gnuplot-py-users Mailing List for Gnuplot.py

Matthew Fulmer wrote:
> I made the changes you requested it and tested it via test.py. I
> also replaced occurences of the deprecated apply function with
> the recommended syntax. I did this because it makes it easier to
> call the _NewFileItem with keyword parameters that are not part
> of keyw (like filename). I will reverse that change if necessary
> (tested with python 2.4).
> 
> I am subscribed to the mailing list now (I had a procmail bug
> that sent the subscription confirmation to the wrong place).
> Here is the new patch

Thanks again.  I just committed this patch to CVS.

Michael

I think Python 2.2 is quite reasonable. In fact, I would say that Python
2.3 is quite reasonable too, but maybe some platforms still don't
support it?

I mean, if some software is still using a version of Python that came
out X years ago, is it still being actively maintained?

Regards,
 Noel

On Thu, 2005-12-15 at 23:12 -0500, Alan G Isaac wrote:
> On Thu, 15 Dec 2005, Michael Haggerty apparently wrote: 
> > I also ask whether there are people who would be upset by 
> > jumping the Python version requirement directly to Python 
> > 2.2. 
> 
> How about 2.3?  ;-)
> http://www.python.org/2.3/highlights.html
> Anyway, go for it!
> 
> Cheers,
> Alan Isaac
> 
> 
> 
> 
> 
> 
> -------------------------------------------------------
> 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
> _______________________________________________
> Gnuplot-py-users mailing list
> Gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-py-users

On Thu, 15 Dec 2005, Michael Haggerty apparently wrote: 
> I also ask whether there are people who would be upset by 
> jumping the Python version requirement directly to Python 
> 2.2. 

How about 2.3?  ;-)
http://www.python.org/2.3/highlights.html
Anyway, go for it!

Cheers,
Alan Isaac

Matthew Fulmer wrote:
> I made the changes you requested it and tested it via test.py. I
> also replaced occurences of the deprecated apply function with
> the recommended syntax. I did this because it makes it easier to
> call the _NewFileItem with keyword parameters that are not part
> of keyw (like filename). I will reverse that change if necessary
> (tested with python 2.4).
> 
> I am subscribed to the mailing list now (I had a procmail bug
> that sent the subscription confirmation to the wrong place).
> Here is the new patch

Thanks for the revised patch.  I may not have time to look it over
before leaving on a Xmas trip to America, but I'll try to get to it
early in January.

Lurkers should also feel free to jump in and help reviewing this (or any
other) code.

Yours,
Michael

Hello,

Presently, I think that Gnuplot.py should work with Python 1.5.2.

Who would object to requiring Python version 2.0 for Gnuplot.py?

Who would object to requiring Python version 2.2 for Gnuplot.py?

The reason for asking this question now is that Matthew Fulmer has
submitted a patch with aspects that would require Python 2.0.  These
aspects could be removed if necessary, but it is probably reasonable to
require 2.0 anyway.

Python 2.2 has been out almost as long and added a few other nice
features; therefore I also ask whether there are people who would be
upset by jumping the Python version requirement directly to Python 2.2.

Speak now or forever hold your peace.

Michael

I made the changes you requested it and tested it via test.py. I
also replaced occurences of the deprecated apply function with
the recommended syntax. I did this because it makes it easier to
call the _NewFileItem with keyword parameters that are not part
of keyw (like filename). I will reverse that change if necessary
(tested with python 2.4).

I am subscribed to the mailing list now (I had a procmail bug
that sent the subscription confirmation to the wrong place).
Here is the new patch

-- 
Matthew Fulmer

Matthew Fulmer wrote:
> Attached is a patch that adds the keyword parameter 'filename'
> to the functions GridData and Data. Passing this parameter saves
> the passed data to a permanent file instead of a temporary file
> or a fifo. Note that I also changed the name of the class
> _TempFileItem to _NewFileItem since it now handles all non-fifo
> file creation, not just temporary files. 
> 
> Please CC me in your reply; I was unable to subscribe to the
> list due to (I believe) the current Sourceforge flux.

Thanks for the patch!  I think that this is a good idea in general, and
I'd like to incorporate it into Gnuplot.py.  But there are a few things
to be fixed before I could merge your change:

1. Please use spaces instead of tabs for indenting Python code, for
consistency with the rest of the code.  And please limit line lengths to
79 characters and uniformly use 4 spaces for indenting.

