Faster (archive level) skipping for dicom-loader

[jas88]

Currently dicom-loader will open and read all archives, skipping the individual DICOM objects within which have already been loaded. It would be faster and generally the correct behaviour (ie unless a previous run aborted mid-archive) to skip all contents of an archive if any item from that archive has already been processed, i.e. skip "123/123456.7z" if any entry of the form "123/123456.7z!%" exists in Mongo.