gnuplot-py-users Mailing List for Gnuplot.py

chuck clark writes:
> I think my problem lies in the fact of I'm not sure which PlotItem
> type from gnuplot-py I should use.  I was originally trying to use
> Data but of course "12:01" can't be converted to a Numeric Float
> array.  So now I'm taking a step back and wondering if I need to
> create my own PlotItem types of Date, Time and DateTime to do this?
> [...]
> gnuplot> set xdata time
> gnuplot> set timefmt "%H:%M"
> gnuplot> set format x "%H:%M"
> gnuplot> set xrange ["12:00":"12:05"]
> gnuplot> plot "/home/cclark/test.dat" using 1:2

A PlotItem object represents a whole plot including all of the
associated data (if any), and the plot sub-command that has to be sent
to gnuplot (e.g., '"/tmp/sxhhfy" using 1:3 with linespoints').  You
can use any of the data-based PlotItems defined in the CVS version
(_InlineFileItem, _FIFOFileItem, or _TempFileItem); you just have to
call its constructor using the content that needs to be sent to
gnuplot as one of the arguments.  For example, this should work
(untested):

>>> content = """\
... 12:01    4.63
... 12:02    6.74
... 12:03    0.0
... 12:04    6.32
... """
>>> plot_item = PlotItems._TempFileItem(content, with='linespoints')
>>> g = Gnuplot.Gnuplot()
>>> g('set xdata time')
>>> g('set xdata time')
>>> g('set timefmt "%H:%M"')
>>> g('set format x "%H:%M"')
>>> g('set xrange ["12:00":"12:05"]')
>>> g.plot(plot_item)
>>> 

The only trick is getting your data into a "content" string that
gnuplot can understand.  You can do that any way you want.
PlotItems.Data() is now a helper function that does mainly that; you
can use it as a starting point.

I haven't decided whether to make the _InlineFileItem etc classes a
part of the Gnuplot.py published interface, which is why their names
start with '_'.  I might rename them before the next Gnuplot.py
release.

Michael

--
Michael Haggerty
hag...@jp...

Leo,
The mxdate package which Michael speaks of is a great package indeed and
handles datetime data very well and has almost any date/time functions you
could think of.  I think my problem lies in the fact of I'm not sure which
PlotItem type from gnuplot-py I should use.  I was originally trying to use
Data but of course "12:01" can't be converted to a Numeric Float array.  So
now I'm taking a step back and wondering if I need to create my own PlotItem
types of Date, Time and DateTime to do this?

I thought about doing as you say about converting my hh:mm:ss string into a
float but it isn't really clean because time is base 60 and not decimal and
I lose a lot of readability because suddenly my x-axis is not regular times
that the user would expect to see but some decimal representation of it.  If
I somehow try to preserve the time by making 12:01 into 1201 then I'll have
funny gaps from 1260 to 1300 and on every other hour.

Gnuplot itself handles Date/Time types as seen in the 3.7.1 manual.  If you
had the data below in a file called test.dat you could plot it in Gnuplot
with the following:

12:01    4.63
12:02    6.74
12:03    0.0
12:04    6.32

gnuplot> set xdata time
gnuplot> set timefmt "%H:%M"
gnuplot> set format x "%H:%M"
gnuplot> set xrange ["12:00":"12:05"]
gnuplot> plot "/home/cclark/test.dat" using 1:2

So I guess what I was wondering was if anyone had plotted date/time data
using gnuplot-py with the existing PlotItem types or if people who were more
familiar with the package than I am thought I'd need to come up with a new
type.  Any thoughts would be greatly appreciated.

cheers,
chuck

Leonardo Milano writes:
> Is there a date-format data type in Python ?. I browsed the docs
> and couldn't find one. Or is there a class for that ?.

There is a package called mx.DateTime or something like that with a
vary capable date class, but unfortunately it is not part of standard
Python.  Standard Python tends to use doubles or tuples to represent
date/times and the time module has lots of routines for manipulating
dates/times.

