matplotlib-devel Mailing List for matplotlib (Page 7)

The help for xlim() says:

     Set/Get the xlimits of the current axes::

       xmin, xmax = xlim()   # return the current xlim
       xlim( (xmin, xmax) )  # set the xlim to xmin, xmax
       xlim( xmin, xmax )    # set the xlim to xmin, xmax


but it also has the unexpected behavior of turning off autoscaling if used:

-----
import matplotlib.pyplot as plt

plt.clf()
ax = plt.subplot(211)
plt.draw()
print 'autoscale X on: ',ax._autoscaleXon,' xlim: ',plt.xlim()
ax.plot([0,.5,1,1.5,2],[0,1,0,1,0])
plt.draw()
print 'autoscale X on: ',ax._autoscaleXon,' xlim: ',ax.get_xlim(),'\n'

ax = plt.subplot(212)
plt.draw()
print 'autoscale X on: ',ax._autoscaleXon,' xlim: ',ax.get_xlim()
plt.plot([0,.5,1,1.5,2],[0,1,0,1,0])
plt.draw()
print 'autoscale X on: ',ax._autoscaleXon,' xlim: ',ax.get_xlim(),'\n'
-----

returns:

 >>> import xlim_unautoscale
autoscale X on:  True  xlim:  (0.0, 1.0)
autoscale X on:  False  xlim:  (0.0, 1.0)

autoscale X on:  True  xlim:  (0.0, 1.0)
autoscale X on:  True  xlim:  (0.0, 2.0)


I assume that this is because xlim() calls set_xlim() which has 
auto=False as a default keyword...

expected behavior: xlim() should behave exactly like get_xlim()
ditto for ylim()

M

I'll stay on the lookout for the github py3k branch and will do as
suggested. Thanks for the update.

-- Fernando



On Thu, Feb 3, 2011 at 06:05, Michael Droettboom <md...@st...> wrote:
> On 02/03/2011 07:48 AM, Darren Dale wrote:
>> On Thu, Feb 3, 2011 at 12:12 AM, Fernando Garcia Bermudez
>> <mo...@gm...>  wrote:
>>
>>> Before continuing the debug process, though, I wanted to check with
>>> someone knowledgeable of this branch regarding its current status. Is
>>> it possible to compile matplotlib on python3?
>>>
>> Mike D. did some initial work for py3 support a while back. I don't
>> know what platform he was working on, but he was able to produce a
>> simple plot, maybe running the simple_plot demo.
>>
> I was working on Linux (RHEL5), and never got past that platform.
>
> Mike
>
> --
> Michael Droettboom
> Science Software Branch
> Space Telescope Science Institute
> Baltimore, Maryland, USA
>
>
> ------------------------------------------------------------------------------
> Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
> Finally, a world-class log management solution at an even better price-free!
> Download using promo code Free_Logger_4_Dev2Dev. Offer expires
> February 28th, so secure your free ArcSight Logger TODAY!
> http://p.sf.net/sfu/arcsight-sfd2d
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>

On 02/03/2011 07:48 AM, Darren Dale wrote:
> On Thu, Feb 3, 2011 at 12:12 AM, Fernando Garcia Bermudez
> <mo...@gm...>  wrote:
>    
>> Before continuing the debug process, though, I wanted to check with
>> someone knowledgeable of this branch regarding its current status. Is
>> it possible to compile matplotlib on python3?
>>      
> Mike D. did some initial work for py3 support a while back. I don't
> know what platform he was working on, but he was able to produce a
> simple plot, maybe running the simple_plot demo.
>    
I was working on Linux (RHEL5), and never got past that platform.

Mike

-- 
Michael Droettboom
Science Software Branch
Space Telescope Science Institute
Baltimore, Maryland, USA

On Thu, Feb 3, 2011 at 12:12 AM, Fernando Garcia Bermudez
<mo...@gm...> wrote:
> Before continuing the debug process, though, I wanted to check with
> someone knowledgeable of this branch regarding its current status. Is
> it possible to compile matplotlib on python3?

Mike D. did some initial work for py3 support a while back. I don't
know what platform he was working on, but he was able to produce a
simple plot, maybe running the simple_plot demo.

I understand there is still plenty of work to do, and some of the devs
recently discussed it off list. In the next couple of days, I should
be able to finish converting the svn/sourceforge repository over to
git/github, which will make it easier for some (like me) to contribute
to different branches of development. My plan is to rebase the py3k
branch so we have a more recent point of reference, and I will
probably split the py3k branch into its own repository. I'd like to
suggest you clone that repository once it is available, apply your
changes in a new branch, and file a pull request on github.

Darren

[Resending in plain text, for better archiving]

I stumbled across this post on the matplotlib-users archives
(http://www.mailinglistarchive.com/html/mat...@li.../2010-09/msg00359.html)
that linked to the py3k branch. After doing a checkout, though, I ran
into issues running the make.osx script as the README.osx recommends
doing. Since I'm assuming I cannot commit any changes to the branch,
let me comment on the few changes that enabled me to compile the
dependencies so far.

