gnuplot-py-users Mailing List for Gnuplot.py

Hi Michael,

as you told the "using" argument is essential for the 4d plot. With

>
> >>> d = Gnuplot.Data(asd, inline=1, using=(1,2,3,4,))
> >>> g.splot(d)
>

 my plot is now behaving in the way I want.

Many thanks.

Marc

Marc Hodapp wrote:

> I get a colored 3d plot with x,y,z and z as colored values on the 
> surface. Somehow the val- values are not shown on the colored surface. 
> Is this due to the implementation of Data? Or did I misunderstand 
> something?
> [...]
> g.title('test')
> g('set pm3d')
> g('set term X11')
> g.splot(Gnuplot.Data(asd, inline=1))
> raw_input()
>
> All I got on the colored surface where values ranging from 0 to 2  
> instead of values from  10 to 30 .


I guess you also need the "using" argument to tell gnuplot to use all 4 
columns:

 >>> Gnuplot.Data(asd, inline=1, using=(1,2,3,4,))

General advice for debugging problems:

To see exactly what is happening, initialize the Gnuplot object with 
debug=1; i.e.,

 >>> g = Gnuplot.Gnuplot(debug=1)

This causes the Gnuplot object to print out each of the commands that it 
is sending to gnuplot.  Also create the Data object explicitly then ask 
it to print out its content:

 >>> d = Gnuplot.Data(asd, inline=1, using=(1,2,3,4,))
 >>> print d.content
 >>> g.splot(d)

The print statement outputs the same data that will be sent to gnuplot 
via the "inline" mechanism.

Now you know everything that is being sent to gnuplot.  If what you see 
differs from what you need to have sent to gnuplot, let us know the 
difference and we'll try to get you from here to there :-)

Michael

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

Hi all,

Michael's suggestion almost worked like a charm (thanks).  This  means  
when I format  my data  like:

data = [
[
[1,1,z(1,1),val(1,1)],
[1,2,z(1,2),val(1,2)],
[1,3,z(1,3),val(1,3)],
],
[
[2,1,z(2,1),val(2,1)],
[2,2,z(2,2),val(2,2)],
[2,3,z(2,3),val(2,3)],
],
...
]

and then plot it with:

resG.splot(Gnuplot.Data(data))

I get a colored 3d plot with x,y,z and z as colored values on the 
surface. Somehow the val- values are not shown on the colored surface. 
Is this due to the implementation of Data? Or did I misunderstand something?
See below for I test case I tried to run.

import Gnuplot
g = Gnuplot.Gnuplot()

asd =[ \
    [ \
     [1,1,0,10], \
     [1,2,0,10], \
     [1,3,0,10], \
     [1,4,0,10], \
    ], \
    [ \
     [2,1,1,20], \
     [2,2,1,20], \
     [2,3,1,20], \
     [2,4,1,20], \
    ], \
    [ \
     [3,1,2,30], \
     [3,2,2,30], \
     [3,3,2,30], \
     [3,4,2,30], \
    ], \
  ]

print asd

g.title('test')
g('set pm3d')
g('set term X11')
g.splot(Gnuplot.Data(asd, inline=1))
raw_input()


All I got on the colored surface where values ranging from 0 to 2  
instead of values from  10 to 30 .

Every help is appreciated. Many thanks in advance

Marc Hodapp

To get blank lines in the data that is sent to gnuplot, put your data 
into a 3-d array then pass the array to Gnuplot.Data.  In other words, 
you currently have

x = [1,1,1,2,2,2,3,3,3,4,4,4]
y = [1,2,3,1,2,3,1,2,3,1,2,3]
z = [z(1,1),z(1,2),z(1,3),z(2,1),...]
val = [val(1,1), ...]

Instead, create an array as follows:

data = [
    [
        [1,1,z(1,1),val(1,1)],
        [1,2,z(1,2),val(1,2)],
        [1,3,z(1,3),val(1,3)],
    ],
    [
        [2,1,z(2,1),val(2,1)],
        [2,2,z(2,2),val(2,2)],
        [2,3,z(2,3),val(2,3)],
    ],
    ...
    ]

Then plot the data with

    resG.splot(Gnuplot.Data(data))

You could create the data array from your old lists using something like 
(untested!):

    data = Numeric.asarray([x,y,z,val], Numeric.Float32)
    # Change the order of the indices to "zip" the arrays together:
    data = Numeric.transpose(set, (1,0))
    data = Numeric.reshape((4,3,4,), data)

If you generate the data using functions (and if you are using python 2 
or above), an even easier way would be to use

    data = [[[x,y,z(x,y),val(x,y)] for y in yvals] for x in xvals]

Hope that helps.
Michael

Marc Hodapp wrote:

> when using gnuplot directly,  I can create 4d data plots with 
> something like:
>
>    set pm3d
>    splot 'file.dat' using 1:2:3:4
>
> Where file.dat is the data file, 1,2,3 are the columns with the x,y,z 
> data and 4 is the column with the value data which defines the color 
> value.
>
> Now I want to do the same thing using gnuplot.py. In my python 
> program, I define 4 lists: x, y,z,val.  I then use the instructions:
>
>    resG = Gnuplot.Gnuplot()
>    resG('set pm3d')
>    resG.splot(Gnuplot.Data(x,y,z,val))
>
> What I get as a result, is a 3d plot with my points, but no colored 
> surface showing val . So my questions are: Can Gnuplot.Data() receive 
> 4 arguments, which are then passed to splot and pm3d? When using 
> gnuplot directly, the file must contain blank lines between two scans 
> (isolines). In gnuplot.py do I have to include empty strings in the  
> x,y,z,val lists to delimit the isolines? How can I delimit these scans 
> in gnuplot.py? Is there an other way to plot my surface using the 
> xyz,val lists ( like resG('splot ...') )?


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

Hi all,

when using gnuplot directly,  I can create 4d data plots with something 
like:

    set pm3d
    splot 'file.dat' using 1:2:3:4

Where file.dat is the data file, 1,2,3 are the columns with the x,y,z 
data and 4 is the column with the value data which defines the color value.

Now I want to do the same thing using gnuplot.py. In my python program, 
I define 4 lists: x, y,z,val.  I then use the instructions:

    resG = Gnuplot.Gnuplot()
    resG('set pm3d')
    resG.splot(Gnuplot.Data(x,y,z,val))

What I get as a result, is a 3d plot with my points, but no colored 
surface showing val . So my questions are: Can Gnuplot.Data() receive 4 
arguments, which are then passed to splot and pm3d? When using gnuplot 
directly, the file must contain blank lines between two scans 
(isolines). In gnuplot.py do I have to include empty strings in the  
x,y,z,val lists to delimit the isolines? How can I delimit these scans 
in gnuplot.py? Is there an other way to plot my surface using the 
xyz,val lists ( like resG('splot ...') )?

Many thanks in advance.

Marc Hodapp