gnuplot-py-users Mailing List for Gnuplot.py

Is this on windows or linux?

benny

2008/1/31, kk...@ma... <kk...@ma...>:
>
> When I try to open more than 6 gnuplot windows at the same time I get
> following issues:
> - new windows aren't opened
> - gnuplot 'save' doesn't work
>
> I suppose that problem is in gnuplot perfomance. I mean that gnuplot get
> new
> request before old request is not yet proccessed. Or maybe it is SMP issue
> cause I have Intel Core Duo. Anyway, I've found some workaround:
> I've modified __call__ method with short delay. I know it's ugly but it
> works
>
>     def __call__(self, s):
>         """Send a command string to gnuplot.
>
>         Send the string s as a command to gnuplot, followed by a
>         newline.  All communication with the gnuplot process (except
>         for inline data) is through this method.
>
>         """
>
>         self.gnuplot(s)
>         if self.debug:
>             # also echo to stderr for user to see:
>             sys.stderr.write('gnuplot> %s\n' % (s,))
>             time.sleep(0.1);
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Gnuplot-py-users mailing list
> Gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-py-users
>

When I try to open more than 6 gnuplot windows at the same time I get 
following issues:
- new windows aren't opened
- gnuplot 'save' doesn't work

I suppose that problem is in gnuplot perfomance. I mean that gnuplot get new 
request before old request is not yet proccessed. Or maybe it is SMP issue 
cause I have Intel Core Duo. Anyway, I've found some workaround:
I've modified __call__ method with short delay. I know it's ugly but it works

    def __call__(self, s):
        """Send a command string to gnuplot.

        Send the string s as a command to gnuplot, followed by a
        newline.  All communication with the gnuplot process (except
        for inline data) is through this method.

        """

        self.gnuplot(s)
        if self.debug:
            # also echo to stderr for user to see:
            sys.stderr.write('gnuplot> %s\n' % (s,))
            time.sleep(0.1);

[CC to Gnuplot.py users mailing list.  Please post follow ups there.]

Alan Bromborsky wrote:
> In gnuplot.py is there anyway to read the current orientation (elevation
> and azimuth) of a 3D plot.  I am trying to automate the procedure of 
> finding the best orientation of the 3D plot in an  X-window
> interactively then get that information and print a  high quality hard
> copy in postscript or LaTeX.

Currently, the communications between Gnuplot.py and gnuplot are
strictly one-directional, Gnuplot.py -> gnuplot.  There is currently no
way for Gnuplot.py to read values (e.g., elevation and azimuth) back
from gnuplot.

Two-way communications would be interesting for other purposes, too
(e.g., mouse interactions), in the event that somebody is thinking about
maybe implementing it :-)

Michael

bm...@ca... wrote:
> I'm preparing for a 1.8 package.

It's really great that you are working on this.  Don't forget to add
yourself to CREDITS.txt :-)

> I added the svg patch.
> Furthermore, I need to do the patch in Feisty/Gutsy Ubuntu:
> 
> Index: _Gnuplot.py
> ===================================================================
> --- _Gnuplot.py (revision 299)
> +++ _Gnuplot.py (working copy)
> @@ -574,8 +574,8 @@
>                  % (string.join(keyw.keys(), ', '),)
>                  )
> 
> +        self.set_string('output', filename)
>          self(string.join(setterm))
> -        self.set_string('output', filename)
>          # replot the current figure (to the printer):
>          self.refresh()
> 
> to prevent pictures being overwritten with the next hardcopy command. Anybody
> sees that too (eg in new test.py I write a .svg file after ps, and this
> overwrites the gp_test.ps file too with the svg file without above patch)?

The change seems reasonable.

> As Michael said, we should try to keep the test suit working. I'll look into the
> errors I get when time. These errors are present in 1.7 on my box too however.
> Should somebody have time to look at them or have pointers, let me know. Or
> should they not stop a 1.8?

