[Gusdev-gusdev] Orphaned tables using GBParser?

[Ed R. <ed_...@be...>]

We have two databases up and running in GUS.  One, the Crypto database, used GBParser.  The other, for TCruzi, was data in TIGR XML, and used that plugin.  Superficial analysis of the installs, though, seems to indicate that the data from the GBParser loaded into slightly different tables than the TCruzi data, and this data load (for the GBParser) was incomplete for the tables that should have been filled.

One example: Only half the features in NAFeatureImp map to something in NAFeatureNAProtein.

Another example is NAProtein links back to NAFeatureNAProtein, but is doesn't link back to GeneFeature because there are 0 joins from GeneFeature to NAFeatureNAProtein.

Anyway, we have a number of questions we would like to open up about the consistency of these plugins.  Should they both be using the GUS Schema the same ways?  Does the schema contain different tables to be used only be different plugins?  Do we, in fact, have a model of what tables data sources are supposed to load that was used when writting these plugins?  We all assume that the schema is supposed to be used consistently by all plugins no matter what their data source.  But are there old, abandoned tables that we shouldn't be filling any longer?


-ed




Ed Robinson
255 Deerfield Rd
Bogart, GA 30622
(706)425-9181

--Learn more about the face of your neighbor, and less about your own.
         -Sargent Shriver