gnuplot-py-users Mailing List for Gnuplot.py

Under linux I can not use gnuplot's mouse interaction ((un)set grid, 
(un)zoom, etc.). It looks like gnuplot reread the data file on every 
operation, thus it conflicts with the pipe interface introduced in 1.7.

Any work-around (beside going back to 1.6)?

   Nadav.

Good idea.  I just changed the setting.  If anybody wants even more 
restrictive settings, please let me know.

Michael

Leonardo Milano wrote:

>It might be a good idea to restrict the list so that only
>subscribed members can post. This will eliminate most,
>if not all the spam we are receiving through the list.
>  
>

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

Dear Michael

It might be a good idea to restrict the list so that only
subscribed members can post. This will eliminate most,
if not all the spam we are receiving through the list.

Cheers
Leo

This is easy using just Python features (no special support from 
Gnuplot.py required):

Under any Python version:

    apply(g.plot, plot_list)

Under recent Python versions (starting at 2.0?):

    g.plot(*plot_list)

Either of these takes the elements of the sequence as individual 
arguments to the g.plot() method.

Michael

Nate Gelbard wrote:

>I have a database query that returns a variable length list of hosts for
>which I want to plot some data about.
>How can I pass Gnuplot.plot a list of Gnuplot.Data types?
>In the past I knew how may lines I had on the graph so I could just pass
>g.plot(gd1,gd2,gd3,gd4) .... how do I 
>make a variable length variable list?
>
>I was trying this but got a TypeError: bad argument type for built-in
>operation
>
>       plot_list = []
>        for i in res.keys():
>            print i
>            x = Numeric.arange(len(res[i]),typecode=Numeric.Float)
>            y = Numeric.zeros(len(res[i]))
>            for j in range(0,len(res[i])):
>                line = res[i]
>                y[j] = line[j][0]
>                gd = Gnuplot.Data(x,y,title=i,smooth=smoothtype)
>                plot_list.append(gd)
>            g.plot(plot_list)
>  
>


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

Hi Nate

> I have a database query that returns a variable length list of hosts for
> which I want to plot some data about.
> How can I pass Gnuplot.plot a list of Gnuplot.Data types?

I don't think there is a way. Is that right Michael ?

I did hack Gnuplot to do what you want. Michael, could
you please take a look ? Is it safe to apply to CVS ?
It seems to work for me. Replace the "_add_to_queue'
method in _Gnuplot.py, by the following
hacked version:

#### CODE BEGINS
    def _add_to_queue(self, items):
        """Add a list of items to the itemlist (but don't plot them).

        'items' is a sequence of items, each of which should be a
        'PlotItem' of some kind, a string (interpreted as a function
        string for gnuplot to evaluate), or a Numeric array (or
        something that can be converted to a Numeric array).

        """
        def _add_item(item):
            if isinstance(item, PlotItems.PlotItem):
                self.itemlist.append(item)
            elif type(item) is types.StringType:
                self.itemlist.append(PlotItems.Func(item))
            else:
                # assume data is an array:
                self.itemlist.append(PlotItems.Data(item))

        for item in items:
            if type(item) is types.ListType:
                # assume it is a list of PlotItems
                for plot_item in item:
                _add_item(plot_item)
            else:
                # assume it is a regular PlotItem
                _add_item(item)

#### CODE ENDS

Essentially I moved the original functionality to "_add_item()",
and I chech for each "item" whether it is really a PlotItem, or
a list of PlotItems instead. Then I pass the data to "_add_item()
accordingly.

Anyways, hope this helps,
-- Leo

On Thu, 6 Nov 2003, Nate Gelbard wrote:

> Hi,
> 
> I have a database query that returns a variable length list of hosts for
> which I want to plot some data about.
> How can I pass Gnuplot.plot a list of Gnuplot.Data types?
> In the past I knew how may lines I had on the graph so I could just pass
> g.plot(gd1,gd2,gd3,gd4) .... how do I 
> make a variable length variable list?
> 
> I was trying this but got a TypeError: bad argument type for built-in
> operation
> 
>        plot_list = []
>         for i in res.keys():
>             print i
>             x = Numeric.arange(len(res[i]),typecode=Numeric.Float)
>             y = Numeric.zeros(len(res[i]))
>             for j in range(0,len(res[i])):
>                 line = res[i]
>                 y[j] = line[j][0]
>                 gd = Gnuplot.Data(x,y,title=i,smooth=smoothtype)
>                 plot_list.append(gd)
>             g.plot(plot_list)
> 
> Thanks,
> --( Nate Gelbard, QA Engineer     
> --( Tripwire, Inc., The Integrity Assurance Company

Hi,

I have a database query that returns a variable length list of hosts for
which I want to plot some data about.
How can I pass Gnuplot.plot a list of Gnuplot.Data types?
In the past I knew how may lines I had on the graph so I could just pass
g.plot(gd1,gd2,gd3,gd4) .... how do I=20
make a variable length variable list?

I was trying this but got a TypeError: bad argument type for built-in
operation

       plot_list =3D []
        for i in res.keys():
            print i
            x =3D Numeric.arange(len(res[i]),typecode=3DNumeric.Float)
            y =3D Numeric.zeros(len(res[i]))
            for j in range(0,len(res[i])):
                line =3D res[i]
                y[j] =3D line[j][0]
                gd =3D Gnuplot.Data(x,y,title=3Di,smooth=3Dsmoothtype)
                plot_list.append(gd)
            g.plot(plot_list)

Thanks,
--( Nate Gelbard, QA Engineer    =20
--( Tripwire, Inc., The Integrity Assurance Company

> I can't get started with version 1.7 on Windows XP Professional 


Fixed.  I got Gnuplot-py working on Windows.
The fix is to put an extra set of quotes in the gnuplot_command if there 
is a space in the path:
gnuplot_command = r'"C:\Program Files\gp373w32\pgnuplot.exe"'

My assertion that Gnuplot.funcutils was missing was erroneous.  No excuse.

I also experimented with setting prefer_inline_data to 1, but I found 
that the demo failed to produce surface plots in this case.
-Roger