npnormalize: nonprofit names and addresses as matchable keys

The normalizer shared by npmatch, matchdb and the packages built on them. np_normalize() turns an organization's name and address into keys (name_key, name_core, street_key, the ZIP parts and more), and np_name_freq(), np_token_idf() and np_addr_freq() count statistics over those keys.

Details

A record and the reference it is matched against have to be normalized by the same code, or matches quietly stop being found. np_normalize_signature() is a hash of that code. matchdb stamps it into everything it publishes, and refuses to use a stored artifact under a different normalizer.

This code was moved out of npmatch unchanged (dev/EXTRACTION.md), so the signature it gives is the one npmatch gave.

See also

Useful links:

• https://github.com/Nonprofit-Open-Data-Collective/npnormalize

• https://nonprofit-open-data-collective.github.io/npnormalize/

• Report bugs at https://github.com/Nonprofit-Open-Data-Collective/npnormalize/issues

Author

Maintainer: Jesse Lecy jdlecy@gmail.com