Because I wanted to compile matplotlib for Python 3, I initially
changed the 4th line to:

PYVERSION=3.1

Apparently urllib changed a bit from python 2 to 3, so I had to change
all calls to it to urllib.request, which resulted in the following
changes to lines 34-37:

    ${PYTHON} -c 'import urllib.request;
urllib.request.urlretrieve("http://sourceforge.net/projects/libpng/files/zlib/${ZLIBVERSION}/zlib-${ZLIBVERSION}.tar.gz/download",
"zlib-${ZLIBVERSION}.tar.gz")'  &&\
    ${PYTHON} -c 'import urllib.request;
urllib.request.urlretrieve("http://sourceforge.net/projects/libpng/files/libpng12/older-releases/${PNGVERSION}/libpng-${PNGVERSION}.tar.gz/download",
"libpng-${PNGVERSION}.tar.gz")' &&\
    ${PYTHON} -c 'import urllib.request;
urllib.request.urlretrieve("http://download.savannah.gnu.org/releases/freetype/freetype-${FREETYPEVERSION}.tar.bz2",
"freetype-${FREETYPEVERSION}.tar.bz2")'

Note that I also updated the links for zlib and libpng. The original
zlib link worked, but I couldn't find the sourceforge project (the url
seemed to be old), so I updated it to the one distributed through
libpng's project nowadays (which is what the other seems to have been,
anyway). For libpng, though, the previous link didn't work and instead
of giving an error when downloading, it just created a file that
couldn't be uncompressed. Changing these three lines above enabled the
download and proper compilation of all deps.

I also bumped into a minor typo on line 25, that caused clean not to
delete the downloaded libpng tarball:

    rm -rf zlib-${ZLIBVERSION}.tar.gz libpng-${PNGVERSION}.tar.gz \

After making all these changes, though, the compilation broke when
either making mpl_build or mpl_install.

Before continuing the debug process, though, I wanted to check with
someone knowledgeable of this branch regarding its current status. Is
it possible to compile matplotlib on python3? If so, what are the
limitations to its capabilities?

Thanks so much for your time,

-- Fernando

P.S.: I'm attaching a diff with all these changes for convenience.

Ben,

I had a thought about what you noticed with the paths, so I got in there and
modified plot_directive.py to replace all the '/' with os.sep, but it didn't
seem to do anything.  All the links were still broken even though it seems
like they all point to the right places.  A problem of absolute versus
relative paths?  The paths to the inline plots all seem to take absolute
paths and the RST links are relative.

Ian

----
Ian Bell
Graduate Research Assistant
Herrick Labs
Purdue University
email: ib...@pu...
cell: (607)227-7626


On Mon, Jan 31, 2011 at 1:43 AM, Ian Bell <ib...@pu...> wrote:

> Ben,
>
> Its the main trunk, svn info is below:
>
> Path: .
> URL:
> https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/trunk/matplotlib
> Repository Root: https://matplotlib.svn.sourceforge.net/svnroot/matplotlib
> Repository UUID: f61c4167-ca0d-0410-bb4a-bb21726e55ed
> Revision: 8935
> Node Kind: directory
> Schedule: normal
> Last Changed Author: pivanov314
> Last Changed Rev: 8935
> Last Changed Date: 2011-01-24 04:41:49 -0500 (Mon, 24 Jan 2011)
>
> Ian
>
> ----
> Ian Bell
> Graduate Research Assistant
> Herrick Labs
> Purdue University
> email: ib...@pu...
> cell: (607)227-7626
>
>
> On Mon, Jan 31, 2011 at 1:36 AM, Benjamin Root <ben...@ou...> wrote:
>
>> On Sunday, January 30, 2011, Ian Bell <ib...@pu...> wrote:
>> > Ben,
>> > I'm building with the freshest subversion build of MPL with Sphinx 1.0.7
>> and Windows XP.  I'd be happy to give you any other information that you
>> need, just let me know.
>> >
>> >
>> > Ian----
>> > Ian Bell
>> > Graduate Research Assistant
>> > Herrick Labs
>> > Purdue University
>> > email: ib...@pu...
>> > cell: (607)227-7626
>> >
>>
>> From which branch?  The development branch or the maintenance branch?
>> If you aren't sure, just paste the results of 'svn info' (or whatever
>> equivalent in your windows svn client).
>>
>> Ben Root
>>
>
>

I'm currently trying to add all remaining high level api into freetype-py so FT_Get_Sfnt_Name should be in a few hours hopefully.
If I forgot to add other useful functions, just tell me.

The agg-trick.py is very similar from what I found in ft2font in matploltib: setting a resolution at (10*dpi, dpi) and a transform matri =(.1,0,0,1).
One output is available on: http://www.loria.fr/~rougier/coding/freetype/agg-trick.png
(This is with subpixel antialias)

I will also integrate the ft_utils within freetype-py.


