matplotlib-users Mailing List for matplotlib

Graeme Lufkin wrote:
>     Matplotlib wraps the Antigrain library quite nicely, and now I'd
> like to use the bindings for my own nefarious purposes.

Perhaps it would be easier to use agg-draw:

http://effbot.org/downloads/#aggdraw

If not, I'd love to see a general purpose python binding to Agg. 
Personally, I'd use it with wxPython.

-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...

    Matplotlib wraps the Antigrain library quite nicely, and now I'd
like to use the bindings for my own nefarious purposes.  Right now,
what I'd like is to create a (probably modified) FigureCanvas, and be
able to call Renderer function to draw my own graphics.  I'd embed
this object in my GTK app, so I'm looking at the GTKAgg stuff.  I've
traced down FigureCanvasGTK as inheriting from gtk.DrawingArea, which
I've (awkwardly) used before.  I'd like to call Agg drawing functions
though, to get the nice anti-aliasing.  So I'm looking at RendererAgg.
    If I create a FigureCanvasGTKAgg with a Figure object I'm not
going to use, I can extract the RendererAgg with get_renderer(), at
which point I should be able to call Agg drawing functions.  But I
don't understand what kind of graphics contexts are being used as
parameters to those functions.
    Is this even the right way to go?  I want to do all my own Agg
drawing, preferably offscreen with an update() function I call
periodically.  Eventually I can write my own FigureCanvas-like class
that inherits from the right classes (from my UI toolkit's drawing
widget, and from RendererAgg that exposes the Agg .so) and omits all
the axes stuff that Matplotlib cares about.
    Thoughts, pointers, please?  Thanks
--
    -- Graeme
    gra...@gm...

HI all,

I'm trying to figure out how to draw something in pixel or "real" 
units(inches, mm, etc). what I want is to be able to place an object 
according to axes units, but not have it change size as I zoom in and 
out, kind of like a marker.

The big picture:

I'm trying to make a plot that shows a vector value over time (in the 
case ocean current velocity and direction). The x axis is time, the y 
axis is magnitude, and at each point I want to draw an arrow pointing in 
the direction of the vector at that time. However, I don't want the size 
(or angle) of that arrow to change as you zoom in and out.

Any ideas how I can accomplish this?

See the enclosed PNG for an example with fake data.

-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

Your second method that referred to bbox *worked*!!! =20

# ------------------------------------------------------
from pylab import *
fig =3D Figure()
p =3D fig.add_subplot(111)
plot([1,2,3])
t =3D text(0, 2.2, "Hello", horizontalalignment =3D 'center')
t.set_clip_box(p.bbox)
show()
# ------------------------------------------------------

The first method did *not* for some reason!?!? :(

# ------------------------------------------------------
from pylab import *
plot([1,2,3])
text(0, 2.2, "Hello", horizontalalignment =3D 'center', clip_on =3D True)
show()
# ------------------------------------------------------

I would prefer to make the second method work because it lets me avoid
having to define a Figure object /explicitly/ as above.  The problem
with adding a subplot to a Figure object is that that many defaults
change that I would have to fix. (e.g. grid lines disappear, plot
dimensions are altered in strange way, etc.) =20

