gchq
Indexing improvements
Currently indexing is a bit convoluted, having to define the index fields and then define an xslt to create the abstraction matching those fields. The index doc should be self contained so that it knows how to extract each field from the source data, e.g. by each field defining an xpath or similar to extract that field’s value from an event. This would then remove the need for having to define a separate index xslt in the pipeline. It should also remove the need to specify the extraction pipeline in queries. These abstraction are like views on the cooked events, which can either be used against the events or materialised through lucene (+extraction)
Also when defining an index if StreamId/EventId are required things then they ought to be defaulted as is done at the moment with Solr Indexes. The user should not really have to explicitly add SreamId/EventId as it is not something they really care about.
