matplotlib-users Mailing List for matplotlib (Page 7)

On Tue, Aug 12, 2008 at 5:16 AM, Zainal Abidin <zai...@gm...> wrote:
> I have a trouble when displat a PNG file which generated from matplotlib
> here is the script from
> http://www.scipy.org/Cookbook/Matplotlib/Using_MatPlotLib_in_a_CGI_script
> there is no image in the web browser, is there anything wrog with the
> script?
> please help me, I am a newbie in matplotlib ;)
One mistake I see is that you are calling use('Agg') after importing
pylab. According to the use docstring:
 Note: this function must be called *before* importing pylab for
 the first time; or, if you are not using pylab, it must be called
 before importing matplotlib.backends.
Not sure if this will solve your problem, but it should be fixed.
JDH

I have a trouble when displat a PNG file which generated from matplotlib
here is the script from
http://www.scipy.org/Cookbook/Matplotlib/Using_MatPlotLib_in_a_CGI_script
there is no image in the web browser, is there anything wrog with the
script?
please help me, I am a newbie in matplotlib ;)
--<start>
#!/usr/bin/python
import pylab
import matplotlib
import os,sys
import cgi
import cgitb
cgitb.enable()
os.environ['HOME']='/tmp/'
matplotlib.use('Agg')
form = cgi.FieldStorage()
pylab.plot([1,2,3])
print "Content-Type: image/png\n"
pylab.savefig(sys.stdout,format='png')
--</end>
and then here is my computer specification
CentOS 5.1 -- Linux localhost.localdomain 2.6.18-53.el5 #1 SMP Mon Nov 12
02:22:48 EST 2007 i686 i686 i386 GNU/Linux
python-2.4.3-19.el5
numpy-1.0.3.1-1.el5.kb
scipy-0.6.0-6.el5
matplotlib 0.91.4
-- 
Zainal Abidin, S.Si
Sub Bidang Informasi Meteorologi Publik
Badan Meteorologi dan Geofisika
Jl. Angkasa I No. 2
Jakarta - Indonesia
Visit Indonesia Year 2008 - Celebrating 100 Years of National Awakening

On Mon, Aug 11, 2008 at 1:31 PM, Jonathan Helmus <jjh...@gm...> wrote:
>
[snip]
> instance...> which I use to add lines, set limits, etc but it doesn't
> have any of the pan functions (start_pan, end_pan, etc).
> self.canvas.figure.get_axes() (which is what seems to be used in the
> NavigationToolbar2 class) returns the same Subplot instance. Any
> suggestions of how to get a axes/Subplot instance which has the pan
> functions?
Hi Jonathan,
On earlier versions of Matplotlib, you can use the NavigationToolbar2
class without actually displaying it (the toolbar) in your program.
Instead you can use your custom buttons, keypresses, etc, to tie into
the class' methods. We've done that with our app, and it works quite
nicely. In fact, you can borrow the bits you need from that class to
create your own zoom/pan routines.
Cheers,
Anthony.

Sorry folks, I found my own problem. Looks like the axes pan functions 
were a recent addition to matplotlib and the version I'm using, 0.91.2, 
doesn't have these. Time to bug the system admin to update to the 0.98 
version.
-Jonathan Helmus
Jonathan Helmus wrote:
> I have been working on a GUI application which uses matplotlib as a wx 
> object. I'd like to include a pan button on the toolbar of the 
> application but can't seem to get an axes object which has the pan 
> functions (which I'd like to use to tie into the pan code). I have a 
> self.axes object that identifies itself as <matplotlib.axes.Subplot 
> instance...> which I use to add lines, set limits, etc but it doesn't 
> have any of the pan functions (start_pan, end_pan, etc). 
> self.canvas.figure.get_axes() (which is what seems to be used in the 
> NavigationToolbar2 class) returns the same Subplot instance. Any 
> suggestions of how to get a axes/Subplot instance which has the pan 
> functions?
>
> Thanks,
> -Jonathan Helmus
> 

