Re: [Gnuplot-info] Extracting (approximate) data from a PostScript file containing gnuplot-generated plot

[Thomas S. <t.s...@fz...>]

Andrew DeYoung <adeyoung <at> andrew.cmu.edu> writes:

> 
> Hi,
> 
> Suppose that I have a PostScript file containing a plot which was generated
> using gnuplot.  However, I do not have the source data, nor do I have the
> gnuplot commands that were used to generate the plot.
> 
> Do you know of any way to somehow extract data from a graphic representation
> (i.e., a PostScript file)?  Such code would have to literally "read off of
> the graph" (in particular, I have a smoothed line/scatter XY plot) from the
> pixel representation, and I know that the results would be approximate at
> best (but this would still be very highly desirable).  
> 
> Do you have any experience with this?  Thank you for your time!
> 
> Andrew DeYoung
> Carnegie Mellon University
> 
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and 
> threat landscape has changed and how IT managers can respond. Discussions 
> will include endpoint security, mobile security and the latest in malware 
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> gnuplot-info mailing list
> gnuplot-info <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

because postscript is a programming language, the
postscript file contains all commands to generate
the plot.
you need to determine the box surrounding the plot
(normally the axes) in postscript units (near the end
of the postscript file), and then with the knowledge
of the original axis ranges (which you see when
printing/viewing the postscript file) you may transform
the plotted points or lines back into the original
coordinate system.

with some basic knowledge of the postscript language
it's relatively easy to extract the data.

you may make a copy of the postscript file, open it
with a normal text editor, and try to identify the
components.
sometimes it helps to change the color by inserting
a line like
   1.0 0.0 0.0 setrgbcolor
into the file and see which lines turn to red when
viewing the file with ghostscript.

an example:

somewhere in the prologe section some commands are
redefined (to save filespace):

...
/M {moveto} bind def
/L {lineto} bind def
/R {rmoveto} bind def
/V {rlineto} bind def
/N {newpath moveto} bind def
/Z {closepath} bind def
/C {setrgbcolor} bind def
/f {rlineto fill} bind def
/g {setgray} bind def
...

later in the file these commands are used
(the following is near the end of the file):

...
% End plot #1
stroke
LTb
546 4871 N
546 280 L
6401 0 V
0 4591 V
-6401 0 V
Z stroke
1.000 UP
1.000 UL
LTb
stroke
grestore
end
showpage
...

the above draws the box around the plot

moveto 546 4871
lineto 546 280
relativelineto 6401 0
relativelineto 0 4591
relativelineto -6401 0

that means the xaxis ranges from 546 to 6947
and the yaxis from 280 to 4871.
i plotted a sin function using default ranges
(i.e. xrange [-10:10], yrange [-1:1]), so for
the xaxis
546 corresponds to -10.
6947 corresponds to 10.
and for the yaxis
280 corresponds to -1.
4871 corresponds to 1.

the datapoints or function line segments are
plotted the same way like the above box:

...
546 3824 M
65 -412 V
64 -445 V
65 -462 V
65 -459 V
64 -438 V
...

which is (in original coordinates):

moveto -10.,0.5438
rlineto 0.0203, -0.1795
rlineto 0.0200, -0.1939
a.s.o.