#!/usr/bin/env python3 # © Copyright 2022, Scott Gasch """Wrapper around US Census address geocoder API described here: https://www2.census.gov/geo/pdfs/maps-data/data/Census_Geocoder_User_Guide.pdf https://geocoding.geo.census.gov/geocoder/Geocoding_Services_API.pdf """ import logging from typing import Any, Dict, List, Optional import requests from requests.utils import requote_uri logger = logging.getLogger(__name__) def geocode_address(address: str) -> Optional[Dict[str, Any]]: """Send a single address to the US Census geocoding API. The response is a parsed JSON chunk of data with N addressMatches in the result section and the details of each match within it. Returns None on error. >>> json = geocode_address('4600 Silver Hill Rd,, 20233') >>> json['result']['addressMatches'][0]['matchedAddress'] '4600 SILVER HILL RD, WASHINGTON, DC, 20233' >>> json['result']['addressMatches'][0]['coordinates'] {'x': -76.92743, 'y': 38.84599} """ url = 'https://geocoding.geo.census.gov/geocoder/geographies/onelineaddress' url += f'?address={address}' url += '&returntype=geographies&layers=all&benchmark=4&vintage=4&format=json' url = requote_uri(url) logger.debug('GET: %s', url) try: r = requests.get(url) except Exception as e: logger.exception(e) return None if r.status_code != 200: logger.error('Unexpected response code %d, wanted 200. Fail.', r.status_code) return None # print(json.dumps(r.json(), indent=4, sort_keys=True)) return r.json() def batch_geocode_addresses(addresses: List[str]): """Send up to addresses for batch geocoding. Each line of the input list should be a single address of the form: STREET ADDRESS, CITY, STATE, ZIP. Components may be omitted but the commas may not be. Result is an array of the same size as the input array with one answer record per line. Returns None on error. >>> batch_geocode_addresses( ... [ ... '4600 Silver Hill Rd, Washington, DC, 20233', ... '935 Pennsylvania Avenue, NW, Washington, DC, 20535-0001', ... '1600 Pennsylvania Avenue NW, Washington, DC, 20500', ... '700 Pennsylvania Avenue NW, Washington, DC, 20408', ... ] ... ) ['"1"," 4600 Silver Hill Rd, Washington, DC, 20233","Match","Exact","4600 SILVER HILL RD, WASHINGTON, DC, 20233","-76.92743,38.84599","76355984","L","24","033","802405","2004"', '"2"," 935 Pennsylvania Avenue, NW, Washington, DC","No_Match"', '"3"," 1600 Pennsylvania Avenue NW, Washington, DC, 20500","Match","Exact","1600 PENNSYLVANIA AVE NW, WASHINGTON, DC, 20500","-77.03534,38.898754","76225813","L","11","001","980000","1034"', '"4"," 700 Pennsylvania Avenue NW, Washington, DC, 20408","Match","Exact","700 PENNSYLVANIA AVE NW, WASHINGTON, DC, 20408","-77.02304,38.89362","76226346","L","11","001","980000","1025"'] """ # TODO: use list_utils.shard to break up the input if it's >10k records # b/c the census website has a hard limit at 10k. url = 'https://geocoding.geo.census.gov/geocoder/geographies/addressbatch' payload = {'benchmark': '4', 'vintage': '4'} raw_file = '' for n, address in enumerate(addresses): raw_file += f'{n+1}, {address}\n' files = {'addressFile': ('input.csv', raw_file)} logger.debug('POST: %s', url) try: r = requests.post(url, files=files, data=payload) except Exception as e: logger.exception(e) return None if r.status_code != 200: print(r.text) logger.error('Unexpected response code %d, wanted 200. Fail.', r.status_code) return None out = [] for line in r.text.split('\n'): line = line.strip() if len(line) > 0: out.append(line) return out if __name__ == '__main__': import doctest doctest.testmod()