How fix second way to avoid a Figure object? =20


Chris



On Wed, 2006-01-04 at 12:05 -0600, John Hunter wrote:
> >>>>> "Christian" =3D=3D Christian Seberino <seb...@sp...> wr=
ites:
>=20
>     Christian> John Thanks!  I tried set_clip_on but it did not work
>     Christian> for me.  I made the little script below which shows my
>     Christian> point.  "Hello" *still* hangs over the left side.  Is
>     Christian> this because the coordinates of the text are on the
>     Christian> edge of the plot?  Can clip_on be somehow applied to
>     Christian> text with x =3D 0 as well?
>=20
> If you want to clip to the axes bounding box, you either need to set
> clip_on when the text instance is created
>=20
>   text(x,y,s, clip_on=3DTrue)
>=20
> or later if you have an existing text instance 't' you can set the clip
> bounding box manually
>=20
>   t.set_clip_box(ax.bbox)
>=20
> JDH
>=20
>=20
>=20
>     Christian> from pylab import * plot([1,2,3]) t =3D text(0, 2.2,
>     Christian> "Hello", horizontalalignment =3D 'center')
>     Christian> t.set_clip_on(True) show()
>=20
>=20
>=20
>=20
>=20
>=20
>=20
>     Christian> On Wed, 2006-01-04 at 08:32 -0600, John Hunter wrote:
>     >> >>>>> "Christian" =3D=3D Christian Seberino
>     >> <seb...@sp...> writes:
>     >>=20
>     Christian> How control whether text & plot elements can go beyond
>     Christian> EDGE of plot?  Is it possible to have some plot
>     Christian> elements and text be able to go over and others get
>     Christian> chopped off?
>     >>  You can turn the clipping attribute on or off for any artist
>     >> element.  Artist is base class for all matplotlib elements in
>     >> the figure (Figure, Axes, Line2D, Patches, Text...)
>     >>=20
>     >> http://matplotlib.sf.net/matplotlib.artist.html
>     >>=20
>     >> You will want to set the "clip_on" attribute for any line (or
>     >> text or whatever) element
>     >>=20
>     >> t =3D ax.text(x,y,s) t.set_clip_on(False)
>     >>=20
>     >> l, =3D ax.plot([1,2,3]) l.set_clip_on(False)
>     >>=20
>     >> Or you can use setp to control the properties of one or more
>     >> artists
>     >>=20
>     >> setp([t,l], clip_on=3DFalse)
>     >>=20
>     >> By default when clpping is on, the artists are clipped to the
>     >> edges of the Axes rectangle.  You can control the clipping box
>     >> by passing a custom bbox to clip_box attribute
>     >>=20
>     >> from pylab import figure, show, draw, rand from
>     >> matplotlib.transforms import lbwh_to_bbox bbox =3D
>     >> lbwh_to_bbox(100,200,50,70) fig =3D figure() ax =3D
>     >> fig.add_subplot(111)
>     >>=20
>     >> line, =3D ax.plot(rand(100), rand(100)) line.set_clip_box(bbox)
>     >> show()
>     >>=20
>     >> Note that the clip box is in figure coordinates (pixels from
>     >> lower, left)
>     >>=20
>     >> JDH
>     >>=20
>     >>=20
>     >> ------------------------------------------------------- This
>     >> SF.net email is sponsored by: Splunk Inc. Do you grep through
>     >> log files for problems?  Stop!  Download the new AJAX search
>     >> engine that makes searching your log files as easy as surfing
>     >> the web.  DOWNLOAD SPLUNK!
>     >> http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
>     >> _______________________________________________
>     >> Matplotlib-users mailing list
>     >> Mat...@li...
>     >> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>     >>=20
>     Christian> -- _______________________________________
>=20
>     Christian> Christian Seberino, Ph.D.  SPAWAR Systems Center San
>     Christian> Diego Code 2872 49258 Mills Street, Room 158 San Diego,
>     Christian> CA 92152-5385 U.S.A.
>=20
>     Christian> Phone: (619) 553-9973 Fax : (619) 553-6521 Email:
>     Christian> seb...@sp...
>     Christian> _______________________________________
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi=
les
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>=20
--=20
_______________________________________

Christian Seberino, Ph.D.
SPAWAR Systems Center San Diego
Code 2872
49258 Mills Street, Room 158
San Diego, CA 92152-5385
U.S.A.

Phone: (619) 553-9973
Fax  : (619) 553-6521
Email: seb...@sp...
_______________________________________

On Wednesday 04 January 2006 12:14, Darren Dale wrote:
> > I had to make one very small tweak to what he posted (included at the
> > end of this message).
> > There is one line that sets up a command like this:
> > command =3D '/usr/local/bin/ps2eps -l "%s"'% psfile
> > This file did not exist on my machine, but installing the ps2eps
> > package from the ubuntu package manager and changing the command to
> > command =3D 'ps2eps -l "%s"'% psfile
> > worked perfectly (I realize this is a very simple thing, but I took a
> > little bit of work to track it down - the error message generate
> > wasn't all that helpful. =A0So, I include it here in hopes of saving
> > someone else a little trouble.)
>
> I need to get the wiki page back up, this ps2eps problem was discussed
> there. I guess I can work on it during lunch today.

