Redirect recipes during indexing when an earlier-known-origin-URL is discovered

[jayaddison]

Describe the reason for these changes and the problem that they solve

As documented in #71, sometimes we discover an earlier known origin (first place that the page was found) for a recipe during recrawling/reindexing of recipes.

Recipes can be recrawled and reindexed in parallel, so there are few timing guarantees available about when this can and will happen.

However, indexing is the last opportunity we have to detect this situation, occurs for all recipes that appear in the search engine, and is a relatively cheap operation that may be re-run without any recrawling of content from the web.

This change adds a detection step that looks for earlier-known origins when a recipe is indexed. If an earlier origin is found, we (attempt to) redirect the recipe document to the expected document's ID, and hide the current document because otherwise it will appear as a duplicate.

Briefly summarize the changes

1. When an earlier page origin is discovered for a recipe during indexing, set the redirected_id value on the Recipe model to the expected ID of the earlier-origin document.

How have the changes been tested?

1. Pending testing.

List any issues that this change relates to
Relates to #71, #82.