On Mon, Aug 11, 2008 at 4:24 PM, nobody <dai...@bu...> wrote:
>
> Bump... Does no-one have a solution for this?
Sorry for not answering. There is no elegant way to do this so we
should add support. In the meantime, the following hack should work
 >>> figs = [manager.canvas.figure for manager in
pylab._pylab_helpers.Gcf.get_all_fig_managers()]
And you asked about testing to see if a figure exists -- do you mean
by figure number? If so, you could check a number against the "num"
attribute from the figure list
 >>> somenum in [fig.num for fig in figs]
JDH

Bump... Does no-one have a solution for this?
-----
I am a BugMeNot account and postings from this user may not always be from
the same person.
-- 
View this message in context: http://www.nabble.com/Figure-existence-test-tp18760841p18934165.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

On Mon, Aug 11, 2008 at 3:51 PM, Michael Droettboom <md...@st...> wrote:
> Christopher Barker wrote:
>> Is there a guide to translating code from the old to new Transforms
>> structure?
> Yes, in the API_CHANGES file, in the source distribution. (I'd link to
> SVN, but it seems Sourceforge's SVN browser is down.)
We keep a copy of this file at http://matplotlib.sf.net/API_CHANGES
that is updated with every release See also
http://matplotlib.sourceforge.net/MIGRATION.txt
JDH

On Mon, Aug 11, 2008 at 3:51 PM, Michael Droettboom <md...@st...> wrote:
> Christopher Barker wrote:
>> Is there a guide to translating code from the old to new Transforms
>> structure?
> Yes, in the API_CHANGES file, in the source distribution. (I'd link to
> SVN, but it seems Sourceforge's SVN browser is down.)
We keep a copy of this file at http://matplotlib.sf.net/API_CHANGES
that is updated with every release See also
http://matplotlib.sourceforge.net/MIGRATION.txt
JDH

Christopher Barker wrote:
> Is there a guide to translating code from the old to new Transforms 
> structure?
> 
Yes, in the API_CHANGES file, in the source distribution. (I'd link to 
SVN, but it seems Sourceforge's SVN browser is down.)
Cheers,
Mike

Christopher Barker wrote:
> have you ported it over yet? Can anyone else help out with a port?
Note: it looks pretty easy, if yu know what you are doing:
from matplotlib.transforms import Interval, Value
Then in the code, I inly see Interval and Value used here:
 self.rintv = Interval(Value(0), Value(1))
 self.rintd = Interval(Value(0), Value(1))
Is there a guide to translating code from the old to new Transforms 
structure?
-CHB
-- 
Christopher Barker, Ph.D.
Oceanographer
Emergency Response Division
NOAA/NOS/OR&R (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception
Chr...@no...

Thanks John, that worked perfectly (with JJ's correction). I'm fairly
new to python and hadn't considered writing a separate class for it.
-Jeffrey
On Mon, Aug 11, 2008 at 3:53 PM, Jae-Joon Lee <lee...@gm...> wrote:
> A minor comment.
> John's code may give incorrect results when exponents are negative.
> int() truncates a floating point argument towards zero, e.g.,
> int(-1.5) == -1 not -2. I guess calling floor() before int() will
> work.
>
> fx = int(np.floor(np.log(abs(val))/np.log(self._base) +0.5))
>
> -JJ
>
>
> On Mon, Aug 11, 2008 at 3:23 PM, John Hunter <jd...@gm...> wrote:
>> On Mon, Aug 11, 2008 at 1:45 PM, Jeffrey Fogel
>> <mat...@je...> wrote:
>>
>>> The two things I have been unable to figure out are how to add a major
>>> tick at all of the other magnitudes (those without a label) and how to
>>> change the format of the labels so that only the exponent is showing.
>>> I'm sure both of these are straightforward, but I have been unable to
>>> figure them out. Thanks for your help.
>>
>> You can accomplish both in one swoop using a custom formatter instead
>> of a custom locator. That is, instead of using a
>> LogLocator(base=100.0) which isn't giving you ticks where you want
>> them, use the default log locator, and make a custom formatter to
>> suppress strings where you don't want them and to use the format you
>> want. Eg, something like
>>
>> import numpy as np
>> import matplotlib.pyplot as plt
>> import matplotlib.ticker as ticker
>>
>> fig = plt.figure()
>> ax = fig.add_subplot(111)
>>
>> N = 1000
>> x = np.arange(N)
>> y = np.random.rand(N)*1e8
>>
>> ax.semilogy(x, y)
>>
>> class MyFormatter(ticker.LogFormatter):
>>
>> def __call__(self, val, pos=None):
>> fx = int(np.log(abs(val))/np.log(self._base) +0.5)
>> isDecade = self.is_decade(fx)
>> if not isDecade and self.labelOnlyBase:
>> return ''
>> if (fx%2)==1: # odd, skip
>> return ''
>>
>> return '%d'%fx
>>
>>
>> ax.yaxis.set_major_formatter(MyFormatter())
>> plt.show()
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>
>