Michael

--
Michael Haggerty
hag...@jp...

Lothar Birk writes:
> > It looks like you are using a home-cooked extension to Gnuplot.py,
> > so I could only guess what your problem might be.  Possibly the
> > argument that you are passing to gnuplot.__call__ is not a
> > sensible string as it should be.
> 
> Yes it is a crooked extension, which maintains a list of all gnuplot
> commands and allows you to build up a plot command by adding more
> and more curves which finally get plotted into one diagrams.

Just a quick note: you can add PlotItems one after another in the
standard Gnuplot.py by calling Gnuplot.replot() with additional
arguments as documented in _Gnuplot.py.  If you want to store other
kinds of commands too (though I don't know why because other commands
are not cumulative) you could use the "filename=" argument to the
Gnuplot constructor.

Michael

--
Michael Haggerty
hag...@jp...

Hi Chuck

> > Is there a way to plot Time/Data data in Gnuplot-py?
> > I've got some two dimensional data that looks like
> > the following:
> > 
> > 12:01:30    4.63
> > 12:02:30    6.74
> > 12:03:30    0.0
> > 12:04:30    6.32

[...]

> I don't have any experience myself with using date-formats in gnuplot,

Is there a date-format data type in Python ?. I browsed the docs
and couldn't find one. Or is there a class for that ?.

Anyway, it seems that the more straightforwars way to go would
be to write a small python function that converts a string "hh:mm:ss"
into a floating point. It should be doable in a few lines
of code.

A little fancier would be a date class, with conversion methods.
But may be that is overkill depending on what you need.

Hope this helps,
-- leo

Hello Micheal,

thanks for the quick answer and pointing me to the gnuplot-py-user list.
I somehow failed to notice there is one.

Michael Haggerty wrote:
> Lothar,
> 
> [I have cc'ed this email to the gnuplot-py-users mailing list.  Please
> subscribe to that list if you are interested in following Gnuplot.py
> developments:
> 
>     http://lists.sourceforge.net/lists/listinfo/gnuplot-py-users
> 
> .]
> 
> Lothar Birk <bi...@en...> writes:
> 
>>Dear Michael,
>>
>>first of all, thank your for your great tool which is very useful to
>>me creating plots on the fly for my lectures and papers.
>>
>>I recently had to move to an Win2k PC :-( and try to get some of my
>>Python scripts running there,
>>In all my scripts using Gnuplot I run into the following error message.
>>
>>Traceback (most recent call last):
>>   File "C:\Home\birk\src\pyprop\b-series\TESTBS~1.PY", line 46, in ?
>>     PDlist=[0.5,0.6,0.8,1.0,1.2,1.4])
>>   File "C:\Home\birk\src\pyprop\b-series\propseries.py", line 269, in 
>>plotKTKQ
>>     g.additem('set xrange [%f:%f]' % (Jmin,Jmax))
>>   File "C:\Home\birk\src\pyprop\b-series\IncrGnuplot.py", line 173, in 
>>additem
>>     self.__call__(plotitem)
>>   File "C:\program 
>>files\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 206, in __call__
>>     self.gnuplot(s)
>>   File "C:\PROGRA~1\python22\lib\site-packages\Gnuplot\gp_win32.py", 
>>line 121, in __call__
>>     self.write(s + '\n')
>>IOError: [Errno 22] Invalid argument
>>
>>The weird thing is that it does not happen at the same place in the
>>script at different runs from a command terminal. May be this error
>>is not a Gnuplot.py issue at all, but related to system or python 
>>installation. I hope, however, you can provide me with hints where to look.
>>
>>I have python-2.2.1, gnuplot-py-1.6.1, Numeric-21.3 and a lot of
>>other stuff which runs as far as I know.
>>
>>Best regards
>>Lothar
> 
> 
> I don't think there's ever been a version 1.6.1 of Gnuplot.py
> released.  Are you using 1.6 or a more recent CVS version or ...?

