matplotlib-users Mailing List for matplotlib (Page 10)

Excellent!  Thanks again!
>   from matplotlib.transforms import unit_bbox, get_bbox_transform
>   trans =3D get_bbox_transform(unit_bbox(), ax.bbox)

--
Charles R. Twardy

     They are statically included with windows just as they were before.
     Virtually every linux distribution has png/freetype/zlib as part
of they base package set, and if not they are easily installable, so
these are dynamically linked in the eggs I posted on sf.
     In the case of osx I linked against static libraries of png and
freetype and added them to the egg.  Zlib is included with osx.  So,
no you don't have to worry about png/freetype.  Here is the link again
for the osx.4 eggs: http://euclid.uits.iupui.edu/~cmoad/mpleggs/ .  I
would really like to hear from anyone if they attempt to use these
whether they worked or not.  That way I know whether they should be
posted on sf as well.

On 1/13/06, Tom Loredo <lo...@as...> wrote:
>
> Charlie wrote:
>
> > I will post osx eggs after some more testing.
>
> Can you clarify the use of eggs for me?  Does the user still
> have to be concerned with installing libraries such as
> FreeType, libpng and zlip, or does the egg include these
> (statically linked?) or install them somewhere?  If the
> user must have the libraries, must they be in some special
> place (e.g., /usr/local vs. Fink's path)?
>
> Thanks,
> Tom
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi=
les
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Charlie wrote:

> I will post osx eggs after some more testing.

Can you clarify the use of eggs for me?  Does the user still
have to be concerned with installing libraries such as
FreeType, libpng and zlip, or does the egg include these
(statically linked?) or install them somewhere?  If the
user must have the libraries, must they be in some special
place (e.g., /usr/local vs. Fink's path)?

Thanks,
Tom

On 10/19/05, Alexander Borghgraef <ale...@gm...>
wrote:
>
> On 10/13/05, Nadezhda Dencheva <den...@st...> wrote:
> >
> > As a final desperate act, install libpng from source in a new directory
> > and try linking to it (using MPLIB_BASE).
>
>
>  So, if I got it right, I compile and install libpng in a local directory=
,
> point an environment variable MPLIB_BASE
> to the /usr/local equivalent (not to the actual directory containing
> libpng.so, usually /usr/local/lib) and then run
>
> python setup.py build
> python setup.py install --prefix $HOME/local
>
> This is getting really frustrating, did it, same error:
>
> ImportError:
> /homes/morlet/aborghgr/local/lib/python2.3/site-packages/matplotlib/_nc_i=
mage.so:
> undefined symbol: png_set_sBIT
>
> And png_set_sBIT is present in the local install of libpng, that one isn'=
t
> stripped of symbols.
>

Ok, tried again with matplotlib 0.86, same error. But this time I dug a bit
deeper, and I noticed that the MPLIB_BASE environment variable was
only taken into account in the case of a solaris system, and I'm working on
linux. Which means that the build hadn't linked to my locally installed
libpng, but to the apparently defective global one. So I edited the linux2
entry in "basedir" from setupext.py, and tried again, with success this
time. So I think we can classify this as solved. Again, thanks for the help
all.

--
Alex Borghgraef

Hi list,

does anyone know a way to increase all font sizes at once e.g.
tickmarks, legend, title, text, ... everything. I was not able to find
the rc-parameter for this purpose.

Joost

Charlie Moad wrote:
> On 1/12/06, Steve Schmerler <el...@gm...> wrote:
> 
>>Hi
>>
>>I discoverd some problems while playing with my fresh installed 0.86
>>(much faster then 0.82 ... great work!):
>>
>>rcParams['text.usetex']=False;plot([1,2,3]); xlabel("x-axis");
>>ylabel("y-label"); savefig("image.eps")
>>
>>produces the expected result (nice bounding box etc.) but using
>>usetex=True
>>
>>rcParams['text.usetex']=True; plot([1,2,3]); xlabel("x-axis");
>>ylabel("y-label"); savefig("image.eps")
>>
>>results in an .eps where the axes tick numbering and the labels are
>>missing (bad bounding box when viewing it with gv). Saving it as .ps
>>
>>rcParams['text.usetex']=True; plot([1,2,3]); xlabel("x-axis");
>>ylabel("y-label"); savefig("image.ps")
>>
>>produces a .ps file whose BB is much bigger then it should be (looks
>>like A4).
>>
> 
> 
> The 0.86.1 release had some usetex related issues updated.  Please try
> it and let us know if the issue is still there.
> 

