[Chimera-users] Extra Chimera marker attributes

Tom Goddard goddard at cgl.ucsf.edu
Tue Apr 10 12:47:47 PDT 2007

Hi Jeff,

  The Chimera path tracer markers accommodate extra attributes that are 
read and written to the XML marker files.  There is no user interface to 
define new marker attributes in Chimera.  A user interface to define new 
marker attributes does not seem hard.  You could specify an attribute 
name and value in a command or dialog that assigns that attribute for 
all selected markers.  For these new attributes to be of any use within 
Chimera more capabilities would be needed.  For example some command or 
dialog that would allow you to select all markers with a certain 
attribute and value (e.g. is_filament = true).

  These kinds of capabilities already exist for attributes of atoms of 
molecules.  For example, see the "setattr" and "defattr" commands, and 
"Render/Select by Attribute" tool, and atom specifiers that can test for 
specific attribute values.


These will work with path tracer markers since markers are implemented 
as atoms.  But there are problems.  For example "defattr" uses names of 
atoms and the names of the path tracer atoms are not meaningful.  
Probably "setattr" is more useful and you can set an attribute for all 
currently selected atoms.  The main weakness is that these atom 
attributes will not be saved in marker XML files.  That is probably easy 
to fix but I would need to investigate how atom attributes are 
implemented.  For markers the attributes are kept in a Python dictionary 
marker.extra_attributes.  I do not know if atom attributes on markers 
will be saved in Chimera session files.  Would have to test that.


Jeff wrote:
> Tom,
> how hard would it be to add other attribute(s) to the markers 
> generated with the volume path tracer tool, aside from 
> marker-to-marker connectivity? Such as the ability to enumerate 
> individual filaments/objects within a single marker set, or a flag 
> that would tell whether a set/group is composed of filaments, or 
> surfaces, or just a point cloud? Either of of these (enumeration, or 
> object type) could be stored as an integer.
> -Jeff

More information about the Chimera-users mailing list