Hi all I spent most of the last night and todays morning with implementing a Clerezza in-memory TripleCollection that supports SPO, POS, OSP indexes for fast filtered iterators.
See https://issues.apache.org/jira/browse/CLEREZZA-683 for details I plan to change the Enhancer and the Entityhub to use this implementation over the weekend. This should give us good performance improvements especially for larger documents. I also expect the WebUI of the enhancer to be much more responsive. However I assume that performance improvements for Ontonet, Reasoning and Rules could be even more significant if this components where actually using Clerezza in-memory models. If yes it would be great if you could check such use cases and provide some performance comparisons. Finally the IndexedMGraph model should also provide good JSON-LD serialization performance for big graphs as this serializer heavily depends on filtered iterators! best Rupert -- | Rupert Westenthaler [email protected] | Bodenlehenstraße 11 ++43-699-11108907 | A-5500 Bischofshofen
