FuzzDex

Fuzzy Index for Python, written in Rust. Works like an error-tolerant dict, keyed by a human input.

Usecases

I use it for matching parts of user supplied physical addresses to data extracted from OSM map to find streets and cities. Original solution used Elasticsearch database with a fuzzy query, which worked - but was 21x slower.

Example

import fuzzdex
cities = fuzzdex.FuzzDex(max_distance=2)
cities.add_phrase("Warsaw", 1, constraints={1, 2, 3})
cities.add_phrase("Wrocław", 2, constraints={4})

streets = fuzzdex.FuzzDex()
streets.add_phrase("Czerniakowska", 1, constraints={1})
streets.add_phrase("Nowy Świat", 2, constraints={1})
streets.add_phrase("Wawelska", 3, constraints={1})

streets.add_phrase("Czerniawska", 4, constraints={2})

cities.finish()
streets.finish()

cities.search("Warszawa", [], max_distance=