OK it is the plain 1.6 distribution. I got confused by the numbering
of the rpms on sourceforge (1.6-1)

> 
> It looks like you are using a home-cooked extension to Gnuplot.py, so
> I could only guess what your problem might be.  Possibly the argument
> that you are passing to gnuplot.__call__ is not a sensible string as
> it should be.
> 
Yes it is a crooked extension, which maintains a list of all gnuplot
commands and allows you to build up a plot command by adding more and
more curves which finally get plotted into one diagrams.

However the error also occurs when I try to run the Gnuplot.py
demo:

C:\Program Files\python22\Lib\site-packages\Gnuplot>python demo.py
gnuplot> set title "A simple example"
gnuplot> set data style linespoints
gnuplot> plot 'C:\DOCUME~1\birk\LOCALS~1\Temp\~1200-0' notitle
Please press return to continue...

Traceback (most recent call last):
   File "demo.py", line 121, in ?
     demo()
   File "demo.py", line 47, in demo
     g.reset()
   File "C:\program 
files\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 3
62, in reset
     self('reset')
   File "C:\program 
files\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 2
06, in __call__
     self.gnuplot(s)
   File "C:\PROGRA~1\python22\lib\site-packages\Gnuplot\gp_win32.py", 
line 121, i
n __call__
     self.write(s + '\n')
IOError: [Errno 22] Invalid argument

> If you send us more information we might be able to help you.
> 
> Michael
> 
> --
> Michael Haggerty
> hag...@jp...
> 
Any help appreciated
Lothar
-- 
-- Lothar Birk -------------------------------------------------------
Visiting Professor, Dept. of Naval Architecture and Marine Engineering
University of Michigan
Phone +1 734 615 8061, Fax +1 734 936 8820, E-mail bi...@um...

Lothar,

[I have cc'ed this email to the gnuplot-py-users mailing list.  Please
subscribe to that list if you are interested in following Gnuplot.py
developments:

    http://lists.sourceforge.net/lists/listinfo/gnuplot-py-users

.]

Lothar Birk <bi...@en...> writes:
> Dear Michael,
> 
> first of all, thank your for your great tool which is very useful to
> me creating plots on the fly for my lectures and papers.
> 
> I recently had to move to an Win2k PC :-( and try to get some of my
> Python scripts running there,
> In all my scripts using Gnuplot I run into the following error message.
> 
> Traceback (most recent call last):
>    File "C:\Home\birk\src\pyprop\b-series\TESTBS~1.PY", line 46, in ?
>      PDlist=[0.5,0.6,0.8,1.0,1.2,1.4])
>    File "C:\Home\birk\src\pyprop\b-series\propseries.py", line 269, in 
> plotKTKQ
>      g.additem('set xrange [%f:%f]' % (Jmin,Jmax))
>    File "C:\Home\birk\src\pyprop\b-series\IncrGnuplot.py", line 173, in 
> additem
>      self.__call__(plotitem)
>    File "C:\program 
> files\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 206, in __call__
>      self.gnuplot(s)
>    File "C:\PROGRA~1\python22\lib\site-packages\Gnuplot\gp_win32.py", 
> line 121, in __call__
>      self.write(s + '\n')
> IOError: [Errno 22] Invalid argument
> 
> The weird thing is that it does not happen at the same place in the
> script at different runs from a command terminal. May be this error
> is not a Gnuplot.py issue at all, but related to system or python 
> installation. I hope, however, you can provide me with hints where to look.
> 
> I have python-2.2.1, gnuplot-py-1.6.1, Numeric-21.3 and a lot of
> other stuff which runs as far as I know.
> 
> Best regards
> Lothar

I don't think there's ever been a version 1.6.1 of Gnuplot.py
released.  Are you using 1.6 or a more recent CVS version or ...?