Lionel Roubeyrie wrote:
> find with this message a modified version of windrose to be OO compliant. 
Lionel,
I've had this message of yours (from March), and finally had a chance to 
use it.
Unfortunately, it seems to use the old Transforms mechanism, so won't 
work with the latest MPL.
have you ported it over yet? Can anyone else help out with a port?
Thanks,
-Chris
-- 
Christopher Barker, Ph.D.
Oceanographer
Emergency Response Division
NOAA/NOS/OR&R (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception
Chr...@no...

I have been working on a GUI application which uses matplotlib as a wx 
object. I'd like to include a pan button on the toolbar of the 
application but can't seem to get an axes object which has the pan 
functions (which I'd like to use to tie into the pan code). I have a 
self.axes object that identifies itself as <matplotlib.axes.Subplot 
instance...> which I use to add lines, set limits, etc but it doesn't 
have any of the pan functions (start_pan, end_pan, etc). 
self.canvas.figure.get_axes() (which is what seems to be used in the 
NavigationToolbar2 class) returns the same Subplot instance. Any 
suggestions of how to get a axes/Subplot instance which has the pan 
functions?
Thanks,
-Jonathan Helmus

A minor comment.
John's code may give incorrect results when exponents are negative.
int() truncates a floating point argument towards zero, e.g.,
int(-1.5) == -1 not -2. I guess calling floor() before int() will
work.
 fx = int(np.floor(np.log(abs(val))/np.log(self._base) +0.5))
-JJ
On Mon, Aug 11, 2008 at 3:23 PM, John Hunter <jd...@gm...> wrote:
> On Mon, Aug 11, 2008 at 1:45 PM, Jeffrey Fogel
> <mat...@je...> wrote:
>
>> The two things I have been unable to figure out are how to add a major
>> tick at all of the other magnitudes (those without a label) and how to
>> change the format of the labels so that only the exponent is showing.
>> I'm sure both of these are straightforward, but I have been unable to
>> figure them out. Thanks for your help.
>
> You can accomplish both in one swoop using a custom formatter instead
> of a custom locator. That is, instead of using a
> LogLocator(base=100.0) which isn't giving you ticks where you want
> them, use the default log locator, and make a custom formatter to
> suppress strings where you don't want them and to use the format you
> want. Eg, something like
>
> import numpy as np
> import matplotlib.pyplot as plt
> import matplotlib.ticker as ticker
>
> fig = plt.figure()
> ax = fig.add_subplot(111)
>
> N = 1000
> x = np.arange(N)
> y = np.random.rand(N)*1e8
>
> ax.semilogy(x, y)
>
> class MyFormatter(ticker.LogFormatter):
>
> def __call__(self, val, pos=None):
> fx = int(np.log(abs(val))/np.log(self._base) +0.5)
> isDecade = self.is_decade(fx)
> if not isDecade and self.labelOnlyBase:
> return ''
> if (fx%2)==1: # odd, skip
> return ''
>
> return '%d'%fx
>
>
> ax.yaxis.set_major_formatter(MyFormatter())
> plt.show()
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

