matplotlib-users Mailing List for matplotlib

On 2005年10月19日, Jeff Whitaker wrote:
> Oh well. I'm satisified then that there really isn't anything out
> there with acceptable license terms. I'm going to post my natgrid-based
> griddata module on the wiki and call it done (for now).
My view is that it is always worth asking for a different
license before giving up. Many people choose 
licenses thoughtlessly. Some not, of course.
Cheers,
Alan Isaac
PS I tried to find current contact info for the natgrid author
and failed.

Robert Kern wrote:
>Jeff Whitaker wrote:
>
> 
>
>>The LICENSE reads:
>>
>>Bivariate Cubic Spline approximation library + standalone utility
>>Version 1.14
>>
>>Copyright 2002 CSIRO Marine Research
>>GPO 1538 Hobart
>>TAS 7001
>>Australia
>>Please send comments and bugs to Pav...@cs...
>>
>>There is no warranty whatsoever. Use at your own risk.
>>
>>These code may be freely redistributed under the condition that the
>>copyright
>>notices are not removed. You may distribute modified versions of this code
>>UNDER THE CONDITION THAT THIS CODE AND ANY MODIFICATIONS MADE TO IT IN THE
>>SAME FILE REMAIN UNDER COPYRIGHT OF CSIRO, BOTH SOURCE AND OBJECT CODE ARE
>>MADE FREELY AVAILABLE WITHOUT CHARGE, AND CLEAR NOTICE IS GIVEN OF THE
>>MODIFICATIONS.
>>
>>If this sounds OK I'll investigate further.
>> 
>>
>
>No, the CAPITALIZED part is very much not okay. It means we'd have to
>sign over the copyright to our modifications to CSIRO, and that the code
>can't be embedded in a commercial or proprietary project.
>
> 
>
Oh well. I'm satisified then that there really isn't anything out 
there with acceptable license terms. I'm going to post my natgrid-based 
griddata module on the wiki and call it done (for now).
-Jeff
-- 
Jeffrey S. Whitaker Phone : (303)497-6313
Meteorologist FAX : (303)497-6449
NOAA/OAR/CDC R/CDC1 Email : Jef...@no...
325 Broadway Office : Skaggs Research Cntr 1D-124
Boulder, CO, USA 80303-3328 Web : http://tinyurl.com/5telg

Jeff Whitaker wrote:
> The LICENSE reads:
> 
> Bivariate Cubic Spline approximation library + standalone utility
> Version 1.14
> 
> Copyright 2002 CSIRO Marine Research
> GPO 1538 Hobart
> TAS 7001
> Australia
> Please send comments and bugs to Pav...@cs...
> 
> There is no warranty whatsoever. Use at your own risk.
> 
> These code may be freely redistributed under the condition that the
> copyright
> notices are not removed. You may distribute modified versions of this code
> UNDER THE CONDITION THAT THIS CODE AND ANY MODIFICATIONS MADE TO IT IN THE
> SAME FILE REMAIN UNDER COPYRIGHT OF CSIRO, BOTH SOURCE AND OBJECT CODE ARE
> MADE FREELY AVAILABLE WITHOUT CHARGE, AND CLEAR NOTICE IS GIVEN OF THE
> MODIFICATIONS.
> 
> If this sounds OK I'll investigate further.
No, the CAPITALIZED part is very much not okay. It means we'd have to
sign over the copyright to our modifications to CSIRO, and that the code
can't be embedded in a commercial or proprietary project.
-- 
Robert Kern
rk...@uc...
"In the fields of hell where the grass grows high
 Are the graves of dreams allowed to die."
 -- Richard Harter

