Permit a method of retro scanning without manage.py

[jdheyburn]

Background

Currently the only way to retrospectively scan objects in S3 buckets is via the ./manage.py CLI tool. This blocks a 'set and forget' approach whereby we will have to manually execute this for retro scans. You can build it in a pipeline but that is still work required.

Desired Change

Perhaps move/copy the retro logic within manage.py to a lambda which reads the inventory file to push messages to the analyzer SQS queue. It could even be triggered by when inventory files are created.

Relates #164
Relates #134