The BB problems remain plus:

1) the axes labels aren't centered (xlabel on the left side of the
x-axis, ylabel on the bottom of the y-axis)

2) somtimes I got

In [2]: rcParams['text.usetex']=True; plot([1,2,3]); xlabel("x-axis");
ylabel("y-label"); savefig("image.eps")
Out[2]: [<matplotlib.lines.Line2D instance at 0xb541912c>]
Out[2]: <matplotlib.text.Text instance at 0xb54108ec>
Out[2]: <matplotlib.text.Text instance at 0xb541646c>

(ipython:8099): Pango-WARNING **: Invalid UTF-8 string passed to
pango_layout_set_text()

with a crashing plot window, sometimes the call gave no erros, the plot
appeared.

cheers,
steve

-- 
"People like Blood Sausage too. People are Morons!" -- Phil Connors,
Groundhog Day

>One question for Travis. On the web site, the NumPy item is listed
>as "pre-order". I realize it was only what last week that the decision
>to call it NumPy was decided (I like that name as well), but what does
>pre-order mean (e.g. when do you expect the e-book will be ready)?
>I know there has been some discussion of this on the scipy list,
>sorry if i don't recall if this question has already been covered.
>  
>

The (pre-order) status will be there until the e-book is completely 
finished.  However, it is pretty far along already, and I deliver the 
current copy immediately as an email attachment.

The C-API section needs some updating after the change (still some 
references to SciPy Core),  and the masked array and distutils/testing 
chapters need to be completed on the Python side.  On the C-API side, 
the full C-API is documented, but  I need to complete the chapters 
explaining how to extend NumPy in a simple way, a chapter showing how 
some more advanced methods, and the final chapter that explains some 
more difficult sections of the source code.

Basically, 11/16 chapters are finished.   The addition of improved 
record arrays, the data-type descriptor objects, and the renaming set me 
back a bit on finishing the book, but it is still very useful right now 
--- especially on the Python side of things.  

For people who write C-extensions the old Numeric C-API still works 
(just replace Numeric/arrayobject.h  with numpy/arrayobject.h), but some 
of the new possibilities include defining your own data-type in C and 
subclassing in C from the ndarray.   Also, much more of the 
functionality of NumPy is available as a C-call (instead of always going 
through Python method calls). 

I appreciate John's announcement.  I'm very excited that we all might be 
able to convert over to a single array sometime soon.

-Travis

On Thu, 12 Jan 2006, John Hunter wrote:

> 
> As many of you know, Travis has been working heroically on the next
> generation Numeric, now called numpy
> 
> http://sourceforge.net/project/showfiles.php?group_id=1369&package_id=175103
> 
> numpy's goal is to combine the best of Numeric and numarray and unify
> the two camps of array users.  Having two array packages has been a
> major drag on scientific python, since packages available for one may
> not work with the other, splitting developer resources and driving
> folks to Ruby :-) So getting everyone to agree on and use a single
> array package will be a major step forward, and Travis has been
> working with the numarray camp to make sure that numpy will have all
> the features they need. As I understand it, Perry and Todd and crew
> have decided to transition to numpy as it becomes feasible for them.
> 
> The latest release of matplotlib (many thanks to Charlie Moad for
> building and testing across platforms) supports numpy in the numerix
> layer (set numerix in your matplotlibrc file), and the windows build
> has support for all three packages (Numeric, numarray and numpy) built
> in.
> 
> I encourage everyone to download the latest numpy and give it a test
> drive with matplotlib, since ultimately we would like to stop
> supporting three array packages and just adopt one.  This will
> significantly reduce compile times and binary distribution sizes, and
> will make the code and build process cleaner.  We don't have a
> specific roadmap for when the numerix layer will become deprecated,
> but we would like it to happen sooner rather than later.
> 
> You can profile your matplotlib scripts by using the numerix flags
> from the prompt, eg
> 
>   > time python myscript.py --numpy
>   > time python myscripy.py --Numeric
>   > time python myscripy.py --numarray
> 
> Also, note that when building matplotlib from source, the matplotlibrc
> file is now autogenerated, setting the backend and numerix setting
> depending on what was found at compile time.  So make sure the
> generated rc file is what you want, and if not edit it and move it to
> your ~/.matplotlib dir.
> 
> Travis has written an excellent guide to numpy that can be purchased
> at http://www.tramy.us .  If you can afford it, it's a great way to
> learn all the new features and support Travis (and his 5 kids!) and
> scipy development.  I think he will give discounts to the financially
> challenged, so don't hesitate to contact him about it.
>   
> Thanks to Daishi, Eric and Andrew for their work on the numpy support.
> 
> JDH
> 