The usetex wiki entry is back up. (and backed up)

>>>>> "Christian" == Christian Seberino <seb...@sp...> writes:

    Christian> John Thanks!  I tried set_clip_on but it did not work
    Christian> for me.  I made the little script below which shows my
    Christian> point.  "Hello" *still* hangs over the left side.  Is
    Christian> this because the coordinates of the text are on the
    Christian> edge of the plot?  Can clip_on be somehow applied to
    Christian> text with x = 0 as well?

If you want to clip to the axes bounding box, you either need to set
clip_on when the text instance is created

  text(x,y,s, clip_on=True)

or later if you have an existing text instance 't' you can set the clip
bounding box manually

  t.set_clip_box(ax.bbox)

JDH



    Christian> from pylab import * plot([1,2,3]) t = text(0, 2.2,
    Christian> "Hello", horizontalalignment = 'center')
    Christian> t.set_clip_on(True) show()







    Christian> On Wed, 2006-01-04 at 08:32 -0600, John Hunter wrote:
    >> >>>>> "Christian" == Christian Seberino
    >> <seb...@sp...> writes:
    >> 
    Christian> How control whether text & plot elements can go beyond
    Christian> EDGE of plot?  Is it possible to have some plot
    Christian> elements and text be able to go over and others get
    Christian> chopped off?
    >>  You can turn the clipping attribute on or off for any artist
    >> element.  Artist is base class for all matplotlib elements in
    >> the figure (Figure, Axes, Line2D, Patches, Text...)
    >> 
    >> http://matplotlib.sf.net/matplotlib.artist.html
    >> 
    >> You will want to set the "clip_on" attribute for any line (or
    >> text or whatever) element
    >> 
    >> t = ax.text(x,y,s) t.set_clip_on(False)
    >> 
    >> l, = ax.plot([1,2,3]) l.set_clip_on(False)
    >> 
    >> Or you can use setp to control the properties of one or more
    >> artists
    >> 
    >> setp([t,l], clip_on=False)
    >> 
    >> By default when clpping is on, the artists are clipped to the
    >> edges of the Axes rectangle.  You can control the clipping box
    >> by passing a custom bbox to clip_box attribute
    >> 
    >> from pylab import figure, show, draw, rand from
    >> matplotlib.transforms import lbwh_to_bbox bbox =
    >> lbwh_to_bbox(100,200,50,70) fig = figure() ax =
    >> fig.add_subplot(111)
    >> 
    >> line, = ax.plot(rand(100), rand(100)) line.set_clip_box(bbox)
    >> show()
    >> 
    >> Note that the clip box is in figure coordinates (pixels from
    >> lower, left)
    >> 
    >> JDH
    >> 
    >> 
    >> ------------------------------------------------------- This
    >> SF.net email is sponsored by: Splunk Inc. Do you grep through
    >> log files for problems?  Stop!  Download the new AJAX search
    >> engine that makes searching your log files as easy as surfing
    >> the web.  DOWNLOAD SPLUNK!
    >> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
    >> _______________________________________________
    >> Matplotlib-users mailing list
    >> Mat...@li...
    >> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
    >> 
    Christian> -- _______________________________________

    Christian> Christian Seberino, Ph.D.  SPAWAR Systems Center San
    Christian> Diego Code 2872 49258 Mills Street, Room 158 San Diego,
    Christian> CA 92152-5385 U.S.A.

    Christian> Phone: (619) 553-9973 Fax : (619) 553-6521 Email:
    Christian> seb...@sp...
    Christian> _______________________________________

For what its worth, I have ubuntu running on and external USB drive
with only a little bit of pain with the init stuff.

