Re: [OctDev] convhulln options.

[Paul K. <pki...@us...>]

On Dec 17, 2003, at 7:00 PM, Rafael Laboissiere wrote:

> * Pedro Tarrafeta <pt...@te...> [2003-12-17 14:52]:
>
>> - Output. The only output the command gives right now is limited and
>>   biased:
>>
>>   - Biased: Since the output matrix is limited to dimension "dim",=20
>> extra
>>     vertices for non simplicial facets get discarded. My knowledge of=20=

>> C is
>>     very little, but.. =BFis it possible to increase the dimension of=20=

>> the
>>     returning matrix (maybe filling it with zeros) in the case of=20
>> finding a
>>     non-simplicial facet?. It would be something in line 97 like =
that:
>>
>>         if (j > dim) {
>>             dim =3D dim + 1
>>             redim idx(n, dim)
>>             for k =3D 1 to i-1 {
>>                 idx (k, dim) =3D  0
>>             }
>>             idx(i,j++) =3D 1 + qh_pointid(vertex->point)
>>          }
>
> Instead of doing that, we might return a cell array instead of a=20
> matrix when
> non simplicial facets are found.  For instance, the following result:
>
>     $ cat cube
>     3
>     7
>     0 0 0
>     0 0 1
>     0 1 0
>     0 1 1
>     1 0 0
>     1 0 1
>     1 1 0
>     $ cat cube | qhull o
>     3
>     7 7 12
>          0      0      0
>          0      0      1
>          0      1      0
>          0      1      1
>          1      0      0
>          1      0      1
>          1      1      0
>     4 2 3 1 0
>     3 6 3 2
>     4 4 6 2 0
>     3 3 5 1
>     3 5 6 4
>     3 6 5 3
>     4 5 4 0 1
>
> would be represented in Octave like this:
>
>     A =3D {[3 4 2 1],
>          [7 4 3],
>          [5 7 3 1],
>          [4 6 2],
>          [6 7 5],
>          [7 6 4],
>          [6 5 1 2]}
>
>
> (Notice that vertex indices in Octave start at 1 instead of zero.)
>
> I do not know how backward incompatible would be such a change.

I think the most convenient would be to return a cell-array of=20
matrices, one
for each dimension of returned facet.

I think it would also be convenient if the default were to arbitrarily=20=

split
those facets that are non-simplicial so that those who just want=20
triangles
don't have to deal with the occasional square.  I'm guessing that this=20=

is
what matlab does --- convhulln returns this for the cube:

 >> convhulln(cube)

ans =3D

      3     4     1
      4     2     1
      2     4     6
      2     5     1
      5     2     6
      7     5     6
      4     7     6
      7     4     3
      7     3     1
      5     7     1


So long as the default  all-simplicial case is returned as a matrix,=20
backward
compatibility doesn't matter.  The user will need special code to=20
handle the
non-simplicial facets anyway.

Are these two cases different enough that we need different functions?

>
>> - Limited: Parsing a string to qhull allows to obtain some other info=20=

>> on
>>   the hull. Some info that is specially interesting for me is the=20
>> normals
>>   to each facet (option n). Since qhull can return various output=20
>> sets is
>>   it possible that it returns an octave struct with the different=20
>> results?.
>>   (I have tried to parse to qhull the "TO file" string so that I=20
>> could get
>>   the normals from the file created but it does not work. This could=20=

>> also
>>   be a solution, in fact THE solution in the case of OFF output to=20
>> feed
>>   Geomview).
>
> To avoid backward incompatible changes, what about having extra output
> arguments, one for the facet normals, other for the OFF string result,=20=

> and
> so on?

I think you are going to need an interface layer above the QHull options
so that users can ask for them nicely:

[X,Y,Z] =3D convhulln(cube,'output',{'simplicial','ideal','geomview'})

This says that X will contain the simplicial output (as a full matrix),
Y will contain a cell array of facet sizes, and Z will contain a string
used to control geomview.  The default for output must be 'simplicial'
for compatibility.

That means you will need other options for controlling the QHull=20
algorithm,
such as 'qhull','QJ' to feed a raw option string, or if you are feeling
creative, meaningful names for all the options.

Paul Kienzle
pki...@us...