On Mon, Aug 11, 2008 at 1:45 PM, Jeffrey Fogel
<mat...@je...> wrote:
> The two things I have been unable to figure out are how to add a major
> tick at all of the other magnitudes (those without a label) and how to
> change the format of the labels so that only the exponent is showing.
> I'm sure both of these are straightforward, but I have been unable to
> figure them out. Thanks for your help.
You can accomplish both in one swoop using a custom formatter instead
of a custom locator. That is, instead of using a
LogLocator(base=100.0) which isn't giving you ticks where you want
them, use the default log locator, and make a custom formatter to
suppress strings where you don't want them and to use the format you
want. Eg, something like
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
fig = plt.figure()
ax = fig.add_subplot(111)
N = 1000
x = np.arange(N)
y = np.random.rand(N)*1e8
ax.semilogy(x, y)
class MyFormatter(ticker.LogFormatter):
 def __call__(self, val, pos=None):
 fx = int(np.log(abs(val))/np.log(self._base) +0.5)
 isDecade = self.is_decade(fx)
 if not isDecade and self.labelOnlyBase:
 return ''
 if (fx%2)==1: # odd, skip
 return ''
 return '%d'%fx
ax.yaxis.set_major_formatter(MyFormatter())
plt.show()

I am having some trouble getting the ticks and labels on my plot to
match what I want and I was hoping that someone here would be able to
help. Here are what I want (on the yaxis):
logarithmic ticks with major ticks every order of magnitude but labels
ONLY every other magnitude. Additionally, I would like the labels to
be only the exponent (2 instead of 1e2 or 10^2). For the minor ticks,
I want a tick at all 9 values between the major ticks.
Here is what I have so far. For the minor ticks, I have
subsy=[1,2,3,4,5,6,7,8,9] as an argument for semilogy which seems to
work fine. To have labels every other magnitude I used
ticker.LogLocator(base=100.0), which gives me the labels where I want
them.
The two things I have been unable to figure out are how to add a major
tick at all of the other magnitudes (those without a label) and how to
change the format of the labels so that only the exponent is showing.
I'm sure both of these are straightforward, but I have been unable to
figure them out. Thanks for your help.
-Jeffrey

Michael Hearne wrote:
> Jeff - Well, that's the problem - I _do_ need scipy! Not for 
> basemap, but for the rest of my application. Does your sourceforge 
> repository have older copies of the basemap code?
Mike: OK, but I don't think upgrading numpy/maplotlib requires you to 
update scipy. 
The SF site does have older copies of basemap, but beware - I'm not 
backporting bugfixes.
-Jeff
>
> My complaint about scipy has to do with the fact that I spent a good 
> week trying to get scipy's dependencies to install on two Red Hat 
> Enterprise Linux systems: I punted both of them over to an outside 
> contractor, who was able to compile scipy on one system, but not the 
> other. And he specializes in that kind of stuff.
>
> --Mike
>
> Jeff Whitaker wrote:
>> Michael Hearne wrote:
>>> I'm trying to build an application on a newly installed Ubuntu. 
>>> I've used apt-get to install scipy and matplotlib.
>>>
>>> numpy version: 1.0.4
>>> scipy version: 0.6
>>> matplotlib version: 0.91.2
>>>
>>> What version of basemap is safe to download and build with this 
>>> configuration? Doing "apt-cache search basemap" yields no results, 
>>> so I assume there isn't an Ubuntu package for that.
>>>
>>> And if someone tells me I need to compile a newer scipy (and it's 
>>> #$%^&! dependencies) from source, I'll scream.
>>>
>>> Thanks,
>>>
>>> Mike
>>>
>>> 
>> Mike: Basemap's now only supported for matplotlib > 0.98. For that 
>> you need numpy 1.1 (you don't need scipy at all).
>> -Jeff
>>
>
-- 
Jeffrey S. Whitaker Phone : (303)497-6313
Meteorologist FAX : (303)497-6449
NOAA/OAR/PSD R/PSD1 Email : Jef...@no...
325 Broadway Office : Skaggs Research Cntr 1D-113
Boulder, CO, USA 80303-3328 Web : http://tinyurl.com/5telg

