matplotlib-users Mailing List for matplotlib (Page 11)

This is a known bug in ESP Ghostscript, not matplotlib. ESP version 8.15.02 
broke xpdf's pdftops program. I posted a bug report to the ESP website, and 
it has been fixed in their sources, but they havent made a new release since 
that time. Your options are either to drop back to ESP version 7.07 or use 
gpl or afpl ghostscript.

Darren


On Monday 10 July 2006 08:58, Pieter Dumon wrote:
> Sorry, forgot the actual PS file, which contains the actual error. Here it
> is
>
> On 7/10/06, Pieter Dumon <pie...@gm...> wrote:
> > On 7/10/06, Darren Dale <dd...@co...> wrote:
> > > Pieter Dumon wrote:
> > > > For example, this fails for me (it does work without xpdf and doing
> > > > ps2pdf afterwards):
> > > >
> > > > import matplotlib
> > > > matplotlib.use('PS')
> > > > import pylab
> > > >
> > > > pylab.rc('text',usetex=True)
> > > > pylab.rc('ps',usedistiller='xpdf')
> > > > pylab.figure()
> > > > pylab.plot((1,2,3),(4,5,6))
> > > > pylab.savefig("test")
> > > > pylab.show()
> > > >
> > > > When setting ps.usedistiller to xpdf, I get an empty (blank) ps
> > > > image. Without xpdf, I get a proper ps. This did work with
> > > > 0.87-2/numpy-0.9.6 (The Pdf backend works, but not with mathttext.)
> > >
> > > Please set your verbose level to debug-annoying in your rc settings and
> > > post the output. Also post your platform info, along with version info
> > > for xpdf, and ghostscript. I havent had any problems with xpdf.
> >
> > The debug output is in attachment. I don't see any errors, except for
> > the latex warning (but the problem is the same without usetex=True,
> > apparently)
> >
> > ESP Ghostscript 815.02 (2006-04-19)
> > xpdf version 3.01
> > poppler-0.5.3
> >
> > System is a quite standard Gentoo Linux (x86), but with gcc-4.1.1
> > 2.6.16-gentoo-r9 #1 PREEMPT Tue Jun 6 10:38:13 CEST 2006 i686 Mobile
> > Intel(R) Pentium(R) 4 - M CPU 2.00GHz GNU/Linux
> >
> > Pieter

-- 
Darren S. Dale, Ph.D.
Cornell High Energy Synchrotron Source
Cornell University
200L Wilson Lab
Rt. 366 & Pine Tree Road
Ithaca, NY 14853

dd...@co...
office: (607) 255-9894
fax: (607) 255-9001

Sorry, forgot the actual PS file, which contains the actual error. Here it is


On 7/10/06, Pieter Dumon <pie...@gm...> wrote:
> On 7/10/06, Darren Dale <dd...@co...> wrote:
> > Pieter Dumon wrote:
> > > For example, this fails for me (it does work without xpdf and doing
> > > ps2pdf afterwards):
>
> > > import matplotlib
> > > matplotlib.use('PS')
> > > import pylab
> > >
> > > pylab.rc('text',usetex=True)
> > > pylab.rc('ps',usedistiller='xpdf')
> > > pylab.figure()
> > > pylab.plot((1,2,3),(4,5,6))
> > > pylab.savefig("test")
> > > pylab.show()
> > >
> > > When setting ps.usedistiller to xpdf, I get an empty (blank) ps image.
> > > Without xpdf, I get a proper ps. This did work with 0.87-2/numpy-0.9.6
> > > (The Pdf backend works, but not with mathttext.)
> > >
> > Please set your verbose level to debug-annoying in your rc settings and
> > post the output. Also post your platform info, along with version info
> > for xpdf, and ghostscript. I havent had any problems with xpdf.
>
> The debug output is in attachment. I don't see any errors, except for
> the latex warning (but the problem is the same without usetex=True,
> apparently)
>
> ESP Ghostscript 815.02 (2006-04-19)
> xpdf version 3.01
> poppler-0.5.3
>
> System is a quite standard Gentoo Linux (x86), but with gcc-4.1.1
> 2.6.16-gentoo-r9 #1 PREEMPT Tue Jun 6 10:38:13 CEST 2006 i686 Mobile
> Intel(R) Pentium(R) 4 - M CPU 2.00GHz GNU/Linux
>
> Pieter
>
>
>

On 7/10/06, Darren Dale <dd...@co...> wrote:
> Pieter Dumon wrote:
> > For example, this fails for me (it does work without xpdf and doing
> > ps2pdf afterwards):

> > import matplotlib
> > matplotlib.use('PS')
> > import pylab
> >
> > pylab.rc('text',usetex=True)
> > pylab.rc('ps',usedistiller='xpdf')
> > pylab.figure()
> > pylab.plot((1,2,3),(4,5,6))
> > pylab.savefig("test")
> > pylab.show()
> >
> > When setting ps.usedistiller to xpdf, I get an empty (blank) ps image.
> > Without xpdf, I get a proper ps. This did work with 0.87-2/numpy-0.9.6
> > (The Pdf backend works, but not with mathttext.)
> >
> Please set your verbose level to debug-annoying in your rc settings and
> post the output. Also post your platform info, along with version info
> for xpdf, and ghostscript. I havent had any problems with xpdf.