I strongly suspect that the problems shown below are caused by changes
to gnuplot itself.  Several gnuplot commands and options have changed or
been marked deprecated, and Gnuplot.py hasn't been brought up to date.

Of course it would be nice to bring Gnuplot.py up to date with the
current status of gnuplot, but certainly not at the risk of your energy
running out and no release being made at all!

> Doing python test.py gives errors:
> 
> *  test hardcopy:
> 
> Testing hardcopy options: enhanced=1
> Press return to show results...
> 
> gnuplot> set terminal postscript default enhanced
>                                          ^
>          line 714: extraneous argument in set terminal postscript
> 
> *   test GridData and compute_GridData:
> 
> gnuplot> set parametric
> gnuplot> set data style lines
> gnuplot> set hidden
> gnuplot> set contour base
> gnuplot> set xlabel "x"
> gnuplot> set ylabel "y"
> gnuplot> splot "/tmp/tmpdFzv2L.gnuplot/fifo" notitle
> Same thing, saved to a file
> Press return to show results...
> Notice: Cannot contour non grid data. Please use "set dgrid3d".
> 
> gnuplot> splot "/tmp/tmpvEBqT_"
> Same thing, inline data
> Press return to show results...
> Notice: Cannot contour non grid data. Please use "set dgrid3d".
> 
> gnuplot> splot "-" notitle
> The same thing using binary mode
> Press return to show results...
> Notice: Cannot contour non grid data. Please use "set dgrid3d".
> 
> gnuplot> splot "/tmp/tmpcCCPct.gnuplot/fifo" binary notitle
> Same thing, using binary mode and an intermediate file
> Press return to show results...
>          line 1903: File doesn't factorize into full matrix
> 
> Well, that's all for now folks.

All windows gnuplot users.

Just as Michael (in the past at least), I have no Windows box, so cannot test
patches for the windows functionality.

So patch:
http://sourceforge.net/tracker/index.php?func=detail&aid=1722948&group_id=17434&atid=317434
can only be added if some other windows users tune in that the patch indeed
works as advertised.

Benny

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

Hi,

I'm preparing for a 1.8 package.

I added the svg patch.
Furthermore, I need to do the patch in Feisty/Gutsy Ubuntu:

Index: _Gnuplot.py
===================================================================
--- _Gnuplot.py (revision 299)
+++ _Gnuplot.py (working copy)
@@ -574,8 +574,8 @@
                 % (string.join(keyw.keys(), ', '),)
                 )

+        self.set_string('output', filename)
         self(string.join(setterm))
-        self.set_string('output', filename)
         # replot the current figure (to the printer):
         self.refresh()

to prevent pictures being overwritten with the next hardcopy command. Anybody
sees that too (eg in new test.py I write a .svg file after ps, and this
overwrites the gp_test.ps file too with the svg file without above patch)?

As Michael said, we should try to keep the test suit working. I'll look into the
errors I get when time. These errors are present in 1.7 on my box too however.
Should somebody have time to look at them or have pointers, let me know. Or
should they not stop a 1.8?

Doing python test.py gives errors:

*  test hardcopy:

Testing hardcopy options: enhanced=1
Press return to show results...

gnuplot> set terminal postscript default enhanced
                                         ^
         line 714: extraneous argument in set terminal postscript

*   test GridData and compute_GridData:

gnuplot> set parametric
gnuplot> set data style lines
gnuplot> set hidden
gnuplot> set contour base
gnuplot> set xlabel "x"
gnuplot> set ylabel "y"
gnuplot> splot "/tmp/tmpdFzv2L.gnuplot/fifo" notitle
Same thing, saved to a file
Press return to show results...
Notice: Cannot contour non grid data. Please use "set dgrid3d".

gnuplot> splot "/tmp/tmpvEBqT_"
Same thing, inline data
Press return to show results...
Notice: Cannot contour non grid data. Please use "set dgrid3d".

gnuplot> splot "-" notitle
The same thing using binary mode
Press return to show results...
Notice: Cannot contour non grid data. Please use "set dgrid3d".

gnuplot> splot "/tmp/tmpcCCPct.gnuplot/fifo" binary notitle
Same thing, using binary mode and an intermediate file
Press return to show results...
         line 1903: File doesn't factorize into full matrix


Well, that's all for now folks.