Jeff - Well, that's the problem - I _do_ need scipy! Not for basemap, 
but for the rest of my application. Does your sourceforge repository 
have older copies of the basemap code?
My complaint about scipy has to do with the fact that I spent a good 
week trying to get scipy's dependencies to install on two Red Hat 
Enterprise Linux systems: I punted both of them over to an outside 
contractor, who was able to compile scipy on one system, but not the 
other. And he specializes in that kind of stuff.
--Mike
Jeff Whitaker wrote:
> Michael Hearne wrote:
>> I'm trying to build an application on a newly installed Ubuntu. I've 
>> used apt-get to install scipy and matplotlib.
>>
>> numpy version: 1.0.4
>> scipy version: 0.6
>> matplotlib version: 0.91.2
>>
>> What version of basemap is safe to download and build with this 
>> configuration? Doing "apt-cache search basemap" yields no results, 
>> so I assume there isn't an Ubuntu package for that.
>>
>> And if someone tells me I need to compile a newer scipy (and it's 
>> #$%^&! dependencies) from source, I'll scream.
>>
>> Thanks,
>>
>> Mike
>>
>> 
> Mike: Basemap's now only supported for matplotlib > 0.98. For that 
> you need numpy 1.1 (you don't need scipy at all).
> -Jeff
>
-- 
------------------------------------------------------
Michael Hearne
mh...@us...
(303) 273-8620
USGS National Earthquake Information Center
1711 Illinois St. Golden CO 80401
Senior Software Engineer
Synergetics, Inc.
------------------------------------------------------

Michael Hearne wrote:
> I'm trying to build an application on a newly installed Ubuntu. I've 
> used apt-get to install scipy and matplotlib.
>
> numpy version: 1.0.4
> scipy version: 0.6
> matplotlib version: 0.91.2
>
> What version of basemap is safe to download and build with this 
> configuration? Doing "apt-cache search basemap" yields no results, so I 
> assume there isn't an Ubuntu package for that.
>
> And if someone tells me I need to compile a newer scipy (and it's #$%^&! 
> dependencies) from source, I'll scream.
>
> Thanks,
>
> Mike
>
> 
Mike: Basemap's now only supported for matplotlib > 0.98. For that 
you need numpy 1.1 (you don't need scipy at all). 
-Jeff
-- 
Jeffrey S. Whitaker Phone : (303)497-6313
Meteorologist FAX : (303)497-6449
NOAA/OAR/PSD R/PSD1 Email : Jef...@no...
325 Broadway Office : Skaggs Research Cntr 1D-113
Boulder, CO, USA 80303-3328 Web : http://tinyurl.com/5telg

I'm trying to build an application on a newly installed Ubuntu. I've 
used apt-get to install scipy and matplotlib.
numpy version: 1.0.4
scipy version: 0.6
matplotlib version: 0.91.2
What version of basemap is safe to download and build with this 
configuration? Doing "apt-cache search basemap" yields no results, so I 
assume there isn't an Ubuntu package for that.
And if someone tells me I need to compile a newer scipy (and it's #$%^&! 
dependencies) from source, I'll scream.
Thanks,
Mike
-- 
------------------------------------------------------
Michael Hearne
mh...@us...
(303) 273-8620
USGS National Earthquake Information Center
1711 Illinois St. Golden CO 80401
Senior Software Engineer
Synergetics, Inc.
------------------------------------------------------

Hello - I have two problems labeling contour lines in 0.98.3.
First, when I call clabel, it removes all contours that are not labeled
(because the label doesn't fit on the section of contour, I presume).
This seems like a bug to me (or a really odd feature).
Easy example:
>>> x,y = meshgrid( linspace(-10,10,50), linspace(-10,10,50) )
>>> z = log(x**2 + y**2)
>>> cobj = contour(x,y,z)
>>> cobj.clabel()
<a list of 8 text.Text objects>
>>> draw()
Second, when using the new manual labeling of contour labels works (which is
pretty neat!), how do I end this feature?
The doc string says: right click, or potentially click both mouse buttons
together (which already worries me).
Neither works for me on win32, mpl 0.98.3, TkAgg backend, interactive mode.
Does anybody have a solution?
Thanks, Mark