The debug output is in attachment. I don't see any errors, except for
the latex warning (but the problem is the same without usetex=True,
apparently)

ESP Ghostscript 815.02 (2006-04-19)
xpdf version 3.01
poppler-0.5.3

System is a quite standard Gentoo Linux (x86), but with gcc-4.1.1
2.6.16-gentoo-r9 #1 PREEMPT Tue Jun 6 10:38:13 CEST 2006 i686 Mobile
Intel(R) Pentium(R) 4 - M CPU 2.00GHz GNU/Linux

Pieter

Hi Fernando,

I tried opening your example of a bad eps file produced with the xpdf option, 
and it looks identical to the good example you sent. (I opened it with both 
ggv, and gv.)

Darren

On Sunday 09 July 2006 23:33, Fernando Perez wrote:
> On 7/9/06, Darren Dale <dd...@co...> wrote:
> > Would you send me an example of a file with a bad bbox and one that is
> > correct?
>
> Thanks for looking into this!
>
> f

-- 
Darren S. Dale, Ph.D.
Cornell High Energy Synchrotron Source
Cornell University
200L Wilson Lab
Rt. 366 & Pine Tree Road
Ithaca, NY 14853

dd...@co...
office: (607) 255-9894
fax: (607) 255-9001

Pieter Dumon wrote:
> Could the bug I submitted some time ago be related to this ? (1502970
> at http://sourceforge.net/tracker/?group_id=80706&atid=560720)
>
> For example, this fails for me (it does work without xpdf and doing
> ps2pdf afterwards):
> My poppler/xpdf installation should be allright
>
> import matplotlib
> matplotlib.use('PS')
> import pylab
>
> pylab.rc('text',usetex=True)
> pylab.rc('ps',usedistiller='xpdf')
> pylab.figure()
> pylab.plot((1,2,3),(4,5,6))
> pylab.savefig("test")
> pylab.show()
>
> ---
> Summary:
> matplotlib-0.87.3 with xpdf renders empty PS
> upraded to matplotlib-0.87-3/numpy-0.9.8.
>
> When setting ps.usedistiller to xpdf, I get an empty (blank) ps image.
> Without xpdf, I get a proper ps. This did work with 0.87-2/numpy-0.9.6
> (The Pdf backend works, but not with mathttext.)
>   
Please set your verbose level to debug-annoying in your rc settings and 
post the output. Also post your platform info, along with version info 
for xpdf, and ghostscript. I havent had any problems with xpdf.

Could the bug I submitted some time ago be related to this ? (1502970
at http://sourceforge.net/tracker/?group_id=80706&atid=560720)

For example, this fails for me (it does work without xpdf and doing
ps2pdf afterwards):
My poppler/xpdf installation should be allright

import matplotlib
matplotlib.use('PS')
import pylab

pylab.rc('text',usetex=True)
pylab.rc('ps',usedistiller='xpdf')
pylab.figure()
pylab.plot((1,2,3),(4,5,6))
pylab.savefig("test")
pylab.show()

---
Summary:
matplotlib-0.87.3 with xpdf renders empty PS
upraded to matplotlib-0.87-3/numpy-0.9.8.

When setting ps.usedistiller to xpdf, I get an empty (blank) ps image.
Without xpdf, I get a proper ps. This did work with 0.87-2/numpy-0.9.6
(The Pdf backend works, but not with mathttext.)

---

Bye,
Pieter


On 7/10/06, Fernando Perez <fpe...@gm...> wrote:
> On 7/9/06, Darren Dale <dd...@co...> wrote:
> > Would you send me an example of a file with a bad bbox and one that is
> > correct?
>
> Thanks for looking into this!
>
> f
>
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>
>
>

On 7/9/06, Darren Dale <dd...@co...> wrote:
> Would you send me an example of a file with a bad bbox and one that is
> correct?

Thanks for looking into this!

f

Would you send me an example of a file with a bad bbox and one that is 
correct?

