gnuplot-py-users Mailing List for Gnuplot.py

okay... i've googled for about an hour and still don't seem my answer...

I am using "arrows" to plot a time line that has

|-------------|

This kind of style. But since the time line is long i would also like 
to include labels in my data that would show the start -duration - end 
of my data in better precision that i can eyeball it looking at the 
xlabel...

so that i might get something that looks like :


    1.121            (2.1)              3.221
       |-----------------------------|





# ------------------------------[ snip ] 
---------------------------------

#! /usr/bin/env python

"""arrows.py -- a plot of the durations

Usage: $ python arrows.py data.txt


"""

import sys
import Gnuplot

def main():
     # First let's get our data form an external file and stuff it in a 
list
     [filename] = sys.argv[1:]
     all = []
     for l in open(filename).readlines():
		l = l.strip().split()
		all.append( [ float(l[0]), float(l[1]), float(l[2]), int(l[3]) ] )
     # Split the data into four datasets by stream
     # Each data set will be drawn in 'vector' style with a different 
color
     allData = []    # This will contain the four datasets
     for stream in [1, 2, 3, 4]:
         # Select the raw data for the stream
         rawData = [ item for item in all if item[3] == stream ]
         # Make a Gnuplot.Data to contain the raw data
         data = Gnuplot.Data(rawData,
             using=(1, 3, 2, '(0)'), # This gives 1-based index into the
                                     # data for x, y, xdelta, ydelta
                                     # ydelta is a literal 0
             with='vectors arrowstyle %s' % stream)  # This sets the 
style of the dataset styles are defined below
             #title='Voice %s' % stream) # This names the dataset for 
the legend
         allData.append(data)

     # Set up general plot parameters
     g = Gnuplot.Gnuplot(debug=1)
     g.title('Overview') # (optional)
     g.xlabel('Time')
     g.ylabel('Event')
     g('set grid')
     #g('set xtics (10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 
130, 140)')
     g('set ytics (0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12)')

     # These lines create four vector styles that have bars at each end
     # The only difference between the styles is the line type they use,
     # which determines the color
     # Shamelessly cribbed from the gnuplot arrowstyle demo
     g('set style line 1 lt 1 lw 2') # Set a line style for the vectors
     g('set style line 2 lt 2 lw 2')
     g('set style line 3 lt 3 lw 2')
     g('set style line 4 lt 4 lw 2')

     # Set an arrow style for the vectors
     g('set style arrow 1 heads size screen 0.008,90 ls 1')
     g('set style arrow 2 heads size screen 0.008,90 ls 2')
     g('set style arrow 3 heads size screen 0.008,90 ls 3')
     g('set style arrow 4 heads size screen 0.008,90 ls 4')

     #g('set key outside box')   # Include a legend; put it outside the 
graph

     # This actually does the plot
     # The * makes it treat the elements of allData as individual 
function arguments
     # It is the same as g.plot(allData[0], allData[1], allData[2], 
allData[3]
     g.plot(*allData)


if __name__ == '__main__':
     main()


# data here:

1.0 52.98151020408163 1 1
53.98151020408163 360.0 2 2
153.98151020408164 100.0 3 3
254.61807619047627 60.0 4 4
316.94841360544211 60.0 5 1
378.66877120181414 30.0 6 2
409.28659297052155 30.0 7 3
440.74881179138322 24.0 8 4
465.75744104308387 60.0 9 1
496.08856235827682 60.0 10 2
557.0976548752833 60.0 11 3

Hi Michael,

While as an occasional user I cannot offer to be maintainer, 
I can offer to help a new maintainer test numpy support.

Also, I suggest posting your query to the SciPy list,
as my impression is that a numer of Gnuplot.py users
lurk there.

Cheers,
Alan Isaac

Hello,

It should be obvious to everybody by now that I haven't had the time or
energy to invest in Gnuplot.py for quite some time now.  When I
established the project, I used it almost daily.  But since I started my
current job five years ago, I haven't used Gnuplot.py more than a
handful of times.  And now that I am involved in another open-source
project (cvs2svn), I want to spend my limited available time on that
project.

For a long time this wasn't much of a problem, because Gnuplot.py did
most of the things people needed and didn't require much maintenance.
(Not that I can't think of nice new features to add, but...)

But now there are a few things that really need to be done:

- Support for numarray.  This is a relatively small job but very important.

- Get Gnuplot.py listed on http://www.python.org/pypi (The Python Cheese
Shop).

- Updated the documentation and website to point at the new Subversion
repository rather than the old CVS one.

- Look over and possibly integrate some patches that have been submitted
to the mailing list over the past months and years.

- Make a new release.  It's been ages, and a few things have changed.

Sooooo.....

Is there anybody out there who can program in Python and would be
interested in taking over the maintenance of Gnuplot.py?  I'd be happy
to help you get up to speed and answer occasional questions.

Yours,
Michael

Hello.

I've been having trouble using multiple replot commands, ie:

g.plot(Gnuplot.Data(dataset[0]))
for i in xrange(1,len(dataset)):
     g.replot(Gnuplot.Data(dataset[i]))

After about the 40th row in dataset, I start getting errors like this:

gnuplot> 1961992.5 0.0 2.47496008873
          ^
          line 31995: invalid command

gnuplot> e

(if inline == 1) and sometimes about bad filenames (seems to be system 
dependent) if inline = 0.  Eventually the plot seems to come out correctly, 
but I can't verify that all the rows are actually being plotted.  Is this a 
bug, or am I going about this all wrong?  I want to overlay a bunch (~100) 
of different lines on the same plot, and then possibly an averaged line in 
dark black on top of that.

I can put sample code and plots up somewhere for reference...

Oh, this is on Debian Stable (but the bad filename errors tend to crop up on 
a Gentoo box, which I don't have administrative access to).

Thanks for your help.
John

-- 
*************************
John Parejko
Department of Physics and Astronomy
215 895-2786
Drexel University
Philadelphia, PA
**************************

Adriaan . wrote:
> I am looking to do some curve fitting and I am wondering if gnuplot-py
> will be accept the gnuplot curve fit commands via similar methods as you
> would generate a normal plot (ie. using the Gnuplot.Gnuplot() object)?
> 
> Has anyone tried this or know if it is implemented?

This has been discussed on the mailing list a couple of times before.
Please search the archive.

The bottom line is that there is no built-in support in Gnuplot.py for
the "fit" function, but it shouldn't be much effort to invoke it
explicitly and read the output back into Python.

Michael

Hi all,

I am looking to do some curve fitting and I am wondering if gnuplot-py will 
be accept the gnuplot curve fit commands via similar methods as you would 
generate a normal plot (ie. using the Gnuplot.Gnuplot() object)?

Has anyone tried this or know if it is implemented?

Cheers,
Adriaan