WIP: Only parse seed files once

[tw4l]

Work in progress

• After successfully parsing all seeds from a seed file, store those seeds in Redis and do not attempt to re-download the seed file again on subsequent runs (e.g. after a crawl is paused, picked up from serialized state, or otherwise restarted)
• Refactor parseSeeds into its own module to avoid circular imports

Tested with pausing crawls in Browsertrix and with picking up from interrupted crawls via serialized state YAML files with the crawler, but still needs tests.