On Sunday 09 July 2006 10:28 pm, Fernando Perez wrote:
> On 7/9/06, Darren Dale <dd...@co...> wrote:
> > I think this is artifact results from a combination of the ghostscript
> > distiller and the postscript renderer. If you have xpdf on your system,
> > you can set ps.usedistiller : xpdf in rc. That will give you embeddable
> > eps files without converting the fonts to outlines (or whatever
> > ghostscript does), so they scale nicely. As a bonus, if you put the
> > figure in a pdf, the figure text will be searchable.
>
> Thanks for the tip!  The result looks indeed very nice, but
> unfortunately the bounding box is all wrong (using ps.distiller=False
> gives eps files with the proper bounding box).  With an incorrect
> bounding box, these eps files are kind of useless (not completely,
> since ps2eps can correct this, but it's annoying to have to fix every
> plot you make manually...)
>
>
> Any ideas?
>
> Cheers,
>
> f
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier Download IBM WebSphere Application Server v.1.0.1 based on Apache
> Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

-- 
Darren S. Dale, Ph.D.
dd...@co...

On 7/9/06, Darren Dale <dd...@co...> wrote:

> I think this is artifact results from a combination of the ghostscript
> distiller and the postscript renderer. If you have xpdf on your system, you
> can set ps.usedistiller : xpdf in rc. That will give you embeddable eps files
> without converting the fonts to outlines (or whatever ghostscript does), so
> they scale nicely. As a bonus, if you put the figure in a pdf, the figure
> text will be searchable.

Thanks for the tip!  The result looks indeed very nice, but
unfortunately the bounding box is all wrong (using ps.distiller=False
gives eps files with the proper bounding box).  With an incorrect
bounding box, these eps files are kind of useless (not completely,
since ps2eps can correct this, but it's annoying to have to fix every
plot you make manually...)


Any ideas?

Cheers,

f

Wolfgang,

Will you please try using these new files? I think they should solve the 
problems you were having. Please post some feedback as soon as possible. 
Thanks.

Darren


On Monday 12 June 2006 1:18 pm, Wolfgang wrote:
> Hi,
>
> I've changed back all my changes but:
>
>   if exit_status: raise RuntimeError('dvipng was not able to \
> RuntimeError: dvipng was not able to process the flowing file:
> C:\Documents and
> Settings\s0167070\.matplotlib\tex.cache\3bb1bc57100b0c65915bdda3a1f60dae.dv
>i Here is the full report generated by dvipng:

Hello,

It looks like there are a couple of typos in the equations used in the=20
function.

On Friday 07 July 2006 12:17, Faheem Mitha wrote:
> I think the bivariate normal probability distribution function
> implementation in Matplotlib is buggy. See the example code below, and
> note that the change in the x directions and the y directions is not
> uniform as it should be.
> The implementation (in mlab.py) is appended.
[snip]
> =A0 =A0 =A0z =3D Xmu**2/sigmax**2 + Ymu**2/sigmay - 2*rho*Xmu*Ymu/(sigmax=
*sigmay)
                                           ^ **2
> =A0 =A0 =A0return 1.0/(2*pi*sigmax*sigmay*(1-rho**2)) * exp( -z/(2*(1-rho=
**2)))
                                      ^ sqrt()

Compare to equations (1) and (2) from the link in the docstring:

http://mathworld.wolfram.com/BivariateNormalDistribution.html

I've attached a patch.

Mike

On Sunday 09 July 2006 12:53 am, you wrote:
> > The fix took a little more work in order to address a problem Wolfgang
> > reported with win32 a while back. Both texmanager and backend_ps have
> > been improved and tested on a bleeding edge gentoo system and a windows
> > system.
> >
> > Iff you have time, would you please update to svn 2550 and test the new
> > usetex code? Otherwise, I can test it on a RHEL4 system Monday morning.
>
> The joys of X11, I can actually test this (the fedora box is in my
> office).  It works just fine, thanks a lot!
>
> But I did notice something a bit odd, by accident.  I made up a random
> title for a test plot with various symbols, and the exported eps file
> shows a strange thing: look at the integral limits in the attached
> eps.  The -\infty is smooth (as is all the rest of the latex), while
> the \beta is all jaggy.  With ubuntu's gv it's not super visible until
> you zoom in, but with Fedora's it really stands out.
>
> Not a big issue, but I did find it a bit strange.  You know a lot more
> about this, so the reason behind it may be obvious to you.

I think this is artifact results from a combination of the ghostscript 
distiller and the postscript renderer. If you have xpdf on your system, you 
can set ps.usedistiller : xpdf in rc. That will give you embeddable eps files 
without converting the fonts to outlines (or whatever ghostscript does), so 
they scale nicely. As a bonus, if you put the figure in a pdf, the figure 
text will be searchable.

Darren

On 7/8/06, Darren Dale <dd...@co...> wrote:
> Hi Fernando,
>
> On Friday 07 July 2006 10:36 pm, you wrote:
> > On 7/7/06, Darren Dale <dd...@co...> wrote:
> > > Thanks Fernando. I'll work on this in the morning. I've been boozing all
> > > evening...
> >
> > Enjoy the evening, and in the morning, watch out for the
> >
> > cd foo && command
> >
> > syntax I used.  It's valid bash, but I don't know about win32...
>
> The fix took a little more work in order to address a problem Wolfgang
> reported with win32 a while back. Both texmanager and backend_ps have been
> improved and tested on a bleeding edge gentoo system and a windows system.
>
> Iff you have time, would you please update to svn 2550 and test the new usetex
> code? Otherwise, I can test it on a RHEL4 system Monday morning.

