Feature: Create a DataBase class that examines files within a source

[Zeitsperre]

A DataBase class is an object with the following characteristics:

• Initialized with a local address. Address is globbed to find data files (should data file types, ie file suffixes, be initialized as well?).
• Creates a dictionary of data files; If grouping is set via either a list of indicators or a set of kwargs (e.g. dict(date=True, experiment=True, etc.)), dictionary keys are sorted by groups.
• Contain methods for the following operations: archive or tar, transfer or copy, report_stats, and others
• transfer or copy method should depend on local or connected server statistics. If insufficient space, raise a DiskSpaceError.

[Zeitsperre]

This has been partially achieved:

• a Decoder class now exists that will decode files as needed according to discovered metadata.
• a restructure_datasets function now is able to take data facets and create a filetree structure, with option of either copying or moving files.

There are currently no implementations of these in remote or archive operations. Grouping data for archival purposes would be much easier using the current system. This might be easier if we port the grouping code to class methods of an extended decoder.