Normalize content items data

[ledbetterljoshua]

The idea here is that data from different RSS feeds may be returned in different shapes; however, when stored in the database and when accessing the data via the API, the data should be uniform.

One example of this is if the content of the article exists in the "description" key, instead of the "content" key. The lesswrong blog does this. Another is if a description is missing entirely, like in Sam Altmans blog. For this, we may want to create a description from the content (just the first sentence or so), and save that as the description in the DB.

I'm sure there are other examples, but we might want to start collecting them and thinking about them now.