The joys of X11, I can actually test this (the fedora box is in my
office).  It works just fine, thanks a lot!

But I did notice something a bit odd, by accident.  I made up a random
title for a test plot with various symbols, and the exported eps file
shows a strange thing: look at the integral limits in the attached
eps.  The -\infty is smooth (as is all the rest of the latex), while
the \beta is all jaggy.  With ubuntu's gv it's not super visible until
you zoom in, but with Fedora's it really stands out.

Not a big issue, but I did find it a bit strange.  You know a lot more
about this, so the reason behind it may be obvious to you.

In any case, many thanks for getting this up and running again!

Cheers,

f

Hi Fernando,

On Friday 07 July 2006 10:36 pm, you wrote:
> On 7/7/06, Darren Dale <dd...@co...> wrote:
> > Thanks Fernando. I'll work on this in the morning. I've been boozing all
> > evening...
>
> Enjoy the evening, and in the morning, watch out for the
>
> cd foo && command
>
> syntax I used.  It's valid bash, but I don't know about win32...

The fix took a little more work in order to address a problem Wolfgang 
reported with win32 a while back. Both texmanager and backend_ps have been 
improved and tested on a bleeding edge gentoo system and a windows system. 

Iff you have time, would you please update to svn 2550 and test the new usetex 
code? Otherwise, I can test it on a RHEL4 system Monday morning.

Darren

On Friday 07 July 2006 3:30 pm, Eric Firing wrote:
> Martin,
>
> When I try your example with svn matplotlib, I get a 34 MB eps file, and
> looking at it, I don't see much room for making it smaller--there is one
> obvious optimization, abbreviating "marker", but that's it.

Thanks for the suggestion. the abbreviation is "o" as of svn 2548.

On 7/7/06, Darren Dale <dd...@co...> wrote:
> Thanks Fernando. I'll work on this in the morning. I've been boozing all
> evening...

Enjoy the evening, and in the morning, watch out for the

cd foo && command

syntax I used.  It's valid bash, but I don't know about win32...

Take care,

f

Thanks Fernando. I'll work on this in the morning. I've been boozing all 
evening...


On Friday 07 July 2006 6:43 pm, Fernando Perez wrote:
> On 7/7/06, Fernando Perez <fpe...@gm...> wrote:
> > On 7/7/06, Fernando Perez <fpe...@gm...> wrote:
> > > Don't worry about me: I'll fix the mpl sources myself, that's easy
> > > enough as a workaround.  But you may want to take care of this before
> > > the new release is out, which John mentioned this morning as coming
> > > soon.
> >
> > I got it to work by changing in texmanager.py, line 203, from:
> >
> >             command = 'latex -interaction=nonstopmode
> > -output-directory="%s" \ "%s" > "%s"'%(self.texcache, texfile, outfile)
> >
> > to:
> >
> >             command = 'cd "%s" && latex -interaction=nonstopmode "%s"
> >
> > > "%s"' % \
> >
> >                       (self.texcache,texfile, outfile)
> >
> > This will work with any version of latex.
>
> But a similar change is also needed for EPS generation.  I've just
> attached a full patch, in case you want it.
>
> Cheers,
>
> f

-- 
Darren S. Dale, Ph.D.
dd...@co...

On FC4 (and FC5) we've got version 3.0 of tetex:
$ rpm -q tetex
tetex-3.0-10.FC4

So it looks like the Badger's dragging the chain in this case :-)

Fernando Perez wrote:

>But aside from my own troubles, I suspect that if you guys release
>0.87.4 with this bug, I won't be the only one complaining :)  There
>are probably other systems out there still not using tetex-3.0 (I
>think Red Hat enterprise is fairly conservative in its upgrade policy,
>and I don't know what Fedora 4 has, though Fedora 5 is likely OK).
>  
>

On 7/7/06, Fernando Perez <fpe...@gm...> wrote:
> On 7/7/06, Fernando Perez <fpe...@gm...> wrote:
>
> > Don't worry about me: I'll fix the mpl sources myself, that's easy
> > enough as a workaround.  But you may want to take care of this before
> > the new release is out, which John mentioned this morning as coming
> > soon.
>
> I got it to work by changing in texmanager.py, line 203, from:
>
>             command = 'latex -interaction=nonstopmode -output-directory="%s" \
>                        "%s" > "%s"'%(self.texcache, texfile, outfile)
>
> to:
>
>             command = 'cd "%s" && latex -interaction=nonstopmode "%s"
> > "%s"' % \
>                       (self.texcache,texfile, outfile)
>
> This will work with any version of latex.

