The BDD tests cannot make assumptions about the structure of the
word table anymore because it depends on the tokenizer. Use more
abstract descriptions instead that ask for specific kinds of
tokens.
It may happen that two different postcodes normalize to exactly
the same token. In that case we still need two different entries
in the word table. Token lookup will then make sure that the correct
one is choosen.
Fixes#1953.
Many of the postcode nodes are actually derived from
incomplete addresses and are as such not even centroids.
Better let them only take part in the address computation
via the postcode table.
