Class provides a document-scoped context for generating blank node
identifiers (IDs). This class is intended for use to generate blank node
identifiers (IDs) for a source document within a shared prefix. When that
prefix is mapped onto the term2id index, all blank nodes for the same source
document will tend to be mapped onto the same index partition (this is more
efficient since the ordered writes are more local). The prefix itself is a
UUID, which is how we ensure that blank nodes generated for different
sources are understood as distinct blank nodes by the database (this is
required).
Note: All
BigdataValue instances are in fact created by the delegate
instances created by this class and by the delegate will appear to have been
created by the same factory. This is intentional - it makes the instances
reusable by the base factory.
Note:
#nextID() is NOT thread-safe, but different documents that are
being passed concurrently will have a distinct instance of this factory and
processing a single document is generally single-threaded.