Wrappers for storing resources (filters, normalisers, etc...) on Elasticsearch

[jnioche]

We store the resource files for the filters and normalisers alongside the code and into the uberjar, which works fine, but Elasticsearch users could store them into a specific index where they could be updated without having to rebuild the jar and restart the topology.

One option would be to have the content of the file in a stored text field but this means that you would not be able to push the file directly onto ES. Another approach is to use JSON documents for the resources, in which case users could send the resource files directly into the index.

[jnioche]

This could be extended to the parsefilter and urlfilter definitions themselves which are in JSON.

[jnioche]

See #577 for introduction of JSONResource interface.