Repository navigation
What's new
- Addresses near a point:
near=andwithin_km=filters onreal_random_address(),real_random_addresses()andcount().neartakes a(lat, lng)pair or an address from an earlier call. - Groups of nearby addresses:
real_random_cluster(count, within_km=...)returns addresses that lie close together, with no centre to pick. - Nine new states: IL, ID, LA, IN, OH, PA, SC, NM and MN, bringing the dataset to 3,850 addresses across 29 states. Every source was checked against the publisher's own license terms, so every address is still public domain.
- CLI: new
random-address countandrandom-address clustercommands, plus--near LAT,LNGand--within-km KMoptions. - Friendlier inputs: a postal code can be passed as an int (
6040finds06040), and extra spaces in a city name are ignored. - Coverage page:
COVERAGE.mdlists what the dataset holds, state by state. It is generated from the data and checked in CI. - Python 3.15 is supported.
Compatibility
- Everything from 2.1.1 keeps working unchanged.
- Python 3.11 is now the minimum. Python 3.10 reached end of life on 2026-10-01. Nothing breaks for anyone still on it: pip keeps installing 2.1.1 on Python 3.10, and 1.3.0 on 3.9 or older.
Installation
pip install --upgrade random-addressSee the CHANGELOG for the full list of changes.