But a similar change is also needed for EPS generation.  I've just
attached a full patch, in case you want it.

Cheers,

f

On 7/7/06, Fernando Perez <fpe...@gm...> wrote:

> Don't worry about me: I'll fix the mpl sources myself, that's easy
> enough as a workaround.  But you may want to take care of this before
> the new release is out, which John mentioned this morning as coming
> soon.

I got it to work by changing in texmanager.py, line 203, from:

            command = 'latex -interaction=nonstopmode -output-directory="%s" \
                       "%s" > "%s"'%(self.texcache, texfile, outfile)

to:

            command = 'cd "%s" && latex -interaction=nonstopmode "%s"
> "%s"' % \
                      (self.texcache,texfile, outfile)

This will work with any version of latex.

Cheers,

f

On 7/7/06, Darren Dale <dd...@co...> wrote:
> Hi Fernando,
>
> On Friday 07 July 2006 17:56, Fernando Perez wrote:
> > Hi all,
> >
> > with a current SVN checkout, I'm having serious latex problems (run
> > with debug-annoying):
>
> I'm sorry for the trouble. I tried checking my recent changes on windows and
> Linux, and didnt have any trouble, but I didnt try with an older version of
> latex. It turns out that version 3.14159 does not support the --output-dir
> option. Is updating your tetex an option? I'm using tetex-3.0:

No, not really.  I'm trying to finish some work to free up a day or
two so I can update all of our systems to Ubuntu Dapper, but that
means I can't really burn any time on patching pieces of the system.

But aside from my own troubles, I suspect that if you guys release
0.87.4 with this bug, I won't be the only one complaining :)  There
are probably other systems out there still not using tetex-3.0 (I
think Red Hat enterprise is fairly conservative in its upgrade policy,
and I don't know what Fedora 4 has, though Fedora 5 is likely OK).

Don't worry about me: I'll fix the mpl sources myself, that's easy
enough as a workaround.  But you may want to take care of this before
the new release is out, which John mentioned this morning as coming
soon.

Cheers,

f

Hi Fernando,

On Friday 07 July 2006 17:56, Fernando Perez wrote:
> Hi all,
>
> with a current SVN checkout, I'm having serious latex problems (run
> with debug-annoying):

I'm sorry for the trouble. I tried checking my recent changes on windows and 
Linux, and didnt have any trouble, but I didnt try with an older version of 
latex. It turns out that version 3.14159 does not support the --output-dir 
option. Is updating your tetex an option? I'm using tetex-3.0:

latex --version
pdfeTeX 3.141592-1.30.5-2.2 (Web2C 7.5.5)
kpathsea version 3.5.5
Copyright 2005 Peter Breitenlohner (eTeX)/Han The Thanh (pdfTeX).
Kpathsea is copyright 2005 Karl Berry and Olaf Weber.
There is NO warranty.  Redistribution of this software is
covered by the terms of both the pdfeTeX copyright and
the Lesser GNU General Public License.
For more information about these matters, see the file
named COPYING and the pdfeTeX source.
Primary author of pdfeTeX: Peter Breitenlohner (eTeX)/Han The Thanh (pdfTeX).
Kpathsea written by Karl Berry, Olaf Weber, and others.

