Split I/O handling of e-tuff files from ingestion logic

[renato2099]

We should separate i/o handling for e-tuff files into its own specific specialized library. This will facilitate:

• playing around with local e-tuff files (e.g., inspecting metadata, dataset id, tag id, observations, etc).
• users to combine e-tuff files with other sources
• separate of concerns from Tagbase server