Since this is sanctioned by John, I'll definitely buy a copy
(I also have a copy of Numerical Python...which has among others,
Travis's name in the author list...a very nice document).
I have been following both the matplotlib and the scipy lists for 
several years now. I'm _way_ behind on upgrading to the latest 
mpl...but I will soon. Scipy has been something that I have been 
following, but I have not yet tried. I did install numeric...but not
numarray. I am very glad that things appear to be consolidating
to numpy. 

One question for Travis. On the web site, the NumPy item is listed
as "pre-order". I realize it was only what last week that the decision
to call it NumPy was decided (I like that name as well), but what does
pre-order mean (e.g. when do you expect the e-book will be ready)?
I know there has been some discussion of this on the scipy list,
sorry if i don't recall if this question has already been covered.

Thanks,

Jim

As many of you know, Travis has been working heroically on the next
generation Numeric, now called numpy

http://sourceforge.net/project/showfiles.php?group_id=1369&package_id=175103

numpy's goal is to combine the best of Numeric and numarray and unify
the two camps of array users.  Having two array packages has been a
major drag on scientific python, since packages available for one may
not work with the other, splitting developer resources and driving
folks to Ruby :-) So getting everyone to agree on and use a single
array package will be a major step forward, and Travis has been
working with the numarray camp to make sure that numpy will have all
the features they need. As I understand it, Perry and Todd and crew
have decided to transition to numpy as it becomes feasible for them.

The latest release of matplotlib (many thanks to Charlie Moad for
building and testing across platforms) supports numpy in the numerix
layer (set numerix in your matplotlibrc file), and the windows build
has support for all three packages (Numeric, numarray and numpy) built
in.

I encourage everyone to download the latest numpy and give it a test
drive with matplotlib, since ultimately we would like to stop
supporting three array packages and just adopt one.  This will
significantly reduce compile times and binary distribution sizes, and
will make the code and build process cleaner.  We don't have a
specific roadmap for when the numerix layer will become deprecated,
but we would like it to happen sooner rather than later.

You can profile your matplotlib scripts by using the numerix flags
from the prompt, eg

  > time python myscript.py --numpy
  > time python myscripy.py --Numeric
  > time python myscripy.py --numarray

Also, note that when building matplotlib from source, the matplotlibrc
file is now autogenerated, setting the backend and numerix setting
depending on what was found at compile time.  So make sure the
generated rc file is what you want, and if not edit it and move it to
your ~/.matplotlib dir.

Travis has written an excellent guide to numpy that can be purchased
at http://www.tramy.us .  If you can afford it, it's a great way to
learn all the new features and support Travis (and his 5 kids!) and
scipy development.  I think he will give discounts to the financially
challenged, so don't hesitate to contact him about it.
  
Thanks to Daishi, Eric and Andrew for their work on the numpy support.

JDH

It may of not hit the sf mirror that was randomly picked.  Sometimes
the files take about a day to make the rounds.  Try running it a few
times, and I bet you will hit one that has it.

On 1/12/06, Charles R. Twardy <ct...@gm...> wrote:
> > "easy_install -U matplotlib" from linux or windows and it will
> Downloading....
> error: Unexpected HTML page found at
> http://prdownloads.soruceforge.net/matplotlib/matplotlib-0.86-1.tar.gz?do=
wnload
>
> Debian Linux/ppc.
>
> --
> Charles R. Twardy
>

>>>>> "Charles" == Charles R Twardy <ct...@gm...> writes:

    >> trans = blend_xy_sep_transform(ax.transAxes, ax.transAxes)
    >> 
    >> This is basically a no-op; if the axes transform is the one you
    >> want, just use it
    >> 
    >> ax.text(0.5, 1, "My Title", ha='center',
    >> va='top',transform=ax.transAxes)