> In [5]: plot([1,2,3])
> FigureCanvasAgg.draw
> RendererAgg._get_agg_font
>         findfont failed New Century Schoolbook
>         findfont failed Century Schoolbook L
>         findfont failed Utopia
>         findfont failed ITC Bookman
>         findfont failed Bookman
>         findfont found Bitstream Vera Serif, normal, normal 500, normal,
> 12.0 findfont returning
> /home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/mpl-data/Vera
>Se.ttf Found dvipng version 1.8
> latex -interaction=nonstopmode
> -output-directory="/home/fperez/.matplotlib/tex.cache"
>       
> "/home/fperez/.matplotlib/tex.cache/7ba2a6d994e57f6ad218b4d10097765e.tex"
>
> > "/home/fperez/.matplotlib/tex.cache/7ba2a6d994e57f6ad218b4d10097765e.outp
> >ut"
>
> latex: unrecognized option
> `-output-directory=/home/fperez/.matplotlib/tex.cache'
> Try `latex --help' for more information.
> ---------------------------------------------------------------------------
> exceptions.RuntimeError                              Traceback (most
> recent call last)
>
> /home/fperez/research/papers/nsform/scripts/<ipython console>
>
> /home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/pylab.py
> in plot(*args=([1, 2, 3],), **kwargs={})
>    2021     except:
>    2022         hold(b)
> -> 2023         raise
>    2024
>    2025     hold(b)
>
> [... snip ...]
>
>
> /home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/text.py
> in _get_layout(self=<matplotlib.text.TextWithDash instance>,
> renderer=<matplotlib.backends.backend_agg.RendererAgg instance>)
>     184         heightt += 3  # 3 pixel pad
>     185         for line in lines:
> --> 186             w,h = renderer.get_text_width_height(
>         w = undefined
>         h = undefined
>         renderer.get_text_width_height = <bound method
> RendererAgg.get_text_width_height of
> <matplotlib.backends.backend_agg.RendererAgg instance at 0x4183ac4c>>
>         line = '$0$'
>         self._fontproperties = <matplotlib.font_manager.FontProperties
> instance at 0x418402cc>
>         global ismath = undefined
>         self.is_math_text = <bound method TextWithDash.is_math_text of
> <matplotlib.text.TextWithDash instance at 0x4184030c>>
>     187                 line, self._fontproperties,
> ismath=self.is_math_text()) 188
>
> /home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/backends/back
>end_agg.py in
> get_text_width_height(self=<matplotlib.backends.backend_agg.RendererAgg
> instance>, s='$0$', prop=<matplotlib.font_manager.FontProperties
> instance>, ismath='TeX', rgb=(0, 0, 0))
>     237             size = prop.get_size_in_points()
>     238             texmanager = self.get_texmanager()
> --> 239             Z = texmanager.get_rgba(s, size, self.dpi.get(), rgb)
>         Z = undefined
>         texmanager.get_rgba = <bound method TexManager.get_rgba of
> <matplotlib.texmanager.TexManager instance at 0x41bebd2c>>
>         s = '$0$'
>         size = 12.0
>         self.dpi.get = <built-in method get of tuple object at 0x41736dcc>
>         rgb = (0, 0, 0)
>     240             m,n,tmp = Z.shape
>     241             return n,m
>
> /home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/texmanager.py
> in get_rgba(self=<matplotlib.texmanager.TexManager instance>,
> tex='$0$', fontsize=12.0, dpi=100.0, rgb=(0, 0, 0))
>     333         if Z is None:
>     334             # force=True to skip cacheing while debugging
> --> 335             pngfile = self.make_png(tex, fontsize, dpi,
> force=False) pngfile = undefined
>         self.make_png = <bound method TexManager.make_png of
> <matplotlib.texmanager.TexManager instance at 0x41bebd2c>>
>         tex = '$0$'
>         fontsize = 12.0
>         dpi = 100.0
>         global force = undefined
>         global False = undefined
>     336             X = readpng(os.path.join(self.texcache, pngfile))
>     337
>
> /home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/texmanager.py
> in make_png(self=<matplotlib.texmanager.TexManager instance>,
> tex='$0$', fontsize=12.0, dpi=100.0, force=False)
>     235         # see get_rgba for a discussion of the background
>     236         if force or not os.path.exists(pngfile):
> --> 237             dvifile = self.make_dvi(tex, fontsize)
>         dvifile = undefined
>         self.make_dvi = <bound method TexManager.make_dvi of
> <matplotlib.texmanager.TexManager instance at 0x41bebd2c>>
>         tex = '$0$'
>         fontsize = 12.0
>     238             outfile = basefile+'.output'
>     239             command = 'dvipng -bg Transparent -D %s -T tight
> -o "%s" "%s" \
>
> /home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/texmanager.py
> in make_dvi(self=<matplotlib.texmanager.TexManager instance>,
> tex='$0$', fontsize=12.0, force=0)
>     207             fh = file(outfile)
>     208             if exit_status:
> --> 209                 raise RuntimeError('LaTeX was not able to
> process the flowing \
>         global RuntimeError = undefined
>         tex = '$0$'
>         fh.read = <built-in method read of file object at 0x418375a0>
>     210 string:\n%s\nHere is the full report generated by LaTeX:
> \n\n'% tex + fh.read())
>     211             else: verbose.report(fh.read(), 'debug')
>
> RuntimeError: LaTeX was not able to process the flowing string:
> $0$
> Here is the full report generated by LaTeX:
>
>
>
> In [6]:
>
>
> My matplotlibrc file has (this is what I think is relevant):
>
> backend      : TkAgg    # the default backend
> numerix      : numpy  # numpy, Numeric or numarray
> text.usetex         : True  # use latex for all text handling.  For
> more information, see
>                              #
> http://www.scipy.org/Wiki/Cookbook/Matplotlib/UsingTex
> text.dvipnghack     : False  # some versions of dvipng don't handle
>
>
> This is running on a Fedora Core 3 box.  Yes, I'll be upgrading to
> Ubuntu very soon, but in the meantime it would be nice to be able to
> use mpl :)
>
> My latex info:
>
> planck[scripts]> latex --version
> TeX (Web2C 7.4.5) 3.14159
> kpathsea version 3.4.5
> Copyright (C) 1997-2003 D.E. Knuth.
>
>
> Cheers,
>
> f
>
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier Download IBM WebSphere Application Server v.1.0.1 based on Apache
> Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