John Hunter wrote:
>>>>>>"Todd" == Todd Miller <jm...@st...> writes:
>>>>>> 
>>>>>>
>
> Todd> Either an extension in your application forgot to call
> Todd> import_libnumarray() or the numarray distributed as part of
> Todd> your application is incomplete or broken and the import
> Todd> failed. numarray CVS now makes a better report of which is
> Todd> true.
>
>Could that someone be us?
> 
>
I don't think so. For matplotlib we use import_array() because that's 
the init call for numarray's Numeric compatible interface. I think 
Tim's import_libnumarray() message is originating from numarray as it 
tries to call it's own API, most likely because the numarray 
installation is broken so the import failed. 
Here's a couple other facts:
1. The import_libnumarray() message is confusing because I didn't think 
numarray import failures would occur in a sane world... so import 
failures in numarray-1.3.3 look the same as uninitialized extensions. 
2. numarray uses it's own API to share code internally, so when 
numarray's rather complex import fails, the first internal call to it's 
own API generates the cryptic fatal error.
numarray CVS does two new things that are about a week old:
1. It correctly reports numarray's own import failure with a fatal 
error. This should clarify future problems if there are any.
2. Explicit imports of core extensions were added to numarray.__init__ 
so that py2exe can figure out what extensions must be included in an 
installer. This should actually solve the problem.
Those fixes address "exactly the same symptoms" (py2exe + cryptic fatal 
unitialized API error message) seen by another user who tested the fix 
on two different installer tools.
>I tried
>
> > grep -ri import_libnumarray .
>
>in my src tree and got zippo.
>
>We call 
>
> import_array(); 
>
>....
>
>Is this a matplotlib bug. If so, I'll be happy to patch it and roll
>out a bug-fix release.
> 
>
Nope, don't think so. 
Todd

>>>>> "Todd" == Todd Miller <jm...@st...> writes:
 Todd> Either an extension in your application forgot to call
 Todd> import_libnumarray() or the numarray distributed as part of
 Todd> your application is incomplete or broken and the import
 Todd> failed. numarray CVS now makes a better report of which is
 Todd> true.
Could that someone be us?
I tried
 > grep -ri import_libnumarray .
in my src tree and got zippo.
We call 
 import_array(); 
....
Is this a matplotlib bug. If so, I'll be happy to patch it and roll
out a bug-fix release.
JDH

John Hunter wrote:
>>>>>> "Robert" == Robert Kern <rk...@uc...> writes:
>>>>>> 
>
> Robert> In any case, natgrid isn't Delaunay triangulation, but
> Robert> natural neighbors interpolation which is better (and
> Robert> probably state of the art for these kinds of graphics). It
> Robert> might be easiest to contact Dave Watson and ask if we can
> Robert> distribute his code in matplotlib under matplotlib's
> Robert> license. I can't find any contact information for him,
> Robert> however.
>
> I'm all for this but am not optimistic. It's my understanding that
> the NCAR folks are under a fairly heavy legal burden. Jeff: could you
> spearhead this effort and see if there is any interest on their part
> in relicensing their griddata code under a more permissive license?
>
> JDH
>
> 
John: I will do this and let you know. I agree with Robert on qhull - 
it's not the way to go. I did find this however:
http://www.marine.csiro.au/~sakov/
Bivariate spline-based techniques offer robust and efficient 
approximation solutions. Here `csa' (standing for "cubic spline 
approximation") is a small library for approximating 2D irregular 
(scattered) data with bivariate C1-continuous cubic splines 
<http://www.math.uni-mannheim.de/%7Elsmath4/paper/VIS2001.pdf.gz>. For 
uniformly distributed data this algorithm is clearly superior than 
Natural Neighbours interpolation: it is faster (O(n)!), more robust 
(approximation, not interpolation) and better performing (C1!). (But 
take care when dealing with clustered data.) From version 1.09, `csa' 
can take advantage of the known variance of the data.
The LICENSE reads:
Bivariate Cubic Spline approximation library + standalone utility
Version 1.14
Copyright 2002 CSIRO Marine Research
GPO 1538 Hobart
TAS 7001
Australia
Please send comments and bugs to Pav...@cs...
There is no warranty whatsoever. Use at your own risk.
These code may be freely redistributed under the condition that the 
copyright
notices are not removed. You may distribute modified versions of this code
UNDER THE CONDITION THAT THIS CODE AND ANY MODIFICATIONS MADE TO IT IN THE
SAME FILE REMAIN UNDER COPYRIGHT OF CSIRO, BOTH SOURCE AND OBJECT CODE ARE
MADE FREELY AVAILABLE WITHOUT CHARGE, AND CLEAR NOTICE IS GIVEN OF THE
MODIFICATIONS.
If this sounds OK I'll investigate further.
-Jeff
-- 
Jeffrey S. Whitaker Phone : (303)497-6313
NOAA/OAR/CDC R/CDC1 FAX : (303)497-6449
325 Broadway Web : http://www.cdc.noaa.gov/~jsw
Boulder, CO, USA 80305-3328 Office: Skaggs Research Cntr 1D-124

