RSS/Atom feed Twitter
Site is read-only, email is disabled

Gimp <-> prolog interface

This discussion is connected to the gimp-developer-list.gnome.org mailing list which is provided by the GIMP developers and not related to gimpusers.com.

This is a read-only list on gimpusers.com so this discussion thread is read-only, too.

4 of 5 messages available
Toggle history

Please log in to manage your subscriptions.

Gimp <-> prolog interface Ryan Eisele 03 Dec 04:17
  200912031353.16672.tneuer@i... Torsten Neuer 03 Dec 13:53
  Gimp <-> prolog interface Joao S. O. Bueno 03 Dec 19:15
   Gimp <-> prolog interface Ryan Eisele 08 Dec 08:51
    Gimp <-> prolog interface Tor Lillqvist 08 Dec 10:05
Ryan Eisele
2009-12-03 04:17:29 UTC (almost 15 years ago)

Gimp <-> prolog interface

I want to write some plug-ins in Prolog.  I don't want to deal with any C pointers, while-loops or if-thens.  Although the code below doesn't make use of it, I want to make use of Prolog backtracking in order to find a correct solution through trial-and-error, such as filling a region with a texture that melds into all edges with no abrupt color differences.  I want to be able to write code such as this, in Prolog, for the blur function for example

blur(In, Out) :- blur(In, Out, In).   %the function needs to access input pixels more than %once (not just when they are at the head of the input list), so pass on another copy of the %input as the 3rd argument blur([], [], _). %when there is no more input, there is no more output, so stop blur( [X-Y-A-R-G-B | Input_Tail], [OX-OY-OA-OR-OG-OB | Output_Tail], Input) :-     %find all the neighbors of current pixel, average them, and unify result with head of     %the output list
    findall( X2-Y2-A2-R2-G2-B2, (
                                                 neighbor(X-Y, X2-Y2),                                                  member(X2-Y2-A2-R2-G2-B2, Input),              Neighbors),
   average_pixels(Neighbors, OX-OY-OA-OR-OG-OB),      blur(Input_Tail, Output_Tail, Input).   %recursion

I want to be able to deal with the input and output in the form of a Prolog list, one list for each layer.  And for multiple layers, a list of lists.  The code above is written for one layer.

Each pixel could come in the form of X-Y-A-R-G-B (integers separated by hyphens) where X and Y are the location coordinates, A is the alpha channel, and R-G-B are their respective RGB values.

I understand that processing the data in Prolog would come with performance drawbacks, but I want to make use of backtracking, and I don't want to be dealing with for-loops and array pointers when I'm defining the output.  If it needs to be optimized somebody could rewrite it into C. 

Is there anyone who has written plug-ins before in the C language that would be able to help me make a Gimp Prolog interface so that Prolog programs can be run from a Gimp plug-in.  I want to have a new plug-in under Filters->Prolog that will open up a list of *.pl files in the plug-ins directory (*.pl meaning Prolog file extension).  I can then choose a *.pl file and the plug-in will then communicate with Prolog.  So in other words, instead of compiling a plug-in for Gimp every time, this single plug-in will be able to run all Prolog programs in their *.pl form as soon as they are written. 

Preferring to work in Prolog I need some help from a C programmer to write this plug-in.  I'd appreciate any help from someone who has written a plug-in before.

Ryan

Joao S. O. Bueno
2009-12-03 19:15:57 UTC (almost 15 years ago)

Gimp <-> prolog interface

Hi Ryan -

I strongly suggest you learn python instead, and us it for your scripts, instead.

It attends eveeryone of yoru requisites: - no dealing with pointers or memory allocation/deallocation - gimp data structures are provided as high level objects - the language syntax even allows for a funciotnal-like programing if you prefer that.

(for example to interact through layers, you just do: image = gimp.list_images()[0]
for layer in image.layers:
#your code dealing with the "layer" object

When you get to the pixels you have then as a byte sequence, you then convert to a bytearray (python 2.6) - to get all the green bytes for the pixels from such an array on a separate array, I could do:

green =[1::4]

Follow the tutorial in www.python.org docs, you could get proficient in the language in less than 1 hour - then check the pythons cripts that come with gimp for examples of the API use and pixel access.

js ->

Ryan Eisele
2009-12-08 08:51:17 UTC (almost 15 years ago)

Gimp <-> prolog interface

js, 

Thanks, I believe Python is the solution I was looking for.

- Ryan

--- On Thu, 12/3/09, Joao S. O. Bueno wrote:

From: Joao S. O. Bueno Subject: Re: [Gimp-developer] Gimp prolog interface To: "Ryan Eisele"
Cc: gimp-developer@lists.xcf.berkeley.edu Date: Thursday, December 3, 2009, 1:15 PM

Hi Ryan -

I strongly suggest you learn python instead, and us it for your scripts, instead.

It attends eveeryone of yoru requisites: - no dealing with pointers or memory allocation/deallocation - gimp data structures are provided as high level objects - the language syntax even allows for a funciotnal-like programing if you prefer that.

(for example to interact through layers, you just do: image = gimp.list_images()[0]
for layer in image.layers:
    #your code dealing with the "layer" object

When you get to the pixels you have then as a byte sequence, you then convert to a bytearray (python 2.6)  - to get all the green bytes for the pixels from such an array on a separate array, I could do:

green =[1::4]

Follow the tutorial in www.python.org docs, you could get proficient in the language in less than 1 hour - then check the pythons cripts that come with gimp for examples of the API use and pixel access.

  js   ->

Tor Lillqvist
2009-12-08 10:05:11 UTC (almost 15 years ago)

Gimp <-> prolog interface

Thanks, I believe Python is the solution I was looking for.

But if you really wanted to use Prolog, surely Python is quite far from that? Or is there some extension or whatever to Python that allows one to write Prolog-like clauses, a Python-hosted Prolog in a way?

(Apparently, yes, there are several, see for instance http://code.activestate.com/recipes/303057/ , http://christophe.delord.free.fr/pylog/index.html and and http://codespeak.net/pypy/extradoc/paper/prolog-in-python.pdf )

--tml