-- 
Darren S. Dale, Ph.D.
Cornell High Energy Synchrotron Source
Cornell University
200L Wilson Lab
Rt. 366 & Pine Tree Road
Ithaca, NY 14853

dd...@co...
office: (607) 255-9894
fax: (607) 255-9001

Hi all,

with a current SVN checkout, I'm having serious latex problems (run
with debug-annoying):

In [5]: plot([1,2,3])
FigureCanvasAgg.draw
RendererAgg._get_agg_font
        findfont failed New Century Schoolbook
        findfont failed Century Schoolbook L
        findfont failed Utopia
        findfont failed ITC Bookman
        findfont failed Bookman
        findfont found Bitstream Vera Serif, normal, normal 500, normal, 12.0
findfont returning
/home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/mpl-data/VeraSe.ttf
Found dvipng version 1.8
latex -interaction=nonstopmode
-output-directory="/home/fperez/.matplotlib/tex.cache"
       "/home/fperez/.matplotlib/tex.cache/7ba2a6d994e57f6ad218b4d10097765e.tex"
> "/home/fperez/.matplotlib/tex.cache/7ba2a6d994e57f6ad218b4d10097765e.output"
latex: unrecognized option
`-output-directory=/home/fperez/.matplotlib/tex.cache'
Try `latex --help' for more information.
---------------------------------------------------------------------------
exceptions.RuntimeError                              Traceback (most
recent call last)

/home/fperez/research/papers/nsform/scripts/<ipython console>

/home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/pylab.py
in plot(*args=([1, 2, 3],), **kwargs={})
   2021     except:
   2022         hold(b)
-> 2023         raise
   2024
   2025     hold(b)

[... snip ...]


/home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/text.py
in _get_layout(self=<matplotlib.text.TextWithDash instance>,
renderer=<matplotlib.backends.backend_agg.RendererAgg instance>)
    184         heightt += 3  # 3 pixel pad
    185         for line in lines:
--> 186             w,h = renderer.get_text_width_height(
        w = undefined
        h = undefined
        renderer.get_text_width_height = <bound method
RendererAgg.get_text_width_height of
<matplotlib.backends.backend_agg.RendererAgg instance at 0x4183ac4c>>
        line = '$0$'
        self._fontproperties = <matplotlib.font_manager.FontProperties
instance at 0x418402cc>
        global ismath = undefined
        self.is_math_text = <bound method TextWithDash.is_math_text of
<matplotlib.text.TextWithDash instance at 0x4184030c>>
    187                 line, self._fontproperties, ismath=self.is_math_text())
    188

/home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/backends/backend_agg.py
in get_text_width_height(self=<matplotlib.backends.backend_agg.RendererAgg
instance>, s='$0$', prop=<matplotlib.font_manager.FontProperties
instance>, ismath='TeX', rgb=(0, 0, 0))
    237             size = prop.get_size_in_points()
    238             texmanager = self.get_texmanager()
--> 239             Z = texmanager.get_rgba(s, size, self.dpi.get(), rgb)
        Z = undefined
        texmanager.get_rgba = <bound method TexManager.get_rgba of
<matplotlib.texmanager.TexManager instance at 0x41bebd2c>>
        s = '$0$'
        size = 12.0
        self.dpi.get = <built-in method get of tuple object at 0x41736dcc>
        rgb = (0, 0, 0)
    240             m,n,tmp = Z.shape
    241             return n,m

/home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/texmanager.py
in get_rgba(self=<matplotlib.texmanager.TexManager instance>,
tex='$0$', fontsize=12.0, dpi=100.0, rgb=(0, 0, 0))
    333         if Z is None:
    334             # force=True to skip cacheing while debugging
--> 335             pngfile = self.make_png(tex, fontsize, dpi, force=False)
        pngfile = undefined
        self.make_png = <bound method TexManager.make_png of
<matplotlib.texmanager.TexManager instance at 0x41bebd2c>>
        tex = '$0$'
        fontsize = 12.0
        dpi = 100.0
        global force = undefined
        global False = undefined
    336             X = readpng(os.path.join(self.texcache, pngfile))
    337

/home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/texmanager.py
in make_png(self=<matplotlib.texmanager.TexManager instance>,
tex='$0$', fontsize=12.0, dpi=100.0, force=False)
    235         # see get_rgba for a discussion of the background
    236         if force or not os.path.exists(pngfile):
--> 237             dvifile = self.make_dvi(tex, fontsize)
        dvifile = undefined
        self.make_dvi = <bound method TexManager.make_dvi of
<matplotlib.texmanager.TexManager instance at 0x41bebd2c>>
        tex = '$0$'
        fontsize = 12.0
    238             outfile = basefile+'.output'
    239             command = 'dvipng -bg Transparent -D %s -T tight
-o "%s" "%s" \