It looks like you are using a home-cooked extension to Gnuplot.py, so
I could only guess what your problem might be.  Possibly the argument
that you are passing to gnuplot.__call__ is not a sensible string as
it should be.

If you send us more information we might be able to help you.

Michael

--
Michael Haggerty
hag...@jp...

Chuck,

[I have cc'ed this email to the gnuplot-py-users mailing list.  Please
subscribe to that list if you are interested in following Gnuplot.py
developments:

    http://lists.sourceforge.net/lists/listinfo/gnuplot-py-users

.]

chuck clark <cn...@us...> writes:
> Michael,
> I am a decent Python programmer but just started
> using Gnuplot and Gnuplot-py yesterday so please
> forgive me if this is an obvious question with an
> obvious answer that I've just missed.
> Is there a way to plot Time/Data data in Gnuplot-py?
> I've got some two dimensional data that looks like
> the following:
> 
> 12:01:30    4.63
> 12:02:30    6.74
> 12:03:30    0.0
> 12:04:30    6.32
> 
> I can get this to plot directly in Gnuplot but I
> can't seem to create a Gnuplot.Data that can
> handle the time type data.  Is there something I'm
> missing?  If plotting this kind of data is not
> possible then is there any chance Gnuplot-py will
> be supporting it in the near future?

I don't have any experience myself with using date-formats in gnuplot,
so I wouldn't know how to support them in Gnuplot.py.  The Gnuplot.py
Data class stores its data internally as a Numeric array of floating
point data.  It has no way of storing (let alone plotting) data in
another format.  If gnuplot requires date data formatted as you
mentioned, how would you want to store it within Gnuplot.py?  As
strings?  Date objects of some sort?

If you have a rather specific purpose in mind, you could easily write
a class derived from PlotItem that prints its contents in the format
that you need.  If you want it to be more generally useful, you base
it on the Data class but instead of using Numeric arrays of float to
store its data it would use Python lists (or perhaps Numeric arrays of
Python objects, though last time I checked they didn't work too well).

If you decide to derive from PlotItem, then I would strongly recommend
that you base your work on the latest CVS version of Gnuplot.py since
the PlotItem interface has changed significantly.  I haven't gotten
around to making a new release but when I do it will include the new
interface.

In fact, I just realized that Numeric isn't really required for very
much in Gnuplot.py.  It might be easy to provide a non-Numeric mode,
which would be interesting to me at the moment because I'm mostly
using Jython in my current work and JNumeric isn't well-developed.

Yours,
Michael

--
Michael Haggerty
hag...@jp...

Hi there,

I was trying to install Gnuplot-py-1.6 on Windows. After reading the readme file it mentions the wgnuplot.exe file.

I was wondering if there exists a downloadable file which i can download.  As sourceforge only has gnuplot-py-1.6.zip, tar.gz etc.

Thank you,

Johnny



---------------------------------
Yahoo! Messenger for SMS- Always be connected to your Messenger Friends

Hello,

I have forwarded this message to
gnu...@li....  Please join this user group
if you are interested in keeping informed about Gnuplot.py.  Info
about subscribing is at

    http://lists.sourceforge.net/mailman/listinfo/gnuplot-py-users

It looks as if the gnuplot process has died, but I personally haven't
seen this error before.  Can you check (with 'ps', for example)
whether gnuplot is still running just before and just after the above
error occurs?  You might want to introduces pauses into demo.py to
facilitate testing.

