From: Egon W. <e.w...@sc...> - 2007-01-29 12:12:45
|
On Monday 29 January 2007, Nina Jeliazkova wrote: > Some solutions to be discussed: > - introduce something like "translator facility" to perform translation > property_name_in_file -> property_name_to_use_in_this_software > > - separate properties into different Hashmaps - like identifiers (cas, > names), descriptors, measured properties, etc. > > Does something like this already exist in CDK? CML has a good approach in this: use dictionaries. These dictionaries describe what a certain 'label' means. For example, it could define "CAS RC" to be equivalent to "CAS" etc. OWL is very much suited for this. Remains, however, that these dictionaries need to be written, which really is the responsibility of the program (author) who write the file that uses that dictionary... OWL also allows classification of entries, e.g. done in the blue obelisk algorithm dictionary (see the BO article in JCIM). Based on such classification, we could set up a tool to split up properties according to such a classification. Now, MDL molfile is not really the suitable format for this, and suggest the use of CML instead. Nevertheless, we could do something similar for MDL molfiles too, though this would be an undocumented convention. CML at least allows specificing that. Egon -- e.w...@sc... Cologne University Bioinformatics Center (CUBIC) Blog: http://chem-bla-ics.blogspot.com/ GPG: 1024D/D6336BA6 |