calibredb produces all the data in the database in XML format. See the --output-format option. It should be easy enough to process the generated XML and convert it into the catalogs you want in an external tool.
On the other hand you can patch the content server (the file calibre/library/server.py) to generate hierarchical atom feeds for stanza (which I suspect is your use case). If you do decide to patch server.py, send me the patch and I'll consider it for inclusion.