Michael

Kevin Yiu writes:
> I have encountered an error whilst installing gnuplot-py-1.6.
> I have installed Numeric-18.4.1-Python-2.1 beforehand.
> 
> The errors I am getting when running demo.py are:
> 
> sh: gnuplot: command not found
> Traceback (most recent call last):
>   File "demo.py", line 121, in ?
>     demo()
>   File "demo.py", line 40, in demo
>     g.title('A simple example') # (optional)
>   File "/usr/local/lib/python2.1/site-packages/Gnuplot/_Gnuplot.py", line 
> 439, in title
>     self.set_string('title', s)
>   File "/usr/local/lib/python2.1/site-packages/Gnuplot/_Gnuplot.py", line 
> 381, in set_string
>     self('set %s "%s"' % (option, s))
>   File "/usr/local/lib/python2.1/site-packages/Gnuplot/_Gnuplot.py", line 
> 206, in __call__
>     self.gnuplot(s)
>   File "/usr/local/lib/python2.1/site-packages/Gnuplot/gp_unix.py", line 
> 203, in __call__
>     self.flush()
> IOError: [Errno 32] Broken pipe
> 
> Any suggestions to fix this problem?
> 
> 
> Cheers
> =KeVin=

--
Michael Haggerty
hag...@jp...

Andreas Tille writes:
> On Tue, 10 Sep 2002, Michael Haggerty wrote:
> 
> > What about having gnuplot write the output to a temporary file and
> > then read the file?  If you are under Unix, you could even have
> > gnuplot write the output to a fifo (named pipe) and read it directly
> > into your program.
> I fiddled around something with mkfifo but failed to read it correctly.
> Seems there is some magic to do with threads or someting else because
> reading from the pipe fails completely.

Yes, you would probably have to read from the fifo in another python
thread.

> > I don't think there is a way to get at the string any other way.  One
> > could change Gnuplot.py to read gnuplot's standard output, but I'm not
> > sure that output would only have the pure graphics.  If you want to
> > pursue this, you would probably want to do it by providing your own
> > substitute GnuplotProcess object.
> Any hint how to do that?

Yes, instead of starting gnuplot with the os.popen() command, you
would start it with os.popen2().  This returns the file objects
(child_stdin, child_stdout).  You would write the gnuplot commands to
child_stdin, and you would read the child's results from child_stdout.
Presumably child.stdout would have to be read from a different python
thread to prevent deadlocks.  You would probably have to experiment to
check what other output gnuplot writes to its standard output, and
when and if it flushes the output.  You might have to limit yourself
to plotting a single graph per gnuplot process to prevent subsequent
graphs from running together.

Note that os.popen2() is only available as of Python 2.0.  For
previous Python versions you would have to use the popen2 module.  I'm
not sure whether these modules are available for other OSes.

If you implement this and find it to be useful, submit a patch and
I'll add it to the Gnuplot.py source.

Michael

--
Michael Haggerty
hag...@jp...

On Tue, 10 Sep 2002, Michael Haggerty wrote:

> What about having gnuplot write the output to a temporary file and
> then read the file?  If you are under Unix, you could even have
> gnuplot write the output to a fifo (named pipe) and read it directly
> into your program.
I fiddled around something with mkfifo but failed to read it correctly.
Seems there is some magic to do with threads or someting else because
reading from the pipe fails completely.

> I don't think there is a way to get at the string any other way.  One
> could change Gnuplot.py to read gnuplot's standard output, but I'm not
> sure that output would only have the pure graphics.  If you want to
> pursue this, you would probably want to do it by providing your own
> substitute GnuplotProcess object.
Any hint how to do that?

Kind regards

        Andreas.

thank you for your quick reply! I will certainly check out the mailing list=
=20
archive. my sincere apologies for not doing enough research before asking t=
he=20
question..=20

anyway,again, thanks for the bindings!
greetings,
mark dufour.

