Cool!
Are you using the "wrap" logic of Mk4py for plugging in other data formats?
Nothing so fancy. You essentially assign a column to a specific renderer. For instance there is a column named zmol in my compound databases. The plugin architecture says that any column named "zmol" gets viewed as a molecule. This allows joins and selects to work appropriately when viewing the new types.
Eventally I'll probably add meta_data information similar to the TCL KitViewer proposals.
I haven't looked at the source, but with mk4py.wrap(), the result is a MK view, on which all of MK's view relational & set operators can be applied. It might even be possible to make Gordon McMillan's MkSQL run on top, which means that you could tie into various formats - show them with your KitViewer, and do all sorts of pretty powerful stuff *across* multiple datafiles.
My next step is actually to layer atop MkSQL and allow sql expressions. I need to understand the internal format for MkSQL a little better though...
I've already had great success sucking in CSV files and using the scripting window to manipulate the data. Once I figured out how to use it, MetaKit is incredibly powerful.
By the way, I have some additions I would like to make to the python documentation that should ease people into using the python wrapper.
It's probably too soon to actually try this, but IMO there is quite some potential in that direction.
-jcw
-- Brian Kelley [EMAIL PROTECTED] Whitehead Institute for Biomedical Research 617 258-6191
_______________________________________________ metakit mailing list - [EMAIL PROTECTED] http://www.equi4.com/mailman/listinfo/metakit