    Charles> Can't do that.  Puts the text too high, runs into the top
    Charles> bar and the tickmarks.  Can't apply the offset to

Oh right (slaps self on head).  You need to set the offset on the
transform and you can't apply that to ax.transAxes since it will screw
up the axes transform...  So you need to copy it, or duplicate it.
blend is one way to copy the transform, but there is a better way --
just mimic what axes.py does when it creates it.


  from matplotlib.transforms import unit_bbox, get_bbox_transform
  trans = get_bbox_transform(unit_bbox(), ax.bbox)

this transforms the (0,0), (1,1) bounding box to the axes bounding
box.  You can then set the offset of this transform as before...

JDH

>   trans =3D blend_xy_sep_transform(ax.transAxes, ax.transAxes)
>
> This is basically a no-op; if the axes transform is the one you want,
> just use it
>
>   ax.text(0.5, 1, "My Title", ha=3D'center', va=3D'top',transform=3Dax.tr=
ansAxes)

Can't do that.  Puts the text too high, runs into the top bar and the tickm=
arks.
Can't apply the offset to ax.transAxes either  -- it changes
*EVERYTHING*.  The old "assignment is reference" problem.  But if
blend is nearly a no-op, then I'll use it.

I don't need a data component to put the text below the tickmarks. But
I *did* need the magic offset numbers rcParams['xtick.major.size'] and
gcf().dpi/Value(72.), plus your knowledge of scale_transform and
set_offset.  And it seems blend_ gives me a copy operator for
transAxes.

Thanks! Thanks! Thanks!

-C

--
Charles R. Twardy

In addition to the standard downloads I put eggs up for windows/linux
py2.3/2.4 combinations.  The pypi download link for matplotlib has
been set to the sourceforge download page.  Now you can run
"easy_install -U matplotlib" from linux or windows and it will
automatically look at pypi, redirect to mpl sf download page, find the
right binary depending on your platform, and install matplotlib.  This
makes updates/installations pretty trivial.

I will post osx eggs after some more testing.

- Charlie

On 1/12/06, Charlie Moad <cw...@gm...> wrote:
> This minor release fixes the windows installer issues.  A few updates
> snuck in as well.
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> 2006-1-11 Fixed setup.py for win32 build and added rc template to the
> MANIFEST.in
>
> 2006-1-10 Added xpdf distiller option. matplotlibrc ps.usedistiller can n=
ow be
>             none, false, ghostscript, or xpdf. Validation checks for
>             dependencies. This needs testing, but the xpdf option should =
produce
>             the highest-quality output and small file sizes - DSD
>
> 2006-01-10 For the usetex option, backend_ps now does all the LaTeX work =
in the
>             os's temp directory - DSD
>
> 2006-1-10 Added checks for usetex dependencies. - DSD
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>

On 1/12/06, Steve Schmerler <el...@gm...> wrote:
> Hi
>
> I discoverd some problems while playing with my fresh installed 0.86
> (much faster then 0.82 ... great work!):
>
> rcParams['text.usetex']=3DFalse;plot([1,2,3]); xlabel("x-axis");
> ylabel("y-label"); savefig("image.eps")
>
> produces the expected result (nice bounding box etc.) but using
> usetex=3DTrue
>
> rcParams['text.usetex']=3DTrue; plot([1,2,3]); xlabel("x-axis");
> ylabel("y-label"); savefig("image.eps")
>
> results in an .eps where the axes tick numbering and the labels are
> missing (bad bounding box when viewing it with gv). Saving it as .ps
>
> rcParams['text.usetex']=3DTrue; plot([1,2,3]); xlabel("x-axis");
> ylabel("y-label"); savefig("image.ps")
>
> produces a .ps file whose BB is much bigger then it should be (looks
> like A4).
>

The 0.86.1 release had some usetex related issues updated.  Please try
it and let us know if the issue is still there.

This minor release fixes the windows installer issues.  A few updates
snuck in as well.

