Store blogpost preview text into storage during crawling

[velias]

Store blogpost preview text into storage during crawling - analysis of current usage shows it is useful for clients and can reduce network traffic if API is designed correctly.
It should be plain content text content extracted from blogpost HTML body and cropped to defined maximal length (with ... appended if cropped), the same way as in Borg.
Should be stored into field named content_preview which is documented in mongo backend project now (but not implemented yet!)