2. I think that code like this

>      def __init__(self, content, **keyw):
> +        if keyw.has_key('filename'):
> +            filename = keyw['filename']
> +            del keyw['filename']
> +            if filename:
> +                self.temp = False
> +            else:
> +                filename = tempfile.mktemp()
> +                self.temp = True
> +        else:
> +            filename = tempfile.mktemp()
> +            self.temp = True

can be more succinctly written like this:

>      def __init__(self, content, filename=None, **keyw):
> +        if filename:
> +            self.temp = False
> +        else:
> +            filename = tempfile.mktemp()
> +            self.temp = True

3. Please include a patch to the test.py script to test the new feature.

4. Please include a patch to NEWS.txt telling what's new.

Also see notes below within your patch.

> ------------------------------------------------------------------------
> 
> Index: PlotItems.py
> ===================================================================
> RCS file: /cvsroot/gnuplot-py/gnuplot-py/PlotItems.py,v
> retrieving revision 2.16
> diff -u -p -r2.16 PlotItems.py
> --- PlotItems.py	7 Nov 2005 21:10:00 -0000	2.16
> +++ PlotItems.py	10 Dec 2005 08:11:21 -0000
> @@ -302,6 +302,7 @@ class _FileItem(PlotItem):
>  
>          self.filename = filename
>  
> +        # Use single-quotes so that pgnuplot can handle DOS filenames:
>          apply(PlotItem.__init__, (self,), keyw)
>  
>      def get_base_command_string(self):

I don't understand this comment in this context.

> @@ -336,8 +337,20 @@ class _FileItem(PlotItem):
>              self._options['binary'] = (0, None)
>  
>  
> -class _TempFileItem(_FileItem):
> +class _NewFileItem(_FileItem):
>      def __init__(self, content, **keyw):
> +	if keyw.has_key('filename'):
> +	    filename = keyw['filename']
> +	    del keyw['filename']
> +	    if filename:
> +		self.temp = False
> +	    else:
> +		filename = tempfile.mktemp()
> +		self.temp = True
> +	else:
> +	    filename = tempfile.mktemp()
> +	    self.temp = True
> +
>          binary = keyw.get('binary', 0)
>          if binary:
>              mode = 'wb'

I don't see how this part of the patch can work, as filename is
overwritten a few lines later.

> @@ -359,13 +372,14 @@ class _TempFileItem(_FileItem):
>  
>          # If the user hasn't specified a title, set it to None so
>          # that the name of the temporary file is not used:
> -        if not keyw.has_key('title'):
> +        if not keyw.has_key('title') and self.temp:
>              keyw['title'] = None
>  
>          apply(_FileItem.__init__, (self, filename,), keyw)
>  
>      def __del__(self):
> -        os.unlink(self.filename)
> +    	if self.temp:
> +            os.unlink(self.filename)
>  
>  
>  class _InlineFileItem(_FileItem):
> @@ -529,6 +543,8 @@ def Data(*set, **keyw):
>              rather than through a temporary file.  The default is the
>              value of gp.GnuplotOpts.prefer_inline_data.
>  
> +        'filename=<string>' -- save data to a permanent file
> +
>      The keyword arguments recognized by '_FileItem' can also be used
>      here.
>  
> @@ -558,7 +574,17 @@ def Data(*set, **keyw):
>              cols = (cols,)
>          set = Numeric.take(set, cols, -1)
>  
> +    if keyw.has_key('filename'):
> +        filename = keyw['filename']
> +	if not filename:
> +	    del keyw['filename']
> +	    filename = None
> +    else:
> +        filename = None
> +
>      if keyw.has_key('inline'):
> +	if filename:
> +            raise Errors.OptionError('cannot pass data both inline and via a file')
>          inline = keyw['inline']
>          del keyw['inline']
>      else:

I think that the same OptionError test needs to be done a few lines
later, if gp.GnuplotOpts.prefer_inline_data is set.  Or, better yet,
invert the test to check for filename before considering the inline option:

    if filename:
        if keyw.get('inline', 0):
            raise ...
    elif keyw.has_key('inline'):
        inline = keyw['inline']
        del keyw['inline']
    else:
        inline = gp.GnuplotOpts.prefer_inline_data