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
2006-1-11 Fixed setup.py for win32 build and added rc template to the
MANIFEST.in

2006-1-10 Added xpdf distiller option. matplotlibrc ps.usedistiller can now=
 be
            none, false, ghostscript, or xpdf. Validation checks for
            dependencies. This needs testing, but the xpdf option should pr=
oduce
            the highest-quality output and small file sizes - DSD

2006-01-10 For the usetex option, backend_ps now does all the LaTeX work in=
 the
            os's temp directory - DSD

2006-1-10 Added checks for usetex dependencies. - DSD

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

On 11/01/06, John Hunter <jdh...@ac...> wrote:
>
> Hi Jose,
>
> Thanks a lot for your efforts -- 0.86 already in Fedora, now that's
> speed!

  Clearly you are the one to thank about it. ;-)

  The other person to thank is Orion (FE maintainer). :-)

> If you could write up a small blurb about where to find mpl for fedora
> with links, I'll put it in the FAQ.

  Since it is in Extras it is enough to issue from the command line as root=
:

# yum install python-matplotlib

  FWIW soon basemap will be also available using the same method:

# yum install python-basemap

> Thanks,
> JDH

  Obrigado. :-)
--
Jos=E9 Matos

John Hunter wrote:
>>>>>>"Jos=E9" =3D=3D Jos=E9 Matos <jao...@gm...> writes:
>=20
>     Jos=E9>   If you or anyone else has problems with it I read this
>     Jos=E9> list and although I am not its maintaineir I volunteer to
>     Jos=E9> report and got those package bugs fixed In Fedora.
>=20
> Hi Jose,
>=20
> Thanks a lot for your efforts -- 0.86 already in Fedora, now that's
> speed!
>=20
> If you could write up a small blurb about where to find mpl for fedora
> with links, I'll put it in the FAQ.

I think that for Fedora 4 users, it's just

	yum install matplotlib

and whenever they do

	yum update

they'll get the newest version.  These days, Fedora Extras (where matplot=
lib=20
lives, kind of like Ubuntu Universe) is officially on by default.

Cheers,

f

>>>>> "Jos=E9" =3D=3D Jos=E9 Matos <jao...@gm...> writes:
    Jos=E9>   If you or anyone else has problems with it I read this
    Jos=E9> list and although I am not its maintaineir I volunteer to
    Jos=E9> report and got those package bugs fixed In Fedora.

Hi Jose,

Thanks a lot for your efforts -- 0.86 already in Fedora, now that's
speed!

If you could write up a small blurb about where to find mpl for fedora
with links, I'll put it in the FAQ.

Thanks,
JDH

On 11/01/06, Derrick Snowden <Der...@no...> wrote:
> Thanks Charlie, that was it.
>
> yum install gtk2-devel
>
> did the trick.  Incidentally I found python-matplotlib-0.85 at the
> fedora extras repo.  If I remembered that it certainly would have made
> things easier.  Oh well, now I have 0.86.

  0.86 should be soon there as well. It was build already and it only
needs to be signed. (AFAIK)

  If you or anyone else has problems with it I read this list and
although I am not its maintaineir I volunteer to report and got those
package bugs fixed In Fedora.

> Thanks all who replied.
>
> Derrick
--
Jos=E9 Matos

Thanks Charlie, that was it. 

yum install gtk2-devel

did the trick.  Incidentally I found python-matplotlib-0.85 at the 
fedora extras repo.  If I remembered that it certainly would have made 
things easier.  Oh well, now I have 0.86.

Thanks all who replied.

Derrick

Charlie Moad wrote:

>Those should really be part of the gtk2-dev packages.  At least they
>are on ubuntu.
>
>/usr/include/gtk-2.0/gtk/gtk.h
>
>On 1/11/06, Derrick Snowden <Der...@no...> wrote:
>  
>
>>Hi all,
>>
>>I'm trying to install mpl (I've tried both 0.86 and 0.85) on an x86_64
>>machine running Fedora Core 4. I saw one post related to gtk but it
>>referred to a problem different from mine. Below is the last few lines
>>from stderr...
>>
>> > python setup.py build
>>
>>....
>>
>>In file included from src/_nc_backend_gdk.c:12:
>>/usr/include/pygtk-2.0/pygtk/pygtk.h:8:21: error: gtk/gtk.h: No such
>>file or directory
>>In file included from src/_nc_backend_gdk.c:12:
>>/usr/include/pygtk-2.0/pygtk/pygtk.h:14: error: syntax error before 'atom'
>>/usr/include/pygtk-2.0/pygtk/pygtk.h:16: error: syntax error before
>>'GdkRectangle'
>>/usr/include/pygtk-2.0/pygtk/pygtk.h:23: error: syntax error before
>>'GdkAtom'
>>/usr/include/pygtk-2.0/pygtk/pygtk.h:23: warning: no semicolon at end of
>>struct or union
>>/usr/include/pygtk-2.0/pygtk/pygtk.h:24: warning: data definition has no
>>type or storage class
>>src/_nc_backend_gdk.c: In function 'pixbuf_get_pixels_array':
>>src/_nc_backend_gdk.c:32: error: 'GdkPixbuf' undeclared (first use in
>>this function)
>>src/_nc_backend_gdk.c:32: error: (Each undeclared identifier is reported
>>only once
>>src/_nc_backend_gdk.c:32: error: for each function it appears in.)
>>src/_nc_backend_gdk.c:32: error: 'gdk_pixbuf' undeclared (first use in
>>this function)
>>src/_nc_backend_gdk.c:52: warning: cast to pointer from integer of
>>different size
>>error: command 'gcc' failed with exit status 1
>>
>>Now the second line up above says I need gtk.h, which much to my
>>surprise is nowhere on my system. I have gtk2 and gtk+ devel rpms
>>installed and numerous other gtk related packages including pygtk. Is it
>>possible that pygtk installed correctly without gtk or am I missing
>>something subtle (or completely obvious).
>>
>>Thanks for any help,
>>Derrick
>>
>>--
>>Derrick Snowden
>>PhOD/AOML/NOAA
>>4301 Rickenbacker Causeway
>>Miami, FL 33149
>>Office: 305.361.4322  Fax: 305.361.4392
>>
>>
>>
>>
>>    
>>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_idv37&alloc_id865&op=click
>_______________________________________________
>Matplotlib-users mailing list
>Mat...@li...
>https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>  
>


-- 
Derrick Snowden
PhOD/AOML/NOAA
4301 Rickenbacker Causeway
Miami, FL 33149
Office: 305.361.4322  Fax: 305.361.4392

>>>>> "Charles" == Charles R Twardy <ct...@gm...> writes:

    Charles> John, Thanks, that did it.  I don't really need to blend
    Charles> though.  What I need it taken care of by:

  trans = blend_xy_sep_transform(ax.transAxes, ax.transAxes)

This is basically a no-op; if the axes transform is the one you want,
just use it

  ax.text(0.5, 1, "My Title", ha='center', va='top',transform=ax.transAxes)

My example placed the text below the ticks, which is where the data
component came in.

JDH

John,

Thanks, that did it.  I don't really need to blend though.  What I
need it taken care of by:

#trans =3D blend_xy_sep_transform(ax.transData, ax.transAxes)
trans =3D blend_xy_sep_transform(ax.transAxes, ax.transAxes)
scale =3D fig.dpi/Value(72.) # points -> pixels
point_trans =3D scale_transform(scale, scale)
trans.set_offset((0,-ticksize-2),point_trans)
ax.text(0.5, 1, "My Title", ha=3D'center', va=3D'top', transform=3Dtrans)

It seems a bit silly to call the blend_xy transform for that (since
I'm using transAxes for both x and y), but copy.copy doesn't work, so
there we are.

