I'm working on a web app that will need to process DITA documents from persistent storage (likely a JCR). The DITA Open Toolkit is the only DITA implementation I'm aware of, but it requires all of your documents to exist on the filesystem. Ideally, I'd like something that works like the DITA OT, but allows you to provide a resolver (much like an XSLT URIResolver) to pull referenced content from other sources.

If people have other ideas, such as using a virtual filesystem to trick the DITA OT into working, I'd love to hear those too. Thanks!