Refactored metadata pipeline, added archive uploading

[cpritcha]

• Merged the file group and metadata extraction phase in the metadata pipeline
• Brought all shared data structures into __init__.py
• Added support for archive uploads and celery status checking

Refactored CSV loader to only open file once

TODO: better error messages for celery task failure