[GRASS-dev] GNU scientific library in GRASS

[Jarek Jasiewicz]

Soeren Gebbert pisze:
> Hello,
>
> 2010/4/13 Jarosław Jasiewicz <jarekj at amu.edu.pl>:
>   
>> Hi All!
>>
>> What about using GNU scientific library  (gsl) in GRASS module?
>>
>> Will it create additional dependencies?
>>     
>
> Yes.
>
>   
>> What about compiling on windows?
>> Any other contraindications??
>>
>> In general what I want to use I cannot find in gmath library.
>>     
>
> What kind of algorithms are missing in gmath library?
> GRASS7 already includes a numerical library called ccmath. Have a look
> in the gmath-ccmath wrapper in lib/gmath.
> Currently only a few algorithms from ccmath are used: eigenvalue
> computation, LU decomposition and SVD solver. Have a look at the
> ccmath library (http://freshmeat.net/projects/ccmath/). Maybe the
> algorithms you are searching for is present there? If this is the case
> we can add it to grass7.
>
>   
Hi again!

There is a problem:

I wrote new GRASS module, v.surf.trend I think it will be useful and it 
is a part of something bigger (kriging based GRASS contour to raster 
interpolation with terrain feature like streams lakes, discontinuation, 
peaks etc, something like ANU-DEM) but at the beginning I must solve 
more basic problems

v.surf.trend allow to fit set of points to trend surface at any degree 
of polynomial (in fact more than 6 has no sense)
Module v.surf.trend is ready to use (there remained some cosmetics) but 
is based on external dependencies: GNU Scientific library. It uses 
linear multifit with singular value matrix decomposition 
(gsl_multifit_linear). As I understood above this is not especially 
acceptable.

According to Soren's advice I decided to use function from ccmath and 
here is the problem.

There are two approach which allow to use multifit in ccmath:

function qrlsq which uses qr decomposition to solve linear least squares 
fitting and function lsqsv with singular value matrix decomposition

I assume that results shall be similar to that from gsl. In fact there 
is absolutely not.

I prepared testing version of the module. Is someone can help it is 
here: http://heretic.livenet.pl/heretic/v.surf.trend.tar.gz
(may be some problems with server)

To compile it gsl is necessary

This module performs calculating trend coefs with three methods:

gsl_multifit_linear gives perfect result up to 16 degree polynomial (and 
probably more but I stopped testing here)
ccmath qrlsq gives perfect result to 3th degree of polynomial and good 
result on 4th degree but above coef are completely different from gsl 
and results are similar to good but with artifacts (see output)
ccmath svdlsq gives completely wrong results not comparable with gsl

if someone can help there only calc.c file is important: ccmath.c has 
function from ccmath library, vector.c process input vector file there  
is no help however  and there are some  issues like lack of 3D and WHERE 
support

Sorry for such mail. But it is result of previous suggestion. Now I can:

1) publish module with external dependency and limit it usability for 
advanced users only
2) reinvent wheel again and rewrite gsl solution to GRASS code (I prefer 
spend that time on something more useful, simple extracting like with 
ccmath is not possible)
3) maybe I'm wrong about methods from ccmath and solution is simple?

regards
Jarek

BTW:

there are two flags in current version -r and -s with no flag output is 
from gsl with -r  output is from qr reduction with -s flag svd reduction