Ryan

On 1/4/06, Gary <pa...@in...> wrote:
> This is cool.
>
> About 18 months ago, I wanted, for various reasons, a linux that I could
> boot from an external USB drive.  I never got that to work despite hours
> of tinkering with initrd in various distros.  This might be better:
>
> VMware ( http://www.vmware.com/ ) sells a product that makes virtual
> machines that run on WinXP and linux.  For free, they make available
> VMware Player, an app that runs the VMs once constructed (kinda like
> Adobe Acrobat / Acrobat Reader).  However, they maintain a collection of
> Really Useful preconfigured VMs ( http://www.vmware.com/vmtn/vm/ )
> ready to go.
>
> The demo is "Browser Appliance", a VM running Ubuntu Linux,
> preconfigured for web browsing with firefox.  I installed that, and,
> wow, I got what appears to be a full-featured ubuntu VM.  Running in
> full screen mode, it appears to be indistinguishable from a conventional
> linux installation... but it's not.  It's in a VM on an XP host.  XP is
> still running, and I can still use all the XP stuff.
>
> But this ubuntu VM appears to run perfectly.  It just works.  On my T41
> Thinkpad (Centrino) it finds and uses that pesky Centrino wireless
> device, with no effort on my part.   I fired up the Synaptic Package
> Manager and starting adding packages.  No sweat.  I downloaded the
> matplotlib sources. (after grabbing gcc, etc) A couple of symbolic
> links, and a few tweaks of setup.py and matplotlibrc, and bingo! it
> compiles.  And works!
>
> The VM and its memory etc etc live in WinXP files on the WinXP file
> system.  So I plug in an external USB drive (fat32 for historical
> reasons) and copy the VM files over there.  Shoot... the dang thing just
> works.
>
> As far as I can tell, I have a fully functional ubuntu on my external
> USB drive.   I'm in the process of stressing the system in various ways,
> and so far, everything works.  Haven't tried printing, yet.  There's no
> obvious way to mount the WinXP file system, although I ran across a
> reference that says Samba works for communication to the host.  Haven't
> tried it.
>
> I had *no* (as in *zero*) problems installing it and getting it to
> work.  It's *never* this easy.  One thing I'm not sure of:  speed.  But
> it is not obviously slower than a native installation.  None of the
> software installations or the compiling went noticibly slower than expect=
ed.
>
> The irony is that I don't think I need it anymore.
>
> -gary
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi=
les
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

John

Thanks!  I tried set_clip_on but it did not work for me.
I made the little script below which shows my point.
"Hello" *still* hangs over the left side.
Is this because the coordinates of the text are on=20
the edge of the plot?  Can clip_on be somehow applied
to text with x =3D 0 as well?




from pylab import *
plot([1,2,3])
t =3D text(0, 2.2, "Hello", horizontalalignment =3D 'center')
t.set_clip_on(True)
show()







On Wed, 2006-01-04 at 08:32 -0600, John Hunter wrote:
> >>>>> "Christian" =3D=3D Christian Seberino <seb...@sp...> wr=
ites:
>=20
>     Christian> How control whether text & plot elements can go beyond
>     Christian> EDGE of plot?  Is it possible to have some plot
>     Christian> elements and text be able to go over and others get
>     Christian> chopped off?
>=20
> You can turn the clipping attribute on or off for any artist element.
> Artist is base class for all matplotlib elements in the figure
> (Figure, Axes, Line2D, Patches, Text...)
>=20
> http://matplotlib.sf.net/matplotlib.artist.html
>=20
> You will want to set the "clip_on" attribute for any line (or text or
> whatever) element
>=20
>   t =3D ax.text(x,y,s)
>   t.set_clip_on(False)
>=20
>   l, =3D ax.plot([1,2,3])
>   l.set_clip_on(False)
>=20
> Or you can use setp to control the properties of one or more artists
>=20
>    setp([t,l], clip_on=3DFalse)
>=20
> By default when clpping is on, the artists are clipped to the edges of
> the Axes rectangle.  You can control the clipping box by passing a
> custom bbox to clip_box attribute
>=20
>   from pylab import figure, show, draw, rand
>   from matplotlib.transforms import lbwh_to_bbox
>   bbox =3D lbwh_to_bbox(100,200,50,70)
>   fig =3D figure()
>   ax =3D fig.add_subplot(111)
>=20
>   line, =3D ax.plot(rand(100), rand(100))
>   line.set_clip_box(bbox)
>   show()
>=20
> Note that the clip box is in figure coordinates (pixels from lower,
> left)
>=20
> JDH
>=20
>=20
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log fi=
les
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>=20
--=20
_______________________________________