Hi Mike,
> Sounds like a Inkscape/Cairo issue to me, if the PDF it converts doesn't 
> match the SVG. That said, SVG is tricky to get right, and we may be 
> able to produce things in a different way that Inkscape will handle 
> correctly. I'll look into this and file a bug with Inkscape and/or 
> Cairo if it turns out to be a bug on their end.
> 
Thanks!
> Does directly outputting the PDF from matplotlib work as a workaround?
> 
Yes, it works just fine. And importing this pdf to Inkscape works fine too! So 
that would be a convenient way to go.
Thanks!
-K
>> Hi!
>>
>> This may be a pdf issue or an SVG issue and have nothing to do with 
>> matplotlib.
>> If so, sorry...
>>
>> Whenever I produce a figure with matplotlib and 'crop it' using the 
>> axis command
>> the svg comes out fine. However, when I save this svg to pdf using 
>> inkscape's
>> 'cairo' backend the part of the data that was 'cropped' becomes 
>> visible in the
>> pdf. e.g. if I had plotted data from x = -2 to +10, and then I set the 
>> axes to
>> only show the data from +3 to +6, the svg comes out fine, but the pdf 
>> shows the
>> rest of the graph leaking out from the axes.
>>
>> I currently have to manually crop my data to prevent this.
>>
>> Thanks!
>> -K
>>

Sounds like a Inkscape/Cairo issue to me, if the PDF it converts doesn't 
match the SVG. That said, SVG is tricky to get right, and we may be 
able to produce things in a different way that Inkscape will handle 
correctly. I'll look into this and file a bug with Inkscape and/or 
Cairo if it turns out to be a bug on their end.
Does directly outputting the PDF from matplotlib work as a workaround?
Cheers,
Mike
Kaushik Ghose wrote:
> Hi!
>
> This may be a pdf issue or an SVG issue and have nothing to do with matplotlib.
> If so, sorry...
>
> Whenever I produce a figure with matplotlib and 'crop it' using the axis command
> the svg comes out fine. However, when I save this svg to pdf using inkscape's
> 'cairo' backend the part of the data that was 'cropped' becomes visible in the
> pdf. e.g. if I had plotted data from x = -2 to +10, and then I set the axes to
> only show the data from +3 to +6, the svg comes out fine, but the pdf shows the
> rest of the graph leaking out from the axes.
>
> I currently have to manually crop my data to prevent this.
>
> Thanks!
> -K
>
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 

Hi!
This may be a pdf issue or an SVG issue and have nothing to do with matplotlib.
If so, sorry...
Whenever I produce a figure with matplotlib and 'crop it' using the axis command
the svg comes out fine. However, when I save this svg to pdf using inkscape's
'cairo' backend the part of the data that was 'cropped' becomes visible in the
pdf. e.g. if I had plotted data from x = -2 to +10, and then I set the axes to
only show the data from +3 to +6, the svg comes out fine, but the pdf shows the
rest of the graph leaking out from the axes.
I currently have to manually crop my data to prevent this.
Thanks!
-K

Hi!
This may be a pdf issue or an SVG issue and have nothing to do with matplotlib. 
If so, sorry...
Whenever I produce a figure with matplotlib and 'crop it' using the axis command 
the svg comes out fine. However, when I save this svg to pdf using inkscape's 
'cairo' backend the part of the data that was 'cropped' becomes visible in the 
pdf. e.g. if I had plotted data from x = -2 to +10, and then I set the axes to 
only show the data from +3 to +6, the svg comes out fine, but the pdf shows the 
rest of the graph leaking out from the axes.
I currently have to manually crop my data to prevent this.
Thanks!
-K