On Thursday 12 September 2002 11:28, you wrote:
> mark dufour writes:
> > thanks for writing Gnuplot.py! I have a minor problem though.. I
> > hope you don't mind me mailing you directly..
>
> In the future please subscribe to the Gnuplot.py mailing list and send
> your questions there:
>
>     http://lists.sourceforge.net/lists/listinfo/gnuplot-py-users
>
> > but when this method is invoked from within the program, I get the
> > following error:
> >
> > gnuplot> plot '/tmp/@8667.3' notitle
> >               ^
> >          can't read data file "/tmp/@8667.3"
> >          line 0: (No such file or directory)
>
> This is the well-known but not well-enough documented problem that the
> temporary file used to communicate with gnuplot is deleted before
> gnuplot is done with it.  Please see the mailing list archives for
> suggested solutions.
>
> Michael
>
> --
> Michael Haggerty
> mh...@al...

=2D-=20
Mark Dufour.

'If it stinks, it's chemistry. If it moves, it's biology.
                   If it does not work, It's computer science.'

On Tue, 10 Sep 2002, Michael Haggerty wrote:

> What about having gnuplot write the output to a temporary file and
> then read the file?
Besides the possible speed constraint I'm afraid that this does not
work very well in Zope.  I'm a Zope beginner but I'm afraid that its
Philosophy does not like disk access very much.

> If you are under Unix, you could even have
For sure. It's Debian GNU/Linux.

> gnuplot write the output to a fifo (named pipe) and read it directly
> into your program.
This seems to be the best solution.  But how to do that.  Any example
to let GnuPlot write to a named pipe and read this in a Python program?
Sorry, I'm no experienced Python programmer.

> I don't think there is a way to get at the string any other way.  One
> could change Gnuplot.py to read gnuplot's standard output, but I'm not
> sure that output would only have the pure graphics.  If you want to
> pursue this, you would probably want to do it by providing your own
> substitute GnuplotProcess object.
Well at least I tried `my_program.py > image.eps` and got perfectly the
same as when I set the output to image.eps.  This is no proof but might
be an option for further versions to include (at least for testing ...).

Kind regards

         Andreas.

Andreas Tille writes:
> I'd like to obtain the output from GnuPlot as string rather than as
> file or in a xterm for further processing with PIL or to foreward
> the resulting output to Zope.

What about having gnuplot write the output to a temporary file and
then read the file?  If you are under Unix, you could even have
gnuplot write the output to a fifo (named pipe) and read it directly
into your program.

I don't think there is a way to get at the string any other way.  One
could change Gnuplot.py to read gnuplot's standard output, but I'm not
sure that output would only have the pure graphics.  If you want to
pursue this, you would probably want to do it by providing your own
substitute GnuplotProcess object.

Michael

--
Michael Haggerty
mh...@al...

Hello,

I'd like to obtain the output from GnuPlot as string rather than as
file or in a xterm for further processing with PIL or to foreward
the resulting output to Zope.

If output is not set gnuplot forewards it to stdout and I'm seeking for
a way to grab this into any variable (string, PIL.Image or whatever) to
do further processing with it.

Kind regards

         Andreas.

Dear Gnuplot-py-users,

I found this problem was a know issue and discussed at support page
(http://sourceforge.net/tracker/index.php?func=detail&aid=416091&group_id=17434&atid=217434).

Gnuplot.py works on win32 by following instruction in this support page.

Hiroshi

Hiroshi Watabe
wa...@ri...

Dear Hiroshi,

>   File "c:\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 206, in __call__
>     self.gnuplot(s)
>   File "c:\python22\lib\site-packages\Gnuplot\gp_win32.py", line 124, in __call__
>     self.write(s + '\n')
> IOError: [Errno 22] Invalid argument
> 
> It seems to me that python cannot properly pass argument to
> 'pgnuplot.exe'.

It looks to me that it doesn't even find gnuplot.exe
Is this executable in the executable path ?. It must be. 
Otherwise you can specify the full path to the gnuplot_command
in your system (but you will need to do that for each
release)

I do not run windows, but I changed the gnuplot_command to a 
non existent name in my Linux box and I get an error message
somehow similar to yours. 

Best luck,
-- Leo

Hi Michael !