Christian Seberino, Ph.D.
SPAWAR Systems Center San Diego
Code 2872
49258 Mills Street, Room 158
San Diego, CA 92152-5385
U.S.A.

Phone: (619) 553-9973
Fax  : (619) 553-6521
Email: seb...@sp...
_______________________________________

>>>>> "Gary" == Gary  <pa...@in...> writes:

    Gary> VMware ( http://www.vmware.com/ ) sells a product that makes
    Gary> virtual machines that run on WinXP and linux.  For free,
    Gary> they make available VMware Player, an app that runs the VMs
    Gary> once constructed (kinda like Adobe Acrobat / Acrobat
    Gary> Reader).  However, they maintain a collection of Really
    Gary> Useful preconfigured VMs ( http://www.vmware.com/vmtn/vm/ )
    Gary> ready to go.

I'm a huge vmware fan, but my configuration is the opposite.  I have
native linux running windows xp virtually in vmware.  That's where I
do the matplotlib win32 builds, actually.

Thanks for the heads-up on vmware player.

JDH

This is cool. 

About 18 months ago, I wanted, for various reasons, a linux that I could 
boot from an external USB drive.  I never got that to work despite hours 
of tinkering with initrd in various distros.  This might be better:

VMware ( http://www.vmware.com/ ) sells a product that makes virtual 
machines that run on WinXP and linux.  For free, they make available 
VMware Player, an app that runs the VMs once constructed (kinda like 
Adobe Acrobat / Acrobat Reader).  However, they maintain a collection of 
Really Useful preconfigured VMs ( http://www.vmware.com/vmtn/vm/ )  
ready to go.

The demo is "Browser Appliance", a VM running Ubuntu Linux, 
preconfigured for web browsing with firefox.  I installed that, and, 
wow, I got what appears to be a full-featured ubuntu VM.  Running in 
full screen mode, it appears to be indistinguishable from a conventional 
linux installation... but it's not.  It's in a VM on an XP host.  XP is 
still running, and I can still use all the XP stuff.

But this ubuntu VM appears to run perfectly.  It just works.  On my T41 
Thinkpad (Centrino) it finds and uses that pesky Centrino wireless 
device, with no effort on my part.   I fired up the Synaptic Package 
Manager and starting adding packages.  No sweat.  I downloaded the 
matplotlib sources. (after grabbing gcc, etc) A couple of symbolic 
links, and a few tweaks of setup.py and matplotlibrc, and bingo! it 
compiles.  And works!

The VM and its memory etc etc live in WinXP files on the WinXP file 
system.  So I plug in an external USB drive (fat32 for historical 
reasons) and copy the VM files over there.  Shoot... the dang thing just 
works. 

As far as I can tell, I have a fully functional ubuntu on my external 
USB drive.   I'm in the process of stressing the system in various ways, 
and so far, everything works.  Haven't tried printing, yet.  There's no 
obvious way to mount the WinXP file system, although I ran across a 
reference that says Samba works for communication to the host.  Haven't 
tried it.   

I had *no* (as in *zero*) problems installing it and getting it to 
work.  It's *never* this easy.  One thing I'm not sure of:  speed.  But 
it is not obviously slower than a native installation.  None of the 
software installations or the compiling went noticibly slower than expected.

The irony is that I don't think I need it anymore.

-gary

Dave wrote:
> Also, unrelated but is there a property I can check once using matplotlib to
> get the version information (something like matplotlib.version)?

matplotlib.__version__

this has become something of a python standard.

-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...

On Wednesday 04 January 2006 10:59, Ryan Krauss wrote:
> Thanks to Darren and anyone else involved in making the usetex option
> work.  It generates the most beautiful plots I have ever seen.  