Tim Burgess wrote:
> Hi,
> Still tryin' to freeze my wxmpl matplotlib application.... sigh.
>
> This is wxpython + wxmpl + matplotlib. This has turned into one hell 
> of a fight - BUT I think I am about to win. You have to keep the faith 
> :-}
>
> Name the py2exe issue with freezing matplotlib / wxmpl and I have been 
> there fought that. I have slayed many dragons...
>
> I now have a problem that has been asked before on this list. I have 
> tried the answer that was offered- it didnt help. I think I am in a 
> postion to ask a simple solid question...
>
> When I run my application I see this....
>
> Fatal Python error: Call to API function without first calling 
> import_libnumarray() in Src\_convmodule.c
Either an extension in your application forgot to call 
import_libnumarray() or the numarray distributed as part of your 
application is incomplete or broken and the import failed. numarray CVS 
now makes a better report of which is true.
> I added import numarray all over the place... a previous post 
> suggested that maybe Wxpython was calling numarray before I was... no 
> I dont think wxpython and numarray ever need each other.
>
> To help prove this I changed my application to this: (a boa app - so 
> this code occurs early in execution but not before wxpython has been 
> used...
>
> def __init__(self, parent):
>
> import numarray
> import numarray.numarrayall # desperate attempt to stop error 
> api call without first calling import_libnumarray
> self._init_ctrls(parent) # init wxpython controls
>
> print numarray.__version__
>
> and bang my application no longer runs. If I comment out this print 
> line, the application runs perfectly except when you click the button 
> to draw the graph. Then you get the same error message as above.
>
> BTW in the development environments - application works fine - and 
> that's because the development environs are loading the modules from 
> the python installation - is my call...
My guess is that in the development environments numarray is already 
correctly installed. 
>
> In my library file you can see that numarray is definitely included - 
> it's there.
>
> So my question is:
> Q: How do you force the loading (not the inclusion) of the nummary module?
I may not understand you, but when numarray is all there, "import 
numarray" should work just like for any other Python package. I think 
your problem is that numarray is not being included correctly in your 
installer because it's extension dependencies are hard for py2exe to 
detect correctly. I believe this is fixed now in numarray CVS.
> Or how do you get import_libnumarray to run?
import_libnumarray() is something an extension writer must explicitly 
call in their extension init() if they want to use the numarray API.
> Can some one tell me more about the mechanism python uses to load the 
> modules.. What triggers the load call? Where?
>
> I know I am close. Rest of the application is 95% finished. Customer 
> loves it. 2 week delay coz I cant freeze the @#$@$%@ thing.
Try using numarray CVS to generate your installer. Last week someone 
had a similar issue with this and I added code to numarray.__init__ to 
explicitly import all numarray's core extensions. That user reported 
that the extra "hints" in numarray.__init__ helped py2exe detect the 
necessary numarray extension set correctly, numarray was thus installed 
correctly, and the installer worked.
Regards,
Todd

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_ima=
ge.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.
--
Alex Borghgraef

