[hibernate-dev] Similarity approach in Hibernate Search

[Sanne Grinovero]

Each time a new Document is added to the index,
the similarity relevant to that entity is looked up from the
pertaining documentBuilder
and set to the indexwriter:

AddWorkDelegate:

Similarity similarity = documentBuilder.getSimilarity();
writer.setSimilarity( similarity );
writer.addDocument( work.getDocument(), analyzer );

So the analyzer is scoped per document, the similarity is globally set
on the indexwriter.
Does this make sense to update the similarity for each add operation type?

This is a problem as I can't use two (more) threads to add documents to
the same index.

Is there a good use case for which someone might need a different
Similarity implementation
for different entities contained in the same index?
I'd like to change that to an "illegal configuration".

Sanne