That's great to hear. There have been several contributors along the way, 
notably JDH who had the idea in the first place.

> I had to make one very small tweak to what he posted (included at the
> end of this message).
> There is one line that sets up a command like this:
> command = '/usr/local/bin/ps2eps -l "%s"'% psfile
> This file did not exist on my machine, but installing the ps2eps
> package from the ubuntu package manager and changing the command to
> command = 'ps2eps -l "%s"'% psfile
> worked perfectly (I realize this is a very simple thing, but I took a
> little bit of work to track it down - the error message generate
> wasn't all that helpful.  So, I include it here in hopes of saving
> someone else a little trouble.)

I need to get the wiki page back up, this ps2eps problem was discussed there. 
I guess I can work on it during lunch today.

> Is there any way that this option that preserves the text could be an
> rc option?  I can see myself making this change every time a new
> version is released.

Maybe I could test for the necessary deps (ps2eps, xpdf, ghostscript, etc) in 
backend_ps, and if they are all available, then the eps->pdf->ps->eps 
renormalization step is performed, and if not, an error message is passed to 
mpl verbose and the backup distillation step is performed. John, is this 
acceptable?

Darren

This is cool. 

About 18 months ago, I wanted, for various reasons, a linux that I could 
boot from an external USB drive.  I never got that to work despite hours 
of tinkering with initrd in various distros.  This might be better:

VMware ( http://www.vmware.com/ ) sells a product that makes virtual 
machines that run on WinXP and linux.  For free, they make available 
VMware Player, an app that runs the VMs once constructed (kinda like 
Adobe Acrobat / Acrobat Reader).  However, they maintain a collection of 
Really Useful preconfigured VMs ( http://www.vmware.com/vmtn/vm/ )  
ready to go.

The demo is "Browser Appliance", a VM running Ubuntu Linux, 
preconfigured for web browsing with firefox.  I installed that, and, 
wow, I got what appears to be a full-featured ubuntu VM.  Running in 
full screen mode, it appears to be indistinguishable from a conventional 
linux installation... but it's not.  It's in a VM on an XP host.  XP is 
still running, and I can still use all the XP stuff.

But this ubuntu VM appears to run perfectly.  It just works.  On my T41 
Thinkpad (Centrino) it finds and uses that pesky Centrino wireless 
device, with no effort on my part.   I fired up the Synaptic Package 
Manager and starting adding packages.  No sweat.  I downloaded the 
matplotlib sources. (after grabbing gcc, etc) A couple of symbolic 
links, and a few tweaks of setup.py and matplotlibrc, and bingo! it 
compiles.  And works!

The VM and its memory etc etc live in WinXP files on the WinXP file 
system.  So I plug in an external USB drive (fat32 for historical 
reasons) and copy the VM files over there.  Shoot... the dang thing just 
works. 

As far as I can tell, I have a fully functional ubuntu on my external 
USB drive.   I'm in the process of stressing the system in various ways, 
and so far, everything works.  Haven't tried printing, yet.  There's no 
obvious way to mount the WinXP file system, although I ran across a 
reference that says Samba works for communication to the host.  Haven't 
tried it.   

I had *no* (as in *zero*) problems installing it and getting it to 
work.  It's *never* this easy.  One thing I'm not sure of:  speed.  But 
it is not obviously slower than a native installation.  None of the 
software installations or the compiling went noticibly slower than expected.

The irony is that I don't think I need it anymore.

-gary

Thanks to Darren and anyone else involved in making the usetex option
work.  It generates the most beautiful plots I have ever seen.  And
following the advice Darren posted a while back, the text in my axis
labels is searchable when the plots are included in a pdf.

I had to make one very small tweak to what he posted (included at the
end of this message).
There is one line that sets up a command like this:
command =3D '/usr/local/bin/ps2eps -l "%s"'% psfile
This file did not exist on my machine, but installing the ps2eps
package from the ubuntu package manager and changing the command to
command =3D 'ps2eps -l "%s"'% psfile
worked perfectly (I realize this is a very simple thing, but I took a
little bit of work to track it down - the error message generate
wasn't all that helpful.  So, I include it here in hopes of saving
someone else a little trouble.)

