gnuplot-py-users Mailing List for Gnuplot.py

On Tue, 10 Jan 2006, Michael Haggerty apparently wrote: 
> that leaves the more difficult question of how you can 
> output time data from Gnuplot.py.  Unfortunately, there is 
> no support for time data built in.  You could output the 
> data to a file via another mechanism then plot it using 
> a Gnuplot.File object.  Or you could hack Gnuplot.py to 
> add support for time data. 

Note that Allin Cottrell has done this for gretl, which uses 
gnuplot to plot.  So look there in hopes on not having to 
reinvent the wheel.  Also, the Matplotlib date handling 
routines might be useful; they work remarkably well.

fwiw,
Alan Isaac

> I suggest that you find out how to do what you need to do using gnuplot
> and let us know.  Then we might be able to get Gnuplot.py to do the same
> thing.

  You're right, 
  I'm going to investigate it with naked gnuplot, and you'll know about.
  Thanks everybody

-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/

Thank you very much!
It works like charm!
> 
> >>> import Gnuplot
> >>> Gnuplot.GnuplotOpts.gnuplot_command = \
> ...     "gnuplot -xrm 'gnuplot*line2Color:yellow'"
> 
> (before creating your Gnuplot.Gnuplot object).

Albert



-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/

frank h. wrote:
> I am not sure I understand gnuplot.py completely yet
> but I havent found a way to plot a timeseries, e.g.
> 
> set xdata time
> 
> in gnuplot speak
> 
> how would one go about this with gnuplot.py?

In general, you can send any command to gnuplot by using

>>> g = Gnuplot.Gnuplot()
>>> g('set xdata time')

But that leaves the more difficult question of how you can output time
data from Gnuplot.py.  Unfortunately, there is no support for time data
built in.  You could output the data to a file via another mechanism
then plot it using a Gnuplot.File object.  Or you could hack Gnuplot.py
to add support for time data.

This has been discussed before; see for example

http://sourceforge.net/mailarchive/message.php?msg_id=2172189
http://sourceforge.net/mailarchive/message.php?msg_id=11511622
http://sourceforge.net/mailarchive/message.php?msg_id=2179048

Michael

al...@mm... wrote:
> I'm trying to change the color of a function... and I can't.
> I used to call, for example:
> gnuplot -xrm 'gnuplot*line2Color:yellow', 
> 
> but, of course I can't do this now, so
> how can I do it?
> 
> One more thing. I can't change the file .Xdefaults because I'm not the system
> administrator.
> 
> Thank you very much...
> at least there's somebody on the other side... ;)

Assuming that you are on a Unix/Linux system, it is usually the case
that the login process causes a file in your home directory to be read
and added to the X resources.  For example, on my linux setup the magic
file is called ".Xresources".

Alternatively, you can load properties into your X resources manually
using the "xrdb" command.

Finally, if you only want these properties to be set when you are using
Gnuplot.py, you could add the command-line options to the
gnuplot_command variable as follows:

>>> import Gnuplot
>>> Gnuplot.GnuplotOpts.gnuplot_command = \
...     "gnuplot -xrm 'gnuplot*line2Color:yellow'"

(before creating your Gnuplot.Gnuplot object).

Michael

al...@mm... wrote:
>>4. Try a simple example with "naked" gnuplot, rather than Gnuplot.py,
>>using temporary files to hold the data.  Does it do what you expect?
> 
>     No, it does the same... ;P

I suggest that you find out how to do what you need to do using gnuplot
and let us know.  Then we might be able to get Gnuplot.py to do the same
thing.

Michael

> al...@mm... wrote:
> > I want to use gnuplot.py to represent several "4D" surfaces in the same
> plot, so
> > I write something like:
> >
> >   data1 = [
> >     [
> >         [1,1,z1(1,1),val1(1,1)],
> >         [1,2,z1(1,2),val1(1,2)],
> >         [1,3,z1(1,3),val1(1,3)],
> >     ],
> >     [
> >         [2,1,z1(2,1),val1(2,1)],
> >         [2,2,z1(2,2),val1(2,2)],
> >         [2,3,z1(2,3),val1(2,3)],
> >     ],
> >     ...
> >     ]
> >
> >   data2 = [
> >     [
> >         [1,1,z2(1,1),val2(1,1)],
> >         [1,2,z2(1,2),val2(1,2)],
> >         [1,3,z2(1,3),val2(1,3)],
> >     ],
> >     [
> >         [2,1,z2(2,1),val2(2,1)],
> >         [2,2,z2(2,2),val2(2,2)],
> >         [2,3,z2(2,3),val2(2,3)],
> >     ],
> >     ...
> >     ],
> >
> >   ...
> >
> >   dataN = ...
> >
> >   and then, data = [Gnuplot.Data(data1, using=(1,2,3,4), inline = 1),
> >                     Gnuplot.Data(data2, using=(1,2,3,4), inline = 1),
> >                     ...
> >                     Gnuplot.Data(dataN, using=(1,2,3,4), inline = 1)]
> >
> >   So, when I go to plot:
> >     g = Gnuplot.Gnuplot(debug = 1)
> >     g.splot(*data)
> >
> >   it seems something works because you can read:
> >   gnuplot> splot '-' using 1:2:3:4 notitle, '-' using 1:2:3:4 notitle, '-'
> using
> > 1:2:3:4 notitle, '-' using 1:2:3:4 notitle, '-' using 1:2:3:4 notitle
> >   on your terminal, but only the last surface is plotted, maybe because
> the
> > (x,y) pairs are the same in data1, data2, ...
> >
> >   So, how can I solve this?
> >   My target is to see all the 3D points in different colors or something
> like
> > that... do you understand?
>
> I'm a bit out of practice.  How do you expect the four values to be
> used?  Usually splot uses (x,y,z) triplets, no?
    I expect to see surfaces by (x,y,z) coloured in function of val, i.e., if
val is high I want to see the point in yellow and if it is low in black.
    
> Otherwise I don't see why this shouldn't work.  I can't imagine that it
> would be a problem that the (x,y) pairs are the same.  Ideas:
>
> 1. If you have hidden line removal on (which is the default, I think) it
> could be that all of your surfaces are being plotted but the top surface
> is visually obscuring the other surfaces.
    How do I know it? I tried with set & unset hidden3d, but it looks like the
same...  

>
> 2. If your z values are all the same, then you would not see distinct
> surfaces.
    You're right, but they aren't the same  
  
> 3. Try the same thing without the inline=1 option.  Does that work?
> (This would imply a Gnuplot.py bug.)
    It works the same.

> 4. Try a simple example with "naked" gnuplot, rather than Gnuplot.py,
> using temporary files to hold the data.  Does it do what you expect?
    No, it does the same... ;P
 
> Michael


In fact I can see the points colored typping
 splot  "temp.plot" using 1:2:3:4 with points pointtype 7 palette
but with no surface... (temp.plot is a file with the same data)

Maybe I shoud read the entire gnuplot tutorial... but I find it really boring...
;)

thank you very much!
Albert








-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/