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
 <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Suddenly I feel less alone.<br>
<br>
.<br>
I'm engaged in exactly the same process right now,
I've tried moving the whole of numarray into&nbsp; the&nbsp; dist a&nbsp; dir out of
sheer desperation....but no go- my traceback looks like this.....<br>
<br>
Traceback (most recent call last):<br>
Traceback (most recent call last):<br>
&nbsp; File "HacGui.py", line 21, in ?<br>
&nbsp; File "activity.pyc", line 17, in ?<br>
&nbsp; File "matplotlib\axes.pyc", line 5, in ?<br>
&nbsp; File
"C:\DOCUME~1\ROBMAR~1\REPO\Hydra2\hydra\admin\client\HAC\build\dist\__init__.pyc",
line 95, in ?<br>
<br>
ImportError: No module named random_array<br>
<br>
As soon as I remove numarray fropm the dist dir this resolves to <br>
<br>
<br>
Fatal Python error: Call to API function without first calling
import_libnumarray() in Src\_convmodule.c<br>
<br>
I came across a post yesterday&nbsp;
<a class="moz-txt-link-freetext" href="http://aspn.activestate.com/ASPN/Mail/Message/numpy-discussion/2852995">http://aspn.activestate.com/ASPN/Mail/Message/numpy-discussion/2852995</a>&nbsp;
that suggests that it's possible to <br>
'force' the whole of numarray into an app. I'd like to know how though<br>
<br>
<b>Russell E. Owen wrote:
<span class="mail_quotation">&gt; If I convert my python code to an
application (Windows via py2exe or Mac </span><span
 class="mail_quotation">&gt; via bundlebuilder) it fails with the
following error:</span>
<span class="mail_quotation">&gt; </span><span class="mail_quotation">&gt;
Fatal Python error: Call to API function without first calling </span><span
 class="mail_quotation">&gt; import_libnumarray() in Src/_convmodule.c</span>
<span class="mail_quotation">&gt; </span><span class="mail_quotation">&gt;
</span>This currently (1.3.3) happens when a numarray API function is
called before the API is successfully initialized. Although the message
was intended as an aid to extension writers, in this case numarray is
failing to import altogether. At one point numarray had a fatal error
for import failures but I removed it at someone's request. I've
restored it because I think it's most commonly fatal anyway and
removing the message just obfuscated the problem. The non-fatal
behavior is now in the _import_libnumarray() macro.
<span class="mail_quotation">&gt; I can force *all* of numarray into
the application, which avoids the </span><span class="mail_quotation">&gt;
</span><span class="mail_quotation">&gt; </span>
This is what you need to do. It should be possible to factor out (or
not explicitly list) numarray's Packages, but core numarray is not
meant to be distributed in pieces. The many type-specific extensions
were only added to work around a compiler problem, not to lighten
binary distributions.
.</b>
<br>
<br>
<br>
<br>
<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ('.',
glob.glob(r'C:\PROGRA~1\Python24\share\MATPLO~1\matplotlibrc')),<br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ('.',
glob.glob(r'C:\PROGRA~1\Python24\Lib\SITE-P~1\numarray\*.py')), <br>
<br>
.....seems to take care of an earlier error ( can't find matplotlibrc
file) <br>
<br>
You have my complete empathy on this problem. In the event I come up
with some sort of magic bullet, I'll mail you ASAP.<br>
<br>
Have fun <span class="moz-smiley-s1"><span> :-) </span></span><br>
<br>
Rob M.<br>
&nbsp;<br>
<br>
<br>
Tim Burgess wrote:
<blockquote cite="mid...@tb..." type="cite">Hi,
 <br>
Still tryin' to freeze my wxmpl matplotlib application.... sigh.
 <br>
 <br>
This is wxpython + wxmpl + matplotlib. This has turned into one hell of
a fight - BUT I think I am about to win. You have to keep the faith :-}
 <br>
 <br>
Name the py2exe issue with freezing matplotlib / wxmpl and I have been
there fought that. I have slayed many dragons...
 <br>
 <br>
I now have a problem that has been asked before on this list. I have
tried the answer that was offered- it didnt help. I think I am in a
postion to ask a simple solid question...
 <br>
 <br>
When I run my application I see this....
 <br>
 <br>
Fatal Python error: Call to API function without first calling
import_libnumarray() in Src\_convmodule.c
 <br>
 <br>
This application has requested the Runtime to terminate it in an
unusual way.
 <br>
Please contact the application's support team for more information.
 <br>
 <br>
I added import numarray all over the place... a previous post suggested
that maybe Wxpython was calling numarray before I was... no I dont
think wxpython and numarray ever need each other.
 <br>
 <br>
To help prove this I changed my application to this: (a boa app - so
this code occurs early in execution but not before wxpython has been
used...
 <br>
 <br>
&nbsp;&nbsp; def __init__(self, parent):
 <br>
 <br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; import numarray
 <br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; import numarray.numarrayall # desperate attempt to stop error
api call without first calling import_libnumarray
 <br>
&nbsp; &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self._init_ctrls(parent)&nbsp;&nbsp; # init wxpython controls
 <br>
 <br>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print numarray.__version__
 <br>
 <br>
and bang my application no longer runs. If I comment out this print
line, the application runs perfectly except when you click the button
to draw the graph. Then you get the same error message as above.
 <br>
 <br>
BTW in the development environments - application works fine - and
that's because the development environs are loading the modules from
the python installation - is my call...
 <br>
 <br>
In my library file you can see that numarray is definitely included -
it's there.
 <br>
 <br>
So my question is:
 <br>
Q: How do you force the loading (not the inclusion) of the nummary
module? Or how do you get import_libnumarray to run?
 <br>
 <br>
Can some one tell me more about the mechanism python uses to load the
modules.. What triggers the load call? Where?
 <br>
 <br>
I know I am close. Rest of the application is 95% finished. Customer
loves it. 2 week delay coz I cant freeze the @#$@$%@ thing.
 <br>
 <br>
BTW kudos to py2exe developers and maintainers - as I learn more about
the challenges in identifying the modules and the packaging of them etc
- I appreciate it.&nbsp; This is not a sucking up effort to get an answer -
but if it helps.... :-}
 <br>
 <br>
thanks
 <br>
 <br>
PS I have cross posted this to matploblib and py2exe lists - I am
desperate!
 <br>
 <br>
</blockquote>
<br>
<pre class="moz-signature" cols="72">-- 
 O__
 _/`.\
 `=( '
</pre>
</body>
</html>