Nicolas


On Feb 1, 2011, at 8:36 PM, Michael Droettboom wrote:

> I've made a basic first crack at this.
> 
> Thanks to your examples, it was really easy to get basic text rendering with matplotlib's Agg backend (the other backends will need to be updated accordingly, and they may be more involved since they actually need to extract curves).  I've made an SVN branch for this here: (I know, I know, git would make this easier, but afaik we're not quite there yet...)
> 
> https://matplotlib.svn.sf.net/svnroot/matplotlib/branches/ctypes_freetype
> 
> This doesn't use the Agg tricks rendering (finer grained hinting in the horizontal direction etc.) so it's a little lower quality than matplotlib trunk right now.  However, your agg-tricks.py example certainly shows the way forward to making that work.  I've tried to localise all of the higher-level stuff in freetype_util.py (in matplotlib) to make this easier.  It may also make sense to put this kind of higher-level (but extremely common) functionality in freetype-py itself.
> 
> The next stumbling block to removing the existing wrapper is the FT_Get_Sfnt_Name API which doesn't seem to be in your ctypes wrappers.  This is used to extract metadata about the fonts and build the font cache table etc.  Hopefully it's simple to add this?
> 
> Mike
> 
> On 02/01/2011 10:53 AM, Nicolas Rougier wrote:
>> 
>> 
>> 
>> Hi all,
>> 
>> I've bound (part of) the high-level API of freetype using ctypes.
>> 
>> I don't know if it might be useful for matplotib text rendering but  just in case:
>> http://code.google.com/p/freetype-py/
>> 
>> 
>> 
>> Nicolas
>> 
>> ------------------------------------------------------------------------------
>> Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
>> Finally, a world-class log management solution at an even better price-free!
>> Download using promo code Free_Logger_4_Dev2Dev. Offer expires 
>> February 28th, so secure your free ArcSight Logger TODAY! 
>> http://p.sf.net/sfu/arcsight-sfd2d
>> 
>> _______________________________________________
>> Matplotlib-devel mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
> 

On Tue, Feb 1, 2011 at 2:36 PM, Michael Droettboom <md...@st...> wrote:
> I've made a basic first crack at this.
>
> Thanks to your examples, it was really easy to get basic text rendering with
> matplotlib's Agg backend (the other backends will need to be updated
> accordingly, and they may be more involved since they actually need to
> extract curves).  I've made an SVN branch for this here: (I know, I know,
> git would make this easier, but afaik we're not quite there yet...)
>
> https://matplotlib.svn.sf.net/svnroot/matplotlib/branches/ctypes_freetype

I've added this to the list of branches to be converted to the git repo.

I've made a basic first crack at this.

Thanks to your examples, it was really easy to get basic text rendering 
with matplotlib's Agg backend (the other backends will need to be 
updated accordingly, and they may be more involved since they actually 
need to extract curves).  I've made an SVN branch for this here: (I 
know, I know, git would make this easier, but afaik we're not quite 
there yet...)

https://matplotlib.svn.sf.net/svnroot/matplotlib/branches/ctypes_freetype

This doesn't use the Agg tricks rendering (finer grained hinting in the 
horizontal direction etc.) so it's a little lower quality than 
matplotlib trunk right now.  However, your agg-tricks.py example 
certainly shows the way forward to making that work.  I've tried to 
localise all of the higher-level stuff in freetype_util.py (in 
matplotlib) to make this easier.  It may also make sense to put this 
kind of higher-level (but extremely common) functionality in freetype-py 
itself.

The next stumbling block to removing the existing wrapper is the 
FT_Get_Sfnt_Name API which doesn't seem to be in your ctypes wrappers.  
This is used to extract metadata about the fonts and build the font 
cache table etc.  Hopefully it's simple to add this?

Mike

On 02/01/2011 10:53 AM, Nicolas Rougier wrote:
>
>
> Hi all,
>
> I've bound (part of) the high-level API of freetype using ctypes.
>
> I don't know if it might be useful for matplotib text rendering but 
>  just in case:
> http://code.google.com/p/freetype-py/
>
>
>
> Nicolas
>
>
> ------------------------------------------------------------------------------
> Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
> Finally, a world-class log management solution at an even better price-free!
> Download using promo code Free_Logger_4_Dev2Dev. Offer expires
> February 28th, so secure your free ArcSight Logger TODAY!
> http://p.sf.net/sfu/arcsight-sfd2d
>
>
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

Thanks for doing this!  There was certainly nothing like this last time 
I looked.

I'll look further into whether this is a useful replacement for the 
domain-specific freetype wrappers in matplotlib now...

Cheers,
Mike

On 02/01/2011 10:53 AM, Nicolas Rougier wrote:
>
>
> Hi all,
>
> I've bound (part of) the high-level API of freetype using ctypes.
>
> I don't know if it might be useful for matplotib text rendering but 
>  just in case:
> http://code.google.com/p/freetype-py/
>
>
>
> Nicolas
>
>
> ------------------------------------------------------------------------------
> Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
> Finally, a world-class log management solution at an even better price-free!
> Download using promo code Free_Logger_4_Dev2Dev. Offer expires
> February 28th, so secure your free ArcSight Logger TODAY!
> http://p.sf.net/sfu/arcsight-sfd2d
>
>
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

Hi all,

I've bound (part of) the high-level API of freetype using ctypes.

I don't know if it might be useful for matplotib text rendering but  just in case:
http://code.google.com/p/freetype-py/



Nicolas