More flexible Site API

Sort of leading on from #8, I've got a usecase that would need to have more control than provided by the current library implementation. Specifically on the server side, I want to load the index from in memory, and turn file lookups into HTTP requests with `Range` headers. Preferably with the possibility of `async` support.

If I work on this, are you interested in getting it merged upstream?

Start by reviewing the current library implementation and the related discussion in issue #8. Define how server-side in-memory index loading, HTTP Range-based file lookups, and optional async support should work. Done requires an agreed API and behavior for these use cases, but the issue does not name files or tests.

Written by the indexing model from the issue text.