Thanks again.
-C


On 1/11/06, John Hunter <jdh...@ni...> wrote:
> >>>>> "Charles" =3D=3D Charles R Twardy <ct...@gm...> writes:
>
>     Charles> I don't think so.  I looked at that.  Maybe I'm mistaken,
>     Charles> but seems that in the end, it just places the text at y
>     Charles> position 0.9 in Axes coordinates.  That's just below the
>     Charles> markers for small subplots, but WAY below the markers for
>     Charles> tall skinny ones.  I want it to place the text at a fixed
>     Charles> absolute y position of just below the markers. As the
>     Charles> graph gets taller, that will go from .9 to .95 to .99,
>     Charles> etc. in Axes coordinates.
>
>     Charles> So one approach would be to find the Axes coordinates for
>     Charles> the lower limit of the tick markers.  Any ideas?
>
> This is a bit tricky but doable.  The transformation that places the
> xticks is a "blend" of the data transform and the axes transform, in
> that the x value of the x ticks is a data coord, and the y value is an
> axes coord(0 is bottom, 1 is top).  The transform module provides a
> helper function to build such a beast
>
>   from matplotlib.transforms import blend_xy_sep_transform
>   trans =3D blend_xy_sep_transform(ax.transData, ax.transAxes)
>
>
> If you then make a call to text
>
>   ax.text(.2,1,'hi mom', transform=3Dtrans)
>
> it will be placed at the top of the yaxis at 0.2 on the xaxis.
>
>
> What you want to do is offset this by a couple of points below the
> tick like.  To do this, you need to set an offset on the transform,
> where the coordinates of the offset are in points.  The offset is
> always an xy tuple with a transform to transform that tuple into
> figure coords.  In the case of points, you want to do
>
>
>   scale =3D fig.dpi/Value(72.) # points -> pixels
>   point_trans =3D scale_transform(scale, scale)
>   trans.set_offset((0,-ticksize-2),point_trans)
>
>
>
> Here is a complete example: now when you resize the figure window or
> pan/zoom your text will remain 2 points below the ticks...
>
> from matplotlib import rcParams
> import matplotlib.numerix as nx
> from pylab import figure, show
> from matplotlib.transforms import blend_xy_sep_transform, Value, scale_tr=
ansform
>
> ticksize =3D rcParams['xtick.major.size']
>
> fig =3D figure()
> ax =3D fig.add_subplot(111)
> x,y =3D nx.mlab.rand(2,100)
> ax.plot(x,y)
> locs =3D nx.arange(0,1.0,0.2)
> ax.set_xticks(locs)
>
>
> trans =3D blend_xy_sep_transform(ax.transData, ax.transAxes)
> scale =3D fig.dpi/Value(72.) # points -> pixels
> point_trans =3D scale_transform(scale, scale)
>
> # the offset is an xy tup and a transformation instance for that tuple
> trans.set_offset((0,-ticksize-2),point_trans)
>
> for loc in locs:
>     ax.text(loc,1,'%1.1f'%loc, transform=3Dtrans, va=3D'top', ha=3D'cente=
r')
> show()
>
>


--
Charles R. Twardy

Those should really be part of the gtk2-dev packages.  At least they
are on ubuntu.

/usr/include/gtk-2.0/gtk/gtk.h

On 1/11/06, Derrick Snowden <Der...@no...> wrote:
> Hi all,
>
> I'm trying to install mpl (I've tried both 0.86 and 0.85) on an x86_64
> machine running Fedora Core 4. I saw one post related to gtk but it
> referred to a problem different from mine. Below is the last few lines
> from stderr...
>
>  > python setup.py build
>
> ....
>
> In file included from src/_nc_backend_gdk.c:12:
> /usr/include/pygtk-2.0/pygtk/pygtk.h:8:21: error: gtk/gtk.h: No such
> file or directory
> In file included from src/_nc_backend_gdk.c:12:
> /usr/include/pygtk-2.0/pygtk/pygtk.h:14: error: syntax error before 'atom=
'
> /usr/include/pygtk-2.0/pygtk/pygtk.h:16: error: syntax error before
> 'GdkRectangle'
> /usr/include/pygtk-2.0/pygtk/pygtk.h:23: error: syntax error before
> 'GdkAtom'
> /usr/include/pygtk-2.0/pygtk/pygtk.h:23: warning: no semicolon at end of
> struct or union
> /usr/include/pygtk-2.0/pygtk/pygtk.h:24: warning: data definition has no
> type or storage class
> src/_nc_backend_gdk.c: In function 'pixbuf_get_pixels_array':
> src/_nc_backend_gdk.c:32: error: 'GdkPixbuf' undeclared (first use in
> this function)
> src/_nc_backend_gdk.c:32: error: (Each undeclared identifier is reported
> only once
> src/_nc_backend_gdk.c:32: error: for each function it appears in.)
> src/_nc_backend_gdk.c:32: error: 'gdk_pixbuf' undeclared (first use in
> this function)
> src/_nc_backend_gdk.c:52: warning: cast to pointer from integer of
> different size
> error: command 'gcc' failed with exit status 1
>
> Now the second line up above says I need gtk.h, which much to my
> surprise is nowhere on my system. I have gtk2 and gtk+ devel rpms
> installed and numerous other gtk related packages including pygtk. Is it
> possible that pygtk installed correctly without gtk or am I missing
> something subtle (or completely obvious).
>
> Thanks for any help,
> Derrick
>
> --
> Derrick Snowden
> PhOD/AOML/NOAA
> 4301 Rickenbacker Causeway
> Miami, FL 33149
> Office: 305.361.4322  Fax: 305.361.4392
>
>
>
>

>  > That and the are a pain to maintain.
>
> Why's that? I do have a little patch to setup.py and friends to add the
> StaticLibs path, but that's it. I keep meaning to give that patch to
> John, but haven't gotten around to it.

