Add the possibility to define the ES document output format

In `ElasticSearchView` would be quite useful to control what JSON-LD algorithm is applied in order to generate the final Document that is going to be indexed.

Currently we index each resource as posted by the client (adding the metadata fields as compacted). However, some clients might want to have a degree of uniformity across their data in order to be able to run queries easily.

One proposal could be:

```json
{
"outputAlgorithm": "compacted|expanded|none",
"outputContext": {...}
}
```
the `outputContext` could be optional, using the context of each of the resources otherwise to perform compaction/expansion

This affects, for example: @jonathanlurie @pgetta

Start by locating ElasticSearchView and tracing how posted resources are transformed before indexing. Clarify the supported outputAlgorithm values and optional outputContext behavior; done means indexed documents can consistently use the requested compacted, expanded, or unchanged format.

Written by the indexing model from the issue text.