/home/fperez/tmp/local/lib/python2.3/site-packages/matplotlib/texmanager.py
in make_dvi(self=<matplotlib.texmanager.TexManager instance>,
tex='$0$', fontsize=12.0, force=0)
    207             fh = file(outfile)
    208             if exit_status:
--> 209                 raise RuntimeError('LaTeX was not able to
process the flowing \
        global RuntimeError = undefined
        tex = '$0$'
        fh.read = <built-in method read of file object at 0x418375a0>
    210 string:\n%s\nHere is the full report generated by LaTeX:
\n\n'% tex + fh.read())
    211             else: verbose.report(fh.read(), 'debug')

RuntimeError: LaTeX was not able to process the flowing string:
$0$
Here is the full report generated by LaTeX:



In [6]:


My matplotlibrc file has (this is what I think is relevant):

backend      : TkAgg    # the default backend
numerix      : numpy  # numpy, Numeric or numarray
text.usetex         : True  # use latex for all text handling.  For
more information, see
                             #
http://www.scipy.org/Wiki/Cookbook/Matplotlib/UsingTex
text.dvipnghack     : False  # some versions of dvipng don't handle


This is running on a Fedora Core 3 box.  Yes, I'll be upgrading to
Ubuntu very soon, but in the meantime it would be nice to be able to
use mpl :)

My latex info:

planck[scripts]> latex --version
TeX (Web2C 7.4.5) 3.14159
kpathsea version 3.4.5
Copyright (C) 1997-2003 D.E. Knuth.


Cheers,

f

>>>>> "Eric" == Eric Firing <ef...@ha...> writes:

    Eric> Martin, When I try your example with svn matplotlib, I get a
    Eric> 34 MB eps file, and looking at it, I don't see much room for
    Eric> making it smaller--there is one obvious optimization,
    Eric> abbreviating "marker", but that's it.  (The svg file is 456
    Eric> MB!)  So, maybe some major optimization has already been
    Eric> done between mpl 0.87.2 and svn.

Yep, Darren got "draw_markers" properly implemented for backend PS.
This function is much better in time and space; I believe only *Agg
and PS implement it, but it could be ported over to SVG fairly easily
by modifying the PS implementation.

    Eric> The bigger problem is that each file format has basic
    Eric> characteristics and limitations.  If you draw a million
    Eric> markers and line segments, you are inevitably going to have
    Eric> a big postscript file, unless the postscript backend somehow
    Eric> detects the fact that almost all of your points are
    Eric> indistinguishable and therefore deletes most of them--and
    Eric> this is really asking too much of a plotting backend, I

Agg does this for draw_lines -- it drops points in the path that are
less one pixel away from the previous point.  

JDH

Martin,

When I try your example with svn matplotlib, I get a 34 MB eps file, and 
looking at it, I don't see much room for making it smaller--there is one 
obvious optimization, abbreviating "marker", but that's it.  (The svg 
file is 456 MB!)  So, maybe some major optimization has already been 
done between mpl 0.87.2 and svn.

The bigger problem is that each file format has basic characteristics 
and limitations.  If you draw a million markers and line segments, you 
are inevitably going to have a big postscript file, unless the 
postscript backend somehow detects the fact that almost all of your 
points are indistinguishable and therefore deletes most of them--and 
this is really asking too much of a plotting backend, I think.  (An 
alternative is to generate a pixel image and make the postscript from 
that; this is what matlab does under some circumstances, but it can 
result in big files of poor quality.)

Your options include: filter your points beforehand so you only plot 
points that are distinct; or use a pixel-based format like png, which 
keeps the file size under control.

Eric

Martin Manns wrote:
> Hi,
> 
> When I use matplotlib for a scatter plot with both dots and connecting lines,
> the exported eps file is huge, if the distances between many points are small.
> I think of this as a bug, since no preview tiff is included in the generated eps
> and a variety of text processing applications (including OpenOffice) crash
> when I try to import the eps. Ghostscript takes forever, too. Is there anything
> that I can do in order to export reasonable eps files?
> 
> I am using:
> python 2.4.2
> matplotlib 0.87.2
> numpy 0.9.8
> 
> with Linux
> 
> The following small example illustrates the problem:
> --
> import pylab,numpy,random
> random.seed()
> x=[random.gauss(0,1)/float(i)**2 for i in xrange(1,1000000)]
> X=numpy.array(x,numpy.Float32)
> pylab.plot(X[1:],X[:-1],"-", c="#eeeeee")
> pylab.plot(X[1:],X[:-1],"xk")
> pylab.show()
> --
> 
> The resulting eps file:
> -rw-r----- 1 xx users 212190257 Jul  4 09:39 image.eps
> 
> Thanks a lot in advance
> 
> Martin
> 
> 
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users