DAWG is not maintained anymore about tz-trout HOT 2 CLOSED

The simplest way to get rid of it would be to use regular dicts for the zipcode-to-TZ-ids mapping. I think that's what we did in the past and then added the DAWG as way to reduce the memory footprint. However, there should be a much simpler way around it.

There are ~42k zipcodes in the US and each of them points to a list of TZ ids. That obviously adds up if you're using a unique list per zip code. However, most of these values are NOT unique and hence we can probably just make sure that we're reusing the same instance of a TZ ID list for each neighboring ZIP code.

Perhaps we can utilize tztrout.data.deduplicator?

from tz-trout.

Got rid of it in #16

from tz-trout.