Dear Kristen,
Make sure that in Gsview, EPS Clip is "ticked"
See under Options -> EPS Clip
The problem is that the EPS picture in not an A4 (or letter) canvas but
on a bigger canvas...
By editing the EPS, you'll find a line that read something like, e.g.:
...
%%BoundingBox: 162 744 1171 1753
...
The left boundingbox margin is set at 1171/72=3D16.26in (413.1mm), much
wider than an A4 (or letter) Media type
You can also set the Media size to, say, A0...
See under Media -> A0 or User Defined...
But for an EPS picture, EPS Clip would suffice...
Regards,
Peter-Jan
PS:
Because of the width of the EPS picture in LaTeX, you would need a line
like, e.g.
...
\centerline{\includegraphics[width=3D16cm]{myepsfile.eps}}
...
To scale the EPS picture for A4 paper (depending on you margins or
number of columns etc.)
-----Original Message-----
From: Kristen Kaasbjerg [mailto:co...@fy...]=20
Sent: 19 October 2005 10:00
To: Randewijk P-J, Mnr <pjr...@su...>
Subject: Re: [Matplotlib-users] savefig(epsfilename) Bug on XP
Hi again
This solved the problem, however, the figure is
placed in the corner of the eps files, implying that
only the bottom left corner of the figure is actually
visible. Any idea why ?
Kristen
----- Original Message -----=20
From: <pjr...@su...>
To: "Kristen Kaasbjerg" <co...@fy...>
Cc: <jdh...@ac...>
Sent: Wednesday, October 19, 2005 9:44 AM
Subject: RE: [Matplotlib-users] savefig(epsfilename) Bug on XP
Dear Kristen,
Try this one.
I'll put a post on the mailinglist as well.
Kind regards,
Peter-Jan
-----Original Message-----
From: Kristen Kaasbjerg [mailto:co...@fy...]=20
Sent: 19 October 2005 09:30
To: Randewijk P-J, Mnr <pjr...@su...>; John Hunter
Cc: mat...@li...
Subject: Re: [Matplotlib-users] savefig(epsfilename) Bug on XP
Hi folks
I can still not succeed saving my figures in ps or eps format with the
new ps backend file you send out. I'm using python 2.4, mpl .84. Have
tried with both Numeric and numarray.
Kristen
----- Original Message -----=20
From: "John Hunter" <jdh...@ac...>
To: <pjr...@su...>
Cc: <mat...@li...>
Sent: Wednesday, October 19, 2005 4:20 AM
Subject: Re: [Matplotlib-users] savefig(epsfilename) Bug on XP
>>>>>> ""Randewijk" =3D=3D "Randewijk P-J, Mnr <pjr...@su...>"=20
>>>>>> <pjr...@su...> writes:
>
> "Randewijk> Dear (newbie like myself) XP or more specific win32
> "Randewijk> users, The problem was not with the 'dvips' command,
> "Randewijk> but with the 'latex' command in version 0.83.2.
>
> "Randewijk> After changing the line 1086 in backend_ps.py 0.83.2
> "Randewijk> to read:
>
> I've made the suggested changes to backend_ps.py. Because I don't
> have access to a win32 box with TeX, could I ask win32 / TeX users to=20
> test the attached backend_ps file to make sure it is OK?
>
>=20

