Revise extraction of catalog in Scotland

[sgreenbury]

Currently (#58) the catalog is constructed through extraction of a zip to a cache directory, with census tables subsequently read from the cache_dir.

Instead the zipfile/extracted files and/or the extracted files should be handled into census tables either:

• within a single asset
• or by an IO manager

This enables functionality on a cluster rather than single machine.