> @@ -570,10 +596,10 @@ def Data(*set, **keyw):
>      content = f.getvalue()
>      if inline:
>          return apply(_InlineFileItem, (content,), keyw)
> -    elif gp.GnuplotOpts.prefer_fifo_data:
> +    elif (not filename) and gp.GnuplotOpts.prefer_fifo_data:
>          return apply(_FIFOFileItem, (content,), keyw)
>      else:
> -        return apply(_TempFileItem, (content,), keyw)
> +        return apply(_NewFileItem, (content,), keyw)
>  
>  
>  def GridData(data, xvals=None, yvals=None, inline=_unset, **keyw):
> @@ -596,6 +622,8 @@ def GridData(data, xvals=None, yvals=Non
>  
>          'inline=<bool>' -- send data to gnuplot "inline"?
>  
> +        'filename=<string>' -- save data to a permanent file
> +
>      Note the unusual argument order!  The data are specified *before*
>      the x and y values.  (This inconsistency was probably a mistake;
>      after all, the default xvals and yvals are not very useful.)
> @@ -658,8 +686,22 @@ def GridData(data, xvals=None, yvals=Non
>      binary = keyw.get('binary', 1) and gp.GnuplotOpts.recognizes_binary_splot
>      keyw['binary'] = binary
>  
> +    if keyw.has_key('filename'):
> +        filename = keyw['filename']
> +	if not filename:
> +	    del keyw['filename']
> +	    filename = None
> +    else:
> +        filename = None
> +
> +    binary = keyw.get('binary', 1) and gp.GnuplotOpts.recognizes_binary_splot
> +    keyw['binary'] = binary
> +
>      if inline is _unset:
> -        inline = (not binary) and gp.GnuplotOpts.prefer_inline_data
> +        inline = (not binary) and (not filename) and gp.GnuplotOpts.prefer_inline_data
> +    elif inline:
> +       if filename:
> +            raise Errors.OptionError('cannot pass data both inline and via a file')
>  
>      # xvals, yvals, and data are now all filled with arrays of data.
>      if binary:
> @@ -686,10 +728,10 @@ def GridData(data, xvals=None, yvals=Non
>              mout[1:,1:] = Numeric.transpose(data.astype(Numeric.Float32))
>  
>          content = mout.tostring()
> -        if gp.GnuplotOpts.prefer_fifo_data:
> +        if (not filename) and gp.GnuplotOpts.prefer_fifo_data:
>              return apply(_FIFOFileItem, (content,), keyw)
>          else:
> -            return apply(_TempFileItem, (content,), keyw)
> +            return apply(_NewFileItem, (content,), keyw)
>      else:
>          # output data to file as "x y f(x)" triplets.  This
>          # requires numy copies of each x value and numx copies of
> @@ -709,9 +751,9 @@ def GridData(data, xvals=None, yvals=Non
>  
>          if inline:
>              return apply(_InlineFileItem, (content,), keyw)
> -        elif gp.GnuplotOpts.prefer_fifo_data:
> +        elif (not filename) and gp.GnuplotOpts.prefer_fifo_data:
>              return apply(_FIFOFileItem, (content,), keyw)
>          else:
> -            return apply(_TempFileItem, (content,), keyw)
> +            return apply(_NewFileItem, (content,), keyw)
>  
>  

Jeremy Conlin wrote:
> I would like to specify the point size of my plotted data.  I have this:
> 
> Histo = Gnuplot.Data(Bins,Bincounts, with='points')
> 
> I would like my pointsize to be 0.5.  I tried doing:
> 
> Histo = Gnuplot.Data(Bins,Bincounts, with='points', pointsize=0.5)
> 
> but got
> 
> Gnuplot.Errors.OptionError: pointsize=0.5
> 
> What do I have to do to specify a pointsize for my plot(s)?

"pointsize" is not supported as a separate option, but you can treat it
as part of the "with" argument:

    Histo = Gnuplot.Data(Bins,Bincounts, with='points pointsize 0.5')

Undoubtedly this could be improved, but unfortunately I don't have time
to work on Gnuplot.py.

