Takes two streams (fullStream and hashStream) and joins them similar to an InnerJoinStream. The difference
in a HashJoinStream is that the tuples in the hashStream will all be read and hashed when this stream is
opened. This provides a few optimizations iff the hashStream has a relatively small number of documents.
You are expected to provide a set of fields for which the hash will be calculated from. If a tuple does
not contain a value (ie, null) for one of the fields the hash is being computed on then that tuple will
not be considered a match to anything. Ie, all fields which are part of the hash must have a non-null value.