On Thursday 06 July 2006 15:57, Nina Jeliazkova wrote: > The method process(IAtomContainer molecule) throws an exception when some > descriptor can not be calculated, which is quite logical.
In the future (not-so-far-future I hope...) it will write a patch that takes advantage of the DataFeatures technology being incorporated into SVN (it's now being used in SaveAs in Bioclipse). This techology allows querying which data features (eg. has_3d_coordinates) is present in some IChemObject. The IChemFileFormat's support this too, so that one can determine of a IChemObject can be saved into a certain file format. See [1]. Similarly, it's my intention to do this for descriptors too, though in this case not hard coded in the source code, but extracted from the QSAR descriptor ontology. See [2]. > However, if an exception is thrown for a descriptor in the middle of the > list (in my example MomentsOfInertia can't be calculated because of missing > 3D coordinates), this prevents from calculation the rest of descriptors, > which may not need 3D coordinates and can otherwise be safely calculated. > > Is this an intentional behaviour? Dunno. > The next question is more about design - is there a reason for classNames, > descriptors, specList and instantiate* functions to be private rather than > protected? Private means these can't be accessed in a descendant class, so > this essentially forbids creation of a descendant class. If you have the need to subclass it, it likely means that the code can be improved, in which I would welcome any patch. Whatfor do you need to subclass it? Egon 1.http://qsar.sourceforge.net/ontologies/data-features/index.xhtml 2.http://qsar.sourceforge.net/dicts/qsar-descriptors/index.xhtml -- CUBIC blog: http://chem-bla-ics.blogspot.com/ Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Cdk-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/cdk-user