Thank you for the changes, I tried the CVS version and it works
perfectly in my Linux box (I reverted my dirty fix in my overloaded
gnuplot class). I tried plotting very large (thus slow) surface plots with
no problems. I rerun my codes with no problmes. It seems rock solid.
Great work !

> 1. The PlotItem interface is changed a little bit (the constructor

Ok with me, I personally don't use it directly :-)

> 2. The hierarchy of classes derived from PlotItem is changed
>    significantly.  

Same as above.

> The changes to the majority of the user interface are really quite
> minor; I think only power-users will notice any difference at all.

Yes. Still, you may want to consider naming the next release 2.0,
to make clear that there are structural changes. Also, it
may be a good idea to release a release candidate first and 
let people try it for a while (up to you of course) and give
feedback. This list is fairly new and it has a few users ...

> It would be great if Unix users would grab the new version and test it
> out a little bit, especially within your own applications.  This
> request is especially directed at those users who have been
> "agitating" for such a change!

Say what ???. :-)

Thank you so much Michael

Best,
-- Leo

Hi,

Due to popular demand, I just hacked together some changes to
Gnuplot.py to allow data to be sent to gnuplot via FIFOs (named pipes)
under Unix.  This seems to be a reliable solution for knowing when to
delete temporary files.  For example, test.py runs perfectly even if I
take out all of the delays!  I have checked the changes into CVS.

This new FIFO feature is only implemented under Unix because Python
doesn't support FIFOs under other operating systems.  Other OSs
continue to use temporary files and/or inline data.

However, in the process of rewriting, I had to change the internals of
the PlotItems class hierarchy.  Before I incorporate the changes into
a release I would like to check whether they're going to cause
backwards-compatibility problems with users.  So you, the elite
members of the gnuplot-py-users mailing list, get to exert some
influence!  Let me know if any of the following things would upset
you:

1. The PlotItem interface is changed a little bit (the constructor
   arguments change and a couple new member functions have been
   added).  Has anybody derived new classes from PlotItem?  If so,
   your classes would have to be changed a little bit.

2. The hierarchy of classes derived from PlotItem is changed
   significantly.  For example, File and Data have been changed from
   actual classes into helper functions that return instances of a new
   _FileItem class.  This might be a problem for anybody who works
   with the internals of PlotItems, for example, to derive entirely
   new PlotItem classes.

3. The AnyFile and TempFile classes were removed as they are no longer
   useful within the Gnuplot.py implementation.  Does anybody use
   those classes externally and would be sad to see them go?

The changes to the majority of the user interface are really quite
minor; I think only power-users will notice any difference at all.
For example, everything in test.py ran perfectly after changing
essentially only the example that created a TempFile explicitly.

It would be great if Unix users would grab the new version and test it
out a little bit, especially within your own applications.  This
request is especially directed at those users who have been
"agitating" for such a change!

Yours,
Michael

--
Michael Haggerty
mh...@al...

Dear Gnuplot-py users,

I tried to run Gnuplot-py on Win32 but I got following error messages;

python demo.py
Traceback (most recent call last):
  File "demo.py", line 121, in ?
    demo()
  File "demo.py", line 72, in demo
    g.ylabel('x^2') # take advantage of enhanced postscript mode
  File "c:\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 434, in ylabel
    self.set_string('ylabel', s)
  File "c:\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 381, in set_string
    self('set %s "%s"' % (option, s))
  File "c:\python22\Lib\site-packages\Gnuplot\_Gnuplot.py", line 206, in __call__
    self.gnuplot(s)
  File "c:\python22\lib\site-packages\Gnuplot\gp_win32.py", line 124, in __call__
    self.write(s + '\n')
IOError: [Errno 22] Invalid argument

It seems to me that python cannot properly pass argument to
'pgnuplot.exe'.

I use
 gnuplot-py-1.6
 python-2.2
 gnuplot-3.71

How can I solve this problem? 

Thank you in advance.

Hiroshi

Hiroshi Watabe
wa...@ri...