create a model for removing old sites

[gbinal]

Idea:

• For a recent scan date, take the all snapshot, filter for the following scan statuses to all be dns resolution errors: primary, robots_txt, sitemap_xml, and www.
• Take the resulting list of initial domains.
• Repeat this for one scan date per month for the past 12 months.
• Append all of the initial domains
• Count how many times each initial domain appears. If the count = 12, it makes the end product file.
• Put the list here.
• As the second to last step in the index building process (right before the forced-add-in step), remove these initial domains.

I'm open to another model though...