Re: [PyMOL] writing out the surface

[Thomas H. <sp...@us...>]

Hi Michael,

you need to reset the camera view.

def writeSurfaceValue(objname, wrlname='/tmp/tmp.wrl',
         ptsname='/tmp/tmp.pts', pdbname='/tmp/tmp.pdb'):
     cmd.hide('everything')
     cmd.show('surface',objname)
     view = cmd.get_view()
     cmd.set_view([1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0,
         0.0, 0.0, -50.0, 0.0, 0.0, 0.0, 40.0, 100.0, -20.0])
     cmd.save(wrlname)
     points = getIndexedFaceSet(wrlname)
     f = file(ptsname,'w')
     f.writelines(points)
     f.close()
     pts2pdb(ptsname,pdbname)
     cmd.set_view(view)

Cheers,
   Thomas

Michael Lerner wrote, On 07/15/12 04:05:
> I finally got back to this, and I'm still having some trouble. The 
> surface I write out seems to be offset from what I'm viewing in PyMOL. I 
> define the following functions:
> 
> def getIndexedFaceSet(wrlname):
>     f = file(wrlname)
>     for line in f:
>         if 'geometry IndexedFaceSet {' in line:
>             break
>     else:
>         print "Could not find IndexedFaceSet"
>         return None
>     assert 'coord Coordinate' in f.next()
>     assert 'point' in f.next()
>     points = []
>     for line in f:
>         if not line.strip(): continue
>         line = line.strip()
>         if ']' in line:
>             break
>         if line.endswith(','):
>             line = line[:-1]
>         points.append(','.join(line.split()) + '\n')
>     return points
> 
> def pts2pdb(ptsname,pdbname):
>     f = file(pdbname,'w')
>     atomid,resi = 1,1
>     for line in open(ptsname):
>         line = line.strip()
>         if not line: continue
>         c = [float(x) for x in line.split(',')]
>         f.write('ATOM %06s %4s PHO  %04s   
>  %8.3f%8.3f%8.3f\n'%(atomid,resi,'C',c[0],c[1],c[2]))
>         atomid += 1
>     f.close()
>         
> def 
> writeSurfaceValue(objname,wrlname='/tmp/tmp.wrl',ptsname='/tmp/tmp.pts',pdbname='/tmp/tmp.pdb'):
>     cmd.hide('everything')
>     cmd.show('surface',objname)
>     cmd.save(wrlname)
>     points = getIndexedFaceSet(wrlname)
>     f = file(ptsname,'w')
>     f.writelines(points)
>     f.close()
>     pts2pdb(ptsname,pdbname)
> cmd.extend('write_surface_value',writeSurfaceValue)
> 
> and then do
> 
> fetch 1rx1
> set surface_solvent, 1
> show surface
> run <whatever file defines the above functions>
> write_surface_value 1rx1
> load /tmp/tmp.pdb
> show spheres, tmp
> 
> and the tmp object looks like it has the right shape, but is clearly 
> offset from 1rx1.
> 
> I have a feeling I'm doing something simple and obvious wrong, but I'm 
> not sure what.
> 
> Thanks,
> 
> -Michael
> 
> On Mon, Jun 11, 2012 at 10:47 PM, Michael Lerner <mgl...@gm... 
> <mailto:mgl...@gm...>> wrote:
> 
>     Hi all,
> 
>     Tsjerk- Thanks. I'm not sure how I missed that, given that it's even
>     mentioned on the wiki on the surface page under the clear heading
>     "Exporting Surface/Mesh Coordinates to File" :-|.
>     (http://www.pymolwiki.org/index.php/Surface#Exporting_Surface.2FMesh_Coordinates_to_File).
>     I don't know anything about VRML, but I found some documentation
>     online at http://www.c3.hu/cryptogram/vrmltut/part5.html and it
>     looks really straightforward. The surface will show up in the VRML
>     file as an IndexedFaceSet, and the points are just XYZ triples, and
>     I can safely ignore everything else.
> 
>     Darrell- Thanks. I'll build that in as an option.
> 
>     Takanori- Thanks. That's really cool. I happen to want solvent
>     accessible surfaces rather than isosurfaces, but "dump" is
>     definitely going into my bag of tricks.
> 
>     Cheers,
>     -Michael
> 
>     On Mon, Jun 11, 2012 at 8:57 PM, Takanori Nakane
>     <t.n...@ma...
>     <mailto:t.n...@ma...>> wrote:
> 
>         Hi,
> 
>         Another way:
>         There is an undocumented API called 'dump', which can
>         dump coordinates of isomesh/isosurface to a file.
> 
>         For mesh, it dumps a list of vertex coordinates, which
>         can be rendered as GL_LINE_STRIP.
>         For surface, it dumps a list of pairs of a vertex coordinate
>         and its vertex normal vector, which can be rendered as GL_TRIANGLES.
> 
>         Please examine my old post for a script example.
>         http://www.mail-archive.com/pym...@li.../msg10012.html
> 
>         Best regards,
> 
>         Takanori Nakane

-- 
Thomas Holder
MPI for Developmental Biology
Spemannstr. 35
D-72076 Tübingen