Is there any way that this option that preserves the text could be an
rc option?  I can see myself making this change every time a new
version is released.

Thanks,

Ryan

#=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D
You can remove the following block in backend_ps.py, starting around line
1144:

           command =3D 'latex -interaction=3Dnonstopmode "%s"' % texfile

           verbose.report(command, 'debug-annoying')
           stdin, stdout, stderr =3D os.popen3(command)
           verbose.report(stdout.read(), 'debug-annoying')
           verbose.report(stderr.read(), 'helpful')
           command =3D 'dvips -R -T %fin,%fin -o "%s" "%s"' % (pw, ph, psfi=
le,
dvifile)
           verbose.report(command, 'debug-annoying')
           stdin, stdout, stderr =3D os.popen3(command)
           verbose.report(stdout.read(), 'debug-annoying')
           verbose.report(stderr.read(), 'helpful')
           os.remove(epsfile)
           if ext.startswith('.ep'):
               dpi =3D rcParams['ps.distiller.res']

               if sys.platform =3D=3D 'win32':
                   command =3D 'gswin32c -dBATCH -dNOPAUSE -dSAFER -r%d \
                     -sDEVICE=3Depswrite -dLanguageLevel=3D2 -dEPSFitPage \
                     -sOutputFile=3D"%s" "%s"'% (dpi, epsfile, psfile)
               else:
                   command =3D 'gs -dBATCH -dNOPAUSE -dSAFER -r%d \
                   -sDEVICE=3Depswrite -dLanguageLevel=3D2 -dEPSFitPage \
                   -sOutputFile=3D"%s" "%s"'% (dpi, epsfile, psfile)

               verbose.report(command, 'debug-annoying')
               stdin, stdout, stderr =3D os.popen3(command)
               verbose.report(stdout.read(), 'debug-annoying')
               verbose.report(stderr.read(), 'helpful')
               shutil.move(epsfile, outfile)
           else: shutil.move(psfile, outfile)


and replace it with this:

           command =3D 'latex -interaction=3Dnonstopmode "%s"' % texfile
           verbose.report(command, 'debug-annoying')
           stdin, stdout, stderr =3D os.popen3(command)
           verbose.report(stdout.read(), 'debug-annoying')
           verbose.report(stderr.read(), 'helpful')
           command =3D 'dvips -R -T %fin,%fin -o "%s" "%s"' % (pw, ph, psfi=
le,
dvifile)
           verbose.report(command, 'debug-annoying')
           stdin, stdout, stderr =3D os.popen3(command)
           verbose.report(stdout.read(), 'debug-annoying')
           verbose.report(stderr.read(), 'helpful')
           os.remove(epsfile)
           pdffile =3D tmpname + '.pdf'
           if ext.startswith('.ep'):
               command =3D 'ps2pdf "%s"'% psfile
               os.system(command)
               command =3D 'pdftops -level2 "%s" "%s"'% (pdffile, psfile)
               os.system(command)
               os.remove(pdffile)
               command =3D '/usr/local/bin/ps2eps -l "%s"'% psfile
               stdin, stderr =3D os.popen4(command)
               verbose.report(stderr.read(), 'helpful')
               command =3D 'epstopdf "%s"'% epsfile
               os.system(command)
               shutil.move(epsfile, outfile)
               shutil.move(pdffile, basename+'.pdf')
           else:
               command =3D 'ps2pdf "%s" "%s"'% (psfile, pdffile)
               stdin, stderr =3D os.popen4(command)
               verbose.report(stderr.read(), 'helpful')
               os.remove(psfile)
               command =3D 'pdftops -paperw %d -paperh %d -level2 "%s" "%s"=
'% \
                           (int(pw*72), int(ph*72), pdffile, psfile)
               os.system(command)
               shutil.move(psfile, outfile)
#=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D

>>>>> "Christian" == Christian Seberino <seb...@sp...> writes:

    Christian> How control whether text & plot elements can go beyond
    Christian> EDGE of plot?  Is it possible to have some plot
    Christian> elements and text be able to go over and others get
    Christian> chopped off?