Benny

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

bm...@ca... wrote:
> So, I'd try to have a look at the distribution part of gnuplot.py, and bugs
> should they come up.
> I'm user bmcage on sourceforge, developer on the GRAMPS python program
> already
> (you can check it's ChangeLog for Benny). No packaging or other
> distribution
> experience though. I do numerical computing (C++, python, ... ) mostly
> for my
> own needs only.
> Let me know what I should know about making a source package, or if you
> want
> more credentials from me first.

Great!  I just added you as a Gnuplot.py developer.  I think I gave you
all of the permissions that you should need; if not, just let me know.

There is information about making releases of Gnuplot.py in the file
RELEASES.txt in the source tree.  If you have any questions, please
direct them to the mailing list and I'll see if I can remember anything
more :-)

Please note that Gnuplot.py has a test suite (albeit not fully
automated).  Please try to keep it working :-)

Michael

Quoting Michael Haggerty <mh...@al...>:

> Hey, what a brilliant idea!  I never thought of that!  All that is
> missing is for SOMEONE to step forward.  This is an open source project;
> that SOMEONE could be you!
>
> I haven't used Gnuplot.py in years and have only occasionally done tiny
> amounts of maintenance out of some irrational sentimental attachment to
> the project.  Gnuplot.py definitely needs a new maintainer, if only to
> collect some minor changes that have been contributed by the mailing
> list and occasionally make a release.  Volunteers needed!
>
> I'd be happy to get you started making contributions to the project.
> Just let me know that you are interested.  (That applies to everybody
> else, too.)

Ok, I use gnuplot and Gnuplot.py for my work at university. As 
everybody, I have
little time for this and don't mind running from SVN, but on the other hand, I
wouldn't want the project to rot too.

So, I'd try to have a look at the distribution part of gnuplot.py, and bugs
should they come up.
I'm user bmcage on sourceforge, developer on the GRAMPS python program already
(you can check it's ChangeLog for Benny). No packaging or other distribution
experience though. I do numerical computing (C++, python, ... ) mostly for my
own needs only.
Let me know what I should know about making a source package, or if you want
more credentials from me first.

Benny

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

Andreas wrote:
> Now I have spent a whole afternoon, just because I wanted to upgrade
> to python 2.5.1 (from 2.3.5). Everything worked fine, only gnuplot.py
> is not working anymore because of the Numeric -> numpy change. Grrrr...
> 
> I have never before used CVS/SVN, and even went through the hassle
> of installing tortoiseCVS, but still could not check out the newest files
> 
> is this wrong? What is wrong?

Gnuplot.py uses Subversion now, not CVS.  Probably the CVS archive is
still available at SourceForge for read-only access, but it is not up to
date.  Please see the Subversion instructions that somebody else
mentioned for information how to check out the Gnuplot.py source from
Subversion.

> Do I need some "editing rights" or something else to use the CVS/SVN
> of gnuplot-py on sourceforge, or is it open to any SF-user? If not,
> please grant me those rights.

No, everybody has read-only access to this Subversion archive.  If you
want write-access, that is another story.

> And probably easier for users who do not want to spend whole afternoons
> just to get a working version of gnuplot.py, is this solution:
> 
>  PLEASE someone PUBLISH A NEW ARCHIVE 
>  WITH THE NEWEST FILES FROM THE CVS.

Hey, what a brilliant idea!  I never thought of that!  All that is
missing is for SOMEONE to step forward.  This is an open source project;
that SOMEONE could be you!