I was building against frameworks of libpng and freetype.  It was the
more mac way of doing things, but took a little more effort.

> > I have been
> > playing with eggs lately and prefer them.
>
> Maybe it's time to look into that. How do you install an egg?

Once you install setuptools, there is a command line tool called
easy_install.  Its name doesn't do it justice.  You can pass easy
install a module name and it will look on the cheeseshop.  You can
pass a source url and it will download, build, and install it.  You
can pass it a local directory pointing to a source.  You get the
pattern.  It has a lot of logic built in and does the right thing 99%
of the time.

> > I usually stick with Apple's bundled python.  I also build for wx and
> > tk since both come with tiger.
>
> Are you building Tiger only then? Damn, there are a lot of python+OS-X
> combinations to support! Also the wx version that comes with Tiger is
> pretty old, so I'm not sure how helpful that is. Boy, this is a pain!

I don't have access to 10.3.

> I really think the Python-OS-X community needs to have a fairly unified
> pile of packages to install. pythonmac.org is a good place for them,
> only because it's there. My goal is to have a MPL that works with the
> other packages there, so you can do one stop shopping. I'm running 10.3,
> so I currently build versions for Apple's 2.3 and Bob I's 2.4.1. Both of
> those have Numeric, numarray and wxPython packages on that site. Those
> need some updating, but they are there.

I am willing to help, but osx is not my primary/only target.  This is
another reason I like eggs.  They work great on osx/linux/windows.

Hi all,

I'm trying to install mpl (I've tried both 0.86 and 0.85) on an x86_64=20
machine running Fedora Core 4. I saw one post related to gtk but it=20
referred to a problem different from mine. Below is the last few lines=20
from stderr...

 > python setup.py build

....

In file included from src/_nc_backend_gdk.c:12:
/usr/include/pygtk-2.0/pygtk/pygtk.h:8:21: error: gtk/gtk.h: No such=20
file or directory
In file included from src/_nc_backend_gdk.c:12:
/usr/include/pygtk-2.0/pygtk/pygtk.h:14: error: syntax error before =91at=
om=92
/usr/include/pygtk-2.0/pygtk/pygtk.h:16: error: syntax error before=20
=91GdkRectangle=92
/usr/include/pygtk-2.0/pygtk/pygtk.h:23: error: syntax error before=20
=91GdkAtom=92
/usr/include/pygtk-2.0/pygtk/pygtk.h:23: warning: no semicolon at end of=20
struct or union
/usr/include/pygtk-2.0/pygtk/pygtk.h:24: warning: data definition has no=20
type or storage class
src/_nc_backend_gdk.c: In function =91pixbuf_get_pixels_array=92:
src/_nc_backend_gdk.c:32: error: =91GdkPixbuf=92 undeclared (first use in=
=20
this function)
src/_nc_backend_gdk.c:32: error: (Each undeclared identifier is reported=20
only once
src/_nc_backend_gdk.c:32: error: for each function it appears in.)
src/_nc_backend_gdk.c:32: error: =91gdk_pixbuf=92 undeclared (first use i=
n=20
this function)
src/_nc_backend_gdk.c:52: warning: cast to pointer from integer of=20
different size
error: command 'gcc' failed with exit status 1

Now the second line up above says I need gtk.h, which much to my=20
surprise is nowhere on my system. I have gtk2 and gtk+ devel rpms=20
installed and numerous other gtk related packages including pygtk. Is it=20
possible that pygtk installed correctly without gtk or am I missing=20
something subtle (or completely obvious).

Thanks for any help,
Derrick

--=20
Derrick Snowden
PhOD/AOML/NOAA
4301 Rickenbacker Causeway
Miami, FL 33149
Office: 305.361.4322  Fax: 305.361.4392