On Dec 16, 2008, at 11:56 AM, Dina Schneidman wrote:

no need to add, works just fine.
Constructing it directly is slightly unsafe as a few method assume that the keys are initialized in a given order (the problem is unlikely to arise, so go ahead until the patch is committed).

I stuck a patch in kernel/doc/patch to added methods to safetly construct keys from PDB strings (it requires a minor change to Key.h). The new methods check to make sure the atom and residue names are standard ones.


thanks!

On Tue, Dec 16, 2008 at 11:48 AM, Daniel Russel <drussel@gmail.com> wrote:

On Dec 16, 2008, at 11:40 AM, Dina Schneidman wrote:

Hi,

is there any reason that ResidueTypes and AtomTypes are defined as static variables, rather than enum?
Having them as static variables means you have a proper type associated with the identifiers, which is kind of nice. It also means that you can add your own residue names and have the mapping to a string accessible to i/o code.




what is the best way to map between ResidueType and the PDB residue name string and the same for atoms?
ResidueType("pdbname") should work for now. Same for AtomType.

I think I should add a function residue_type_from_pdb_name and atom_type_from_pdb_name to provide a cleaner and clearer route. I'll check in such code shortly.


thanks!
Dina

_______________________________________________
IMP-dev mailing list
IMP-dev@salilab.org
https://salilab.org/mailman/listinfo/imp-dev

_______________________________________________
IMP-dev mailing list
IMP-dev@salilab.org
https://salilab.org/mailman/listinfo/imp-dev

_______________________________________________
IMP-dev mailing list
IMP-dev@salilab.org
https://salilab.org/mailman/listinfo/imp-dev