We've been trying to figure out the best way to parse the Zotero library to find out what we have where, and it's quite a puzzle; the search function gives you results, but doesn't seem to let you find out where those results are in the tree, and the storage system is such that linked files (such as our crucial PDFs of documents) are in a centralized storage system addressed by pointers. Exporting in Zotero RDF or in TEI provides two different views of the data which are useful in different ways, but I'm realizing that if we need to do a lot of this, a custom Zotero exporter would be what we really need.