You can turn the clipping attribute on or off for any artist element.
Artist is base class for all matplotlib elements in the figure
(Figure, Axes, Line2D, Patches, Text...)

http://matplotlib.sf.net/matplotlib.artist.html

You will want to set the "clip_on" attribute for any line (or text or
whatever) element

  t = ax.text(x,y,s)
  t.set_clip_on(False)

  l, = ax.plot([1,2,3])
  l.set_clip_on(False)

Or you can use setp to control the properties of one or more artists

   setp([t,l], clip_on=False)

By default when clpping is on, the artists are clipped to the edges of
the Axes rectangle.  You can control the clipping box by passing a
custom bbox to clip_box attribute

  from pylab import figure, show, draw, rand
  from matplotlib.transforms import lbwh_to_bbox
  bbox = lbwh_to_bbox(100,200,50,70)
  fig = figure()
  ax = fig.add_subplot(111)

  line, = ax.plot(rand(100), rand(100))
  line.set_clip_box(bbox)
  show()

Note that the clip box is in figure coordinates (pixels from lower,
left)

JDH

>>>>> "Charlie" == Charlie Moad <cw...@gm...> writes:

    Charlie> This is what the zorder is used for.  Artists should have
    Charlie> a zorder attribute.  The higher the zorder, the more in
    Charlie> front an object is.  I think they all default to 1.

Different types of "Artists" have different default zorders

  In [2]: from matplotlib.patches import Patch

  In [3]: from matplotlib.lines import Line2D

  In [4]: Line2D.zorder
  Out[4]: 2

  In [5]: Patch.zorder
  Out[5]: 1

So by default lines appear above patches....

See also examples/zorder_demo.py.

JDH

In pylab plot() there are named parameters that can be included to control
line properties such as label, linestyle, etc.  But in psd(), stem(), box()
and other specialized plotting commands these parameters usually do not
work.  It would be useful to have them use the same options when it makes
sense.

Can these options be added to psd() and other plot commands to pass through
these same options to plot() or whatever is used for the final display?
Since these commands do not return the lines like plot() does it is a littl=
e
harder to control the properties.

For now what is the best way to change line properties and add labels when
using stem() or psd()?

Also, unrelated but is there a property I can check once using matplotlib t=
o
get the version information (something like matplotlib.version)?

-- David

How control whether text & plot elements can go beyond EDGE of plot?

Is it possible to have some plot elements and text be able
to go over and others get chopped off?

Chris

This is what the zorder is used for.  Artists should have a zorder
attribute.  The higher the zorder, the more in front an object is.  I
think they all default to 1.

- Charlie

On 1/3/06, Christian Seberino <seb...@sp...> wrote:
> How force certain plot items to be on top and others
> to get covered?
>
> (When you have multiple functions being plotted
> it is not clear how to get certain ones on top.)
>
> Chris
>
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.1 (GNU/Linux)
>
> iQEVAwUAQ7sLh8XAD6FzhZH3AQJp2ggArXPHHyEvorIwdjksZM9X56zB5TEL8YmG
> UFZoRKtMHiXZht91kOdnmzm2eKmeA4PEmMwpQ8jpG6icyqb81mt+sUy7o/zf1w3c
> uVgxR/lFk85E3O+Bq56N0tUufZXiiop+2Lg+Lp2PSmNRUeozbHL+CMOOOgdOXgHA
> 3bTEx6jG95nTz4SDpnjbNAFE7Uti7qvhnAi8ngnkh1g+xgoQackI+Nueqd/gFF1j
> Khhmc2Zyvw7yOAbIEj7ECxEHZIYt8rQw8zQV6VHjlyrKgaTe7VdB7xn/2XQOavq4
> ir3wBYbSzKScXUbksAINBupKF11xCZB30AHtovSWjnztxVsS0Wpy2Q=3D=3D
> =3DgNU1
> -----END PGP SIGNATURE-----
>
>
>