Michael

I would like to specify the point size of my plotted data.  I have this:

Histo =3D Gnuplot.Data(Bins,Bincounts, with=3D'points')

I would like my pointsize to be 0.5.  I tried doing:

Histo =3D Gnuplot.Data(Bins,Bincounts, with=3D'points', pointsize=3D0.5)

but got

Gnuplot.Errors.OptionError: pointsize=3D0.5

What do I have to do to specify a pointsize for my plot(s)?
Thanks,
Jeremy

Attached is a patch that adds the keyword parameter 'filename'
to the functions GridData and Data. Passing this parameter saves
the passed data to a permanent file instead of a temporary file
or a fifo. Note that I also changed the name of the class
_TempFileItem to _NewFileItem since it now handles all non-fifo
file creation, not just temporary files. 

Please CC me in your reply; I was unable to subscribe to the
list due to (I believe) the current Sourceforge flux.

-- 
Matthew Fulmer

I wrote this python script to read a file and get data from it to
plot.  However, it makes all the plot colors the same fuschia color.=20
That, of course, doesn't help me very much.  I have done a similar
script that automatically creates different colors for each plot.  I
was wondering if someone could help me discover what is wrong.
I have pasted my script below.
Thanks,
Jeremy


  1 #! /usr/local/bin/python2.4
  2 import string, re, sys
  3 import Gnuplot
  4
  5 print "Searching... ", sys.argv[1]
  6 Output =3D file(sys.argv[1],'r')
  7
  8 OutLines =3D Output.readlines()
  9
 10 tallysearch =3D r'\ +(\d\.\d{4,5}E[+-]\d{2})\
+(\d\.\d{4,5}E[+-]\d{2})\ ?(\d\.\d{4})'
 11 #                  \ +(\d\.\d{4}E[+-]\d{2})
 12 #                  \ ?(\d\.\d{4})'''
 13 TallySearch =3D re.compile(tallysearch)
 14
 15 Result =3D []
 16 Tallies =3D 0
 17 result =3D {}
 18 result['Energy'] =3D []
 19 result['Flux'] =3D []
 20 result['Error'] =3D []
 21 Result.append(result)
 22
 23 print Result[Tallies]
 24
 25 for line in OutLines:
 26     Tally =3D TallySearch.match(line)
 27     if Tally:
 28         Result[Tallies]['Energy'].append(float(Tally.group(1)))
 29         Result[Tallies]['Flux'].append(float(Tally.group(2)))
 30         Result[Tallies]['Error'].append(float(Tally.group(3)))
 31 #       Result[Tallies].append([energy, flux, error])
 32
 33         if Tally.group(1) =3D=3D '1.0000E+01':
 34            Tallies =3D Tallies+1
 35            Result.append(result)
 36
 37 GP =3D Gnuplot.Gnuplot(debug=3D1)
 38
 39 g0 =3D Gnuplot.Data(Result[0]['Energy'], Result[0]['Flux'],
title=3D'50 cm', with=3D'lines')
 40 g2 =3D Gnuplot.Data(Result[2]['Energy'], Result[2]['Flux'],
title=3D'100 cm', with=3D'histeps')
 41 g4 =3D Gnuplot.Data(Result[4]['Energy'], Result[4]['Flux'],
title=3D'150 cm', with=3D'histeps')
 42 g6 =3D Gnuplot.Data(Result[6]['Energy'], Result[6]['Flux'],
title=3D'200 cm', with=3D'histeps')
 43 g1 =3D Gnuplot.Data(Result[1]['Energy'], Result[1]['Flux'],
title=3D'50 cm', with=3D'histeps')
 44 g3 =3D Gnuplot.Data(Result[3]['Energy'], Result[3]['Flux'],
title=3D'100 cm', with=3D'histeps')
 45 g5 =3D Gnuplot.Data(Result[5]['Energy'], Result[5]['Flux'],
title=3D'150 cm', with=3D'histeps')
 46 g7 =3D Gnuplot.Data(Result[7]['Energy'], Result[7]['Flux'],
title=3D'200 cm', with=3D'histeps')
 47 GP('set logscale y')
 48 GP.xlabel('Energy (MeV)')
 49 GP.ylabel('Flux (cm^{-2})')
 50 GP.plot(g0,g2,g4,g6)