Hi,
Still tryin' to freeze my wxmpl matplotlib application.... sigh.
This is wxpython + wxmpl + matplotlib. This has turned into one hell of 
a fight - BUT I think I am about to win. You have to keep the faith :-}
Name the py2exe issue with freezing matplotlib / wxmpl and I have been 
there fought that. I have slayed many dragons...
I now have a problem that has been asked before on this list. I have 
tried the answer that was offered- it didnt help. I think I am in a 
postion to ask a simple solid question...
When I run my application I see this....
Fatal Python error: Call to API function without first calling 
import_libnumarray() in Src\_convmodule.c
This application has requested the Runtime to terminate it in an unusual 
way.
Please contact the application's support team for more information.
I added import numarray all over the place... a previous post suggested 
that maybe Wxpython was calling numarray before I was... no I dont think 
wxpython and numarray ever need each other.
To help prove this I changed my application to this: (a boa app - so 
this code occurs early in execution but not before wxpython has been used...
 def __init__(self, parent):
 import numarray
 import numarray.numarrayall # desperate attempt to stop error 
api call without first calling import_libnumarray
 
 self._init_ctrls(parent) # init wxpython controls
 
 print numarray.__version__
and bang my application no longer runs. If I comment out this print 
line, the application runs perfectly except when you click the button to 
draw the graph. Then you get the same error message as above.
BTW in the development environments - application works fine - and 
that's because the development environs are loading the modules from the 
python installation - is my call...
In my library file you can see that numarray is definitely included - 
it's there.
So my question is:
Q: How do you force the loading (not the inclusion) of the nummary 
module? Or how do you get import_libnumarray to run?
Can some one tell me more about the mechanism python uses to load the 
modules.. What triggers the load call? Where?
I know I am close. Rest of the application is 95% finished. Customer 
loves it. 2 week delay coz I cant freeze the @#$@$%@ thing.
BTW kudos to py2exe developers and maintainers - as I learn more about 
the challenges in identifying the modules and the packaging of them etc 
- I appreciate it. This is not a sucking up effort to get an answer - 
but if it helps.... :-}
thanks
PS I have cross posted this to matploblib and py2exe lists - I am desperate!
-- 
Tim Burgess
IT Consultant
RedHat Certified Engineer
TBITC Pty Ltd
Professional Computer Support for Business
ti...@tb...
Mobile 0422 942 972
Office 85 662 016
http://www.tbitc.com

Dear win32 users,
Line 1089 in the file Johan send out, needs to be changed to:
 command =3D 'latex -interaction=3Dnonstopmode "%s"' % =
texfile
Or see attached...
Kind regards,
Peter-Jan Randewijk
-----Original Message-----
From: John Hunter [mailto:jdh...@ac...]=20
Sent: 19 October 2005 04:20
To: Randewijk P-J, Mnr <pjr...@su...>
Cc: mat...@li...
Subject: Re: [Matplotlib-users] savefig(epsfilename) Bug on XP
>>>>> ""Randewijk" =3D=3D "Randewijk P-J, Mnr <pjr...@su...>"
>>>>> <pjr...@su...> writes:
 "Randewijk> Dear (newbie like myself) XP or more specific win32
 "Randewijk> users, The problem was not with the 'dvips' command,
 "Randewijk> but with the 'latex' command in version 0.83.2.
 "Randewijk> After changing the line 1086 in backend_ps.py 0.83.2
 "Randewijk> to read:
I've made the suggested changes to backend_ps.py. Because I don't have
access to a win32 box with TeX, could I ask win32 / TeX users to test
the attached backend_ps file to make sure it is OK?

Hi folks
I can still not succeed saving my figures in ps or eps format
with the new ps backend file you send out.
I'm using python 2.4, mpl .84. Have tried with both Numeric
and numarray.
Kristen
----- Original Message ----- 
From: "John Hunter" <jdh...@ac...>
To: <pjr...@su...>
Cc: <mat...@li...>
Sent: Wednesday, October 19, 2005 4:20 AM
Subject: Re: [Matplotlib-users] savefig(epsfilename) Bug on XP
>>>>>> ""Randewijk" == "Randewijk P-J, Mnr <pjr...@su...>" 
>>>>>> <pjr...@su...> writes:
>
> "Randewijk> Dear (newbie like myself) XP or more specific win32
> "Randewijk> users, The problem was not with the 'dvips' command,
> "Randewijk> but with the 'latex' command in version 0.83.2.
>
> "Randewijk> After changing the line 1086 in backend_ps.py 0.83.2
> "Randewijk> to read:
>
> I've made the suggested changes to backend_ps.py. Because I don't
> have access to a win32 box with TeX, could I ask win32 / TeX users to
> test the attached backend_ps file to make sure it is OK?
>
> 