I haven't used Gnuplot.py in years and have only occasionally done tiny
amounts of maintenance out of some irrational sentimental attachment to
the project.  Gnuplot.py definitely needs a new maintainer, if only to
collect some minor changes that have been contributed by the mailing
list and occasionally make a release.  Volunteers needed!

I'd be happy to get you started making contributions to the project.
Just let me know that you are interested.  (That applies to everybody
else, too.)

Michael

On Fri, 11 Jan 2008, Andreas wrote: 
> I have never before used CVS/SVN, and even went through the hassle 
> of installing tortoiseCVS, but still could not check out the newest files 

SVN and CVS are two different systems.
If you installed SVN, follow the directions here:
<URL:http://sourceforge.net/svn/?group_id=17434>

Cheers,
Alan Isaac

Hello!

Now I have spent a whole afternoon, just because I wanted to upgrade
to python 2.5.1 (from 2.3.5). Everything worked fine, only gnuplot.py
is not working anymore because of the Numeric -> numpy change. Grrrr...

I have never before used CVS/SVN, and even went through the hassle
of installing tortoiseCVS, but still could not check out the newest files



is this wrong? What is wrong?

CVSROOT=:pserver:kn...@gn...:443/svnroot/gnuplot-py

I cannot put the password anywhere, using tortoiseCVS, 
and tortoiseCVS doesn't ask for it.

So I also tried 
CVSROOT=:pserver:ano...@gn...:443/svnroot/gnuplot-py

Do I need some "editing rights" or something else to use the CVS/SVN
of gnuplot-py on sourceforge, or is it open to any SF-user? If not,
please grant me those rights.

please help me someone with this versioning stuff, 
I'd really like to start using it.





And probably easier for users who do not want to spend whole afternoons
just to get a working version of gnuplot.py, is this solution:


 PLEASE someone PUBLISH A NEW ARCHIVE 
 WITH THE NEWEST FILES FROM THE CVS.



Thanks a lot!
Andreas

[CC to mailing list]

Andreas Krueger (NoSpam!) wrote:
> Many thanks for your gnuplot.py, it has already helped me a lot.
> 
> BUT:
> 
> It does not work anymore for the current releases
> 
> * Python 2.5.1
> * Numpy 1.0.4 (Numeric is discontinued after Python 2.4
> 
> Please release a version that works with Python 2.5.1

Please direct questions about Gnuplot.py to the mailing list.  I haven't
worked much on it for years now.  Another person wanted to take over
maintenance, but he hasn't been very visible either.

I believe that the current Subversion trunk version of Gnuplot.py should
solve both of these problems.  (The problem that I know of with Python
2.5 is that "with" has become a keyword.  If there are other problems,
please let us know.)

It would be great if somebody would take over maintenance of Gnuplot.py,
for example, making a new release that includes the changes that are
already in SVN.

Michael

Never mind, got it now - g.plot(Gnuplot.Data(data,with=3D'lines lw 5')) wor=
ked! :)

Hi,

I've been trying to set the linewidth using with (or with_ in
svn)=3D'lines linewidth 5', but I get this error when trying to run my
code:
=20
/usr/lib/python2.5/site-packages/Gnuplot/_Gnuplot.py:441:
DeprecationWarning: raising a string exception is deprecated raise
'option %s is not supported' % (k,) Traceback (most recent call last):
  File "brett.py", line 49, in <module>
    g.plot(data, with_=3D'lines linewidth 5')
  File "/usr/lib/python2.5/site-packages/Gnuplot/_Gnuplot.py", line
269, in plot apply(self.set, (), keyw)
  File "/usr/lib/python2.5/site-packages/Gnuplot/_Gnuplot.py", line
441, in set raise 'option %s is not supported' % (k,)
option with_ is not supported

When I remove the with part, everything runs as it should, except that
the lines are too thin. ;)
What am I doing wrong? I've looked at the documentation, but my code
doesn't seem wrong to me...

Cheers,
Valentin