Custom text indexing for known data structures

[mrbrianevans]

In the post processor, there should be an option to set a fixed way of indexing and searching the documents. GH-37 made a generic search which indexes all top-level properties of documents, but this could be improved for known data structures by setting it explicitly.

For example, instead of indexing all the properties of a tweet, it should rather only index the full_text property. This will reduce index size and make results more accurate.

Post processors should also be able to show their own search inputs, rather than the default. This can be a boolean flag in the post processor.