Chris Barker wrote:
> Robert Kern wrote:
> 
>> qhull, for example, is a real PITA in this regard.
> 
> Darn, it looked pretty promising from the web site.
It's not. It's well-known to be a bad library. There's a Python
interface to qhull, but it uses pipes to communicate with a separate
qhull process.
>> In any case, natgrid isn't Delaunay triangulation, but natural neighbors
>> interpolation which is better (and probably state of the art for these
>> kinds of graphics).
> 
> Although there are other uses for delaunay triangulization, so it would
> be nice to have a python lib that does it well.
> 
> What about Jonathan Shewchuk's Triangle? I don't know how clean the code
> is but it's fast and powerful. As for license:
> 
> """
> Please note that although Triangle is freely available, it is
> copyrighted by the author and may not be sold or included in commercial
> products without a license.
> """
> 
> I wish he'd just a pick a license and be done with it, but given that
> Python. MPL, and SciPy are not commercial, he'd probably be OK with it.
MPL and Scipy won't take code that can't be redistributed commercially
(much less proprietarily). There's no indication that Shewchuk will
allow that.
-- 
Robert Kern
rk...@uc...
"In the fields of hell where the grass grows high
 Are the graves of dreams allowed to die."
 -- Richard Harter

Robert Kern wrote:
> qhull, for example, is a real PITA in this regard.
Darn, it looked pretty promising from the web site.
> In any case, natgrid isn't Delaunay triangulation, but natural neighbors
> interpolation which is better (and probably state of the art for these
> kinds of graphics).
Although there are other uses for delaunay triangulization, so it would 
be nice to have a python lib that does it well.
What about Jonathan Shewchuk's Triangle? I don't know how clean the code 
is but it's fast and powerful. As for license:
"""
Please note that although Triangle is freely available, it is 
copyrighted by the author and may not be sold or included in commercial 
products without a license.
"""
I wish he'd just a pick a license and be done with it, but given that 
Python. MPL, and SciPy are not commercial, he'd probably be OK with it.
-Chris
-- 
Christopher Barker, Ph.D.
Oceanographer
 		
NOAA/OR&R/HAZMAT (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception
Chr...@no...

John Hunter wrote:
>>>>>>"Robert" == Robert Kern <rk...@uc...> writes:
> 
> Robert> In any case, natgrid isn't Delaunay triangulation, but
> Robert> natural neighbors interpolation which is better (and
> Robert> probably state of the art for these kinds of graphics). It
> Robert> might be easiest to contact Dave Watson and ask if we can
> Robert> distribute his code in matplotlib under matplotlib's
> Robert> license. I can't find any contact information for him,
> Robert> however.
> 
> I'm all for this but am not optimistic. It's my understanding that
> the NCAR folks are under a fairly heavy legal burden. Jeff: could you
> spearhead this effort and see if there is any interest on their part
> in relicensing their griddata code under a more permissive license?
It's not NCAR's code, it's Dave Watson's. NCAR got permission to
distribute the natgrid code under the GPL. We'd need to get independent
permission from Dave Watson to distribute the code under the matplotlib
license.
-- 
Robert Kern
rk...@uc...
"In the fields of hell where the grass grows high
 Are the graves of dreams allowed to die."
 -- Richard Harter

>>>>> "Robert" == Robert Kern <rk...@uc...> writes:
 Robert> In any case, natgrid isn't Delaunay triangulation, but
 Robert> natural neighbors interpolation which is better (and
 Robert> probably state of the art for these kinds of graphics). It
 Robert> might be easiest to contact Dave Watson and ask if we can
 Robert> distribute his code in matplotlib under matplotlib's
 Robert> license. I can't find any contact information for him,
 Robert> however.
I'm all for this but am not optimistic. It's my understanding that
the NCAR folks are under a fairly heavy legal burden. Jeff: could you
spearhead this effort and see if there is any interest on their part
in relicensing their griddata code under a more permissive license?
JDH

John Hunter wrote:
>>>>>>"Jeff" == Jeff Whitaker <js...@fa...> writes:
> 
> Jeff> John: I've now a got a toolkit built with no CDAT code in
> Jeff> it, using the NCAR natgrid library which is GPL. I've done a
> Jeff> fair amount of searching and haven't found anything else
> Jeff> that provides similar functionality with better
> Jeff> (i.e. BSD-like) licensing. The prototype natgrid toolkit is
> Jeff> at /Public/jsw/natgrid-0.1.tar.gz if you'd like to take a
> Jeff> look. All it does is provide a single function, griddata,
> Jeff> which works much the same as the matlab version. I'm ready
> Jeff> to put it CVS if you give the go ahead.
> 
> I'm hesitant to include any GPL code even as a toolkit. For one
> thing, it decreases the impetus for someone to provide a version that
> is compatible with the mpl license. Perhaps this would be better
> placed on the wiki?
> 
> I know that VTK is BSD compatible and has delaunay triangulation, but
> obviously it is impractical to try and pull anything out of VTK. I
> think the qhull license is less restrictive:
> http://www.qhull.org/COPYING.txt. In short, I wonder if we have
> really exhausted all the possibilities. Something from netlib?
> There was a recent thread on scipy in which Eric Jones posted a link
> to some code (FORTRAN I believe) that does something analogous. It
> would be nice to have something in C/C++ and something that is BSD
> compatible. Perhaps I'm dreaming.
I've spent time looking. Everything that I've found with a BSD-like
license isn't a very good library. Global variables are usually the
culprit. qhull, for example, is a real PITA in this regard. The code
that Eric found was from ACM TOMS, I believe, and so I'd question its
license.
In any case, natgrid isn't Delaunay triangulation, but natural neighbors
interpolation which is better (and probably state of the art for these
kinds of graphics). It might be easiest to contact Dave Watson and ask
if we can distribute his code in matplotlib under matplotlib's license.
I can't find any contact information for him, however.
-- 
Robert Kern
rk...@uc...
"In the fields of hell where the grass grows high
 Are the graves of dreams allowed to die."
 -- Richard Harter

>>>>> "Willi" == Willi Richert <w.r...@gm...> writes:
 >> Does it matter if you put the x and y label calls *after* the
 >> errorbar command?
 Willi> Unfortunately, no.
OK, at this point all I can suggest is that you post your data file
with the complete script so that we can see if we can reproduce it on
our end. If we can reproduce it, there is a good chance we can fix
it.
Thanks,
JDH

>>>>> ""Randewijk" == "Randewijk P-J, Mnr <pjr...@su...>" <pjr...@su...> writes:
 "Randewijk> Dear (newbie like myself) XP or more specific win32
 "Randewijk> users, The problem was not with the 'dvips' command,
 "Randewijk> but with the 'latex' command in version 0.83.2.
 "Randewijk> After changing the line 1086 in backend_ps.py 0.83.2
 "Randewijk> to read:
I've made the suggested changes to backend_ps.py. Because I don't
have access to a win32 box with TeX, could I ask win32 / TeX users to
test the attached backend_ps file to make sure it is OK?

>>>>> "Jeff" == Jeff Whitaker <js...@fa...> writes:
 Jeff> John: I've now a got a toolkit built with no CDAT code in
 Jeff> it, using the NCAR natgrid library which is GPL. I've done a
 Jeff> fair amount of searching and haven't found anything else
 Jeff> that provides similar functionality with better
 Jeff> (i.e. BSD-like) licensing. The prototype natgrid toolkit is
 Jeff> at /Public/jsw/natgrid-0.1.tar.gz if you'd like to take a
 Jeff> look. All it does is provide a single function, griddata,
 Jeff> which works much the same as the matlab version. I'm ready
 Jeff> to put it CVS if you give the go ahead.
I'm hesitant to include any GPL code even as a toolkit. For one
thing, it decreases the impetus for someone to provide a version that
is compatible with the mpl license. Perhaps this would be better
placed on the wiki?
I know that VTK is BSD compatible and has delaunay triangulation, but
obviously it is impractical to try and pull anything out of VTK. I
think the qhull license is less restrictive:
http://www.qhull.org/COPYING.txt. In short, I wonder if we have
really exhausted all the possibilities. Something from netlib?
There was a recent thread on scipy in which Eric Jones posted a link
to some code (FORTRAN I believe) that does something analogous. It
would be nice to have something in C/C++ and something that is BSD
compatible. Perhaps I'm dreaming.
JDH