size=100) for i in range(50)]

ret = ax1.hist(arrlist,histtype=‘barstacked’)

reclist = [patchlist[0] for patchlist in ret[2]]
labellist = [‘data’+str(i) for i in range(50)]
ax2.legend(reclist,labellist,loc=‘upper left’,bbox_to_anchor=(0,0,1,1),borderaxespad=0.,ncol=5,mode=‘expand’)

ax2.set_frame_on(False)

ax2.tick_params(bottom=‘off’,left=‘off’,right=‘off’,top=‘off’)
plt.setp(ax2.get_yticklabels(),visible=False)

plt.setp(ax2.get_xticklabels(),visible=False)

you’re asking some object-oriented way, I personally don’t think using pylab and set_tight_layout are the good way

to be “object-oriented” as pylab is only a bounding wrapper by my understanding (maybe I am wrong!). legend and

hist are all matplotlib.axes.Axes method.

Also, I think it’s unrealistic to ask the figure do a nice job for you if there are 50 legned handlers and you want to show

them in 2 columns with a very high width/height ratio of the figure…

hope it could be of a bit help,

cheers,

Chao

On Mon, May 20, 2013 at 6:43 PM, Martin Mokrejs [via matplotlib] <[hidden email] </user/SendEmail.jtp?type=node&node=41102&i=0>> wrote:

Hi Ben,
Benjamin Root wrote:
>
>
>
> On Mon, May 20, 2013 at 12:02 PM, Martin Mokrejs <[hidden email] <[http://user/SendEmail.jtp?type=node&node=41090&i=0](http://user/SendEmail.jtp?type=node&node=41090&i=0)> <mailto:[hidden email] <[http://user/SendEmail.jtp?type=node&node=41090&i=1](http://user/SendEmail.jtp?type=node&node=41090&i=1)>>> wrote:
>
>     Hi,
>       I am having trouble to get space allocated for a long legend text,
>     lets say spanning 2/3 - 3/4 of the whole output. I would like to have
>     stacked barchart as 1st subplot and the place of remaining 3 subplots
>     to be actually allocated by the legend. Alternatively, could I get the
>     legend saved into a separate figure?
>
>     Or could the space for legend text be allocated automatically minimizing
>     output figure size? For example, the width would be 1120px while height
>     be multiples of 840px (840 for each subplot)?
>
>       Attached is a quick example. It shows also that I tried tight_layout()
>     but wasn't successful with this either. I would be glad for some help,
>     ideally converting the whole thing into an object-oriented approach.
>     I am generating several figures in a row and would like to clear()/del()
>     any previously used data ASAP.
>
>
>     Thank you,
>     Martin
>     Am using mpl-1.2.2
>
>
> Try "fig.savefig('foobar.png', bbox_inches='tight')" when saving the
> image. It will make the figure size such that all the visible
> elements of the figure will fit into the saved output. tight_layout()
> is meant to make sure the elements don't overlap each other, but does
> nothing about making sure nothing gets clipped.
Ah, would be nice to make this clear in the docs. So far was doing
import pylab
F = pylab.gcf()
F.set_tight_layout(True)
which as you say does not help the way I thought.
Unfortunately, while
fig.savefig('foobar.png', bbox_inches='tight')
helped to get everything into the .png file (attached), the barchart itself
should span according to the code I posted just 1/2 of the figure. But somehow
it is enlarged and rescaled so that it occupies *more than* 1/2 of the figure.
What in pylab is resizing my image? Note: the final image is 625x1075.
Martin
------------------------------------------------------------------------------
AlienVault Unified Security Management (USM) platform delivers complete
security visibility with the essential security capabilities. Easily and
efficiently configure, manage, and operate all of your security controls
from a single console and one unified framework. Download a free trial.
[http://p.sf.net/sfu/alienvault_d2d](http://p.sf.net/sfu/alienvault_d2d)
_______________________________________________
Matplotlib-users mailing list
[hidden email] <[http://user/SendEmail.jtp?type=node&node=41090&i=2](http://user/SendEmail.jtp?type=node&node=41090&i=2)>
[https://lists.sourceforge.net/lists/listinfo/matplotlib-users](https://lists.sourceforge.net/lists/listinfo/matplotlib-users)

*foobar.png* (132K) Download Attachment <[http://matplotlib.1069221.n5.nabble.com/attachment/41090/0/foobar.png](http://matplotlib.1069221.n5.nabble.com/attachment/41090/0/foobar.png)>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
If you reply to this email, your message will be added to the discussion below:
[http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41090.html](http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41090.html)
To start a new topic under matplotlib - users, email [hidden email] </user/SendEmail.jtp?type=node&node=41102&i=1>
To unsubscribe from matplotlib, click here.
NAML <[http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml](http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml)>


Chao YUE

Laboratoire des Sciences du Climat et de l’Environnement (LSCE-IPSL)

UMR 1572 CEA-CNRS-UVSQ

Batiment 712 - Pe 119

91191 GIF Sur YVETTE Cedex

Tel: (33) 01 69 08 29 02; Fax:01.69.08.77.16



View this message in context: Re: Making space for a long legend outside of a barchart <http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41102.html>

Sent from the matplotlib - users mailing list archive <http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html> at Nabble.com.


Try New Relic Now & We’ll Send You this Cool Shirt

New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your

browser, app, & servers with just a few lines of code. Try New Relic

and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may


Matplotlib-users mailing list

[hidden email]

https://lists.sourceforge.net/lists/listinfo/matplotlib-users


Try New Relic Now & We’ll Send You this Cool Shirt

New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your

browser, app, & servers with just a few lines of code. Try New Relic

and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may


Matplotlib-users mailing list

[hidden email]

https://lists.sourceforge.net/lists/listinfo/matplotlib-users


If you reply to this email, your message will be added to the discussion below:

http://matplotlib.1069221.n5.nabble.com/Making-space-for-a-long-legend-outside-of-a-barchart-tp41088p41114.html

To start a new topic under matplotlib - users, email [hidden email]

  To unsubscribe from matplotlib, click here.


  [NAML](http://matplotlib.1069221.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml)


Chao YUE
Laboratoire des Sciences du Climat et de l’Environnement (LSCE-IPSL)
UMR 1572 CEA-CNRS-UVSQ
Batiment 712 - Pe 119
91191 GIF Sur YVETTE Cedex

Tel: (33) 01 69 08 29 02; Fax:01.69.08.77.16



View this message in context: Re: Making space for a long legend outside of a barchart

Sent from the matplotlib - users mailing list archive at Nabble.com.