v0.1.1 — np_unit_id() for matchdb

npnormalize 0.1.1 adds one exported function, np_unit_id(). It is the public suite-id function matchdb needs for step 3 of the extraction plan. The normalizer itself is unchanged.

Install

remotes::install_github("Nonprofit-Open-Data-Collective/npnormalize@v0.1.1")

What's new

• np_unit_id() reduces a suite as np_normalize() writes it in street_unit to its identifier. "STE 300", "# 300" and "STE D-2" give "300", "300" and "D2", the street_unit_id column. A test ties it to that column, so the two can't drift apart.
• Why: matchdb's md_unit_id() used npmatch's internal .np_unit_id(). npmatch now imports this package (npmatch#22), so that internal is gone, and matchdb needs a public function to call. matchdb's step 3 can require npnormalize (>= 0.1.1).
• Like the other four thin wrappers (np_tokenize(), np_doc_freq(), np_addr_key(), np_state_abb()), it adds nothing of its own.
• The site is rebuilt for 0.1.1: https://nonprofit-open-data-collective.github.io/npnormalize/

Normalizer signature

np_normalize_signature() is still v1-3e7b9f9d59bbb19e, the same as npnormalize 0.1.0 and npmatch 0.1.1. Every database, bundle and tranche matchdb has published stays valid.

Checks

• R CMD check passes on GitHub on five platforms (macOS, Windows, and Ubuntu devel, release and oldrel-1).
• 40 tests pass, including the pinned signature.