Add the ability to cache the standardizer in the database

[woodbri]

Currently the grammar and lexicon have to be loaded for every address. We need to be able to load it once and cache for the lifetime of the query so it is faster to standardize a table of addresses.

[cvvergara]

what if:
read the whole table of addresses,
then load the grammar and lexicon

[woodbri]

That could work for some tables, but it could be a memory problem if you are trying to standardize 80,000,000 records in a table which is about the size of US and Canada.

I did this before for the C code and it works well to cache these in memory.I just need to work out the logistics of doing this for the C++ structures

[cvvergara]

what if...
you load in chuncks of 100000 rows of table adress
then load grammar & lexicon

then for the 80,000,000 instead of loading 80,000,000 times the grammar & lexicon it would be 800 times