RavenDB: Grammatical and phonetical analysis?

Question

I'm a little confused about the level of integration between Lucene.NET and RavenDB. Lucene supports grammatical and phonetical analysis of texts (like word stemming, Metaphone) to allow searches that match the document based on similarity. Now Lucene is used for querying indices, and from what I take from the documentation, also for building them up (see http://ravendb.net/documentation/docs-http-indexes-querying, the last section).

However, the Lucene analyzers that provide grammatical and phonetic analyisis do not seem to be at work. Or are they and I'm missing the switch? Or, is there a way to install/create a plugin to get text analyzers into RavenDB?

Answer 1

At the moment it's not possible to customise the analyser, see this thread for a more in-depth discussion.

If you contribute to the discussion here it might get added. You're not the 1st person who has asked for this behaviour.

One possible solution is to pre-tokenise or analyse the test yourself. So what gets added to the index is the modified text. For instance you could change "run" to "run runs runner" etc. Then a search for "runs" would find your document. There is also a fuzzy search that might help, see here for more info

Update this functionality has now been added to RavenDB, see here for more details