Implement inverse indexes

Implement inverse index. An inverse index can only cover a single field. It takes contents of the field, splits the contents using a separator, and adds an inverse index entry for every word in the field.

Example:

tuple: [1, 'Sleepless in Seattle', "Seattle, movie"]
tuple: [2, 'Moscow never sleeps', 'Moscow, song']
tuple: [3, 'Moscow tour guide', 'Moscow, book']

-> you can retrieve all tuples where index[1] field is in 'Moscow' (returns tuple 1 and 3) or 'book' (returns tuple 3)

This would be a simple form of inverse indexes on tag clouds.