Move terminology matching to Elasticsearch

[phlax]

Currently the terminology matcher is quite complex and very inefficient, particularly if you have a large terminology DB. There is only so much this can be optimised in sql.

Using an indexer such as Elasticsearch which has built in support for more complex matching, and is faster for this type of query should hopefully fix this issue.

In future we may want to look at direct client to Elasticsearch querying or adding the ability to query terminology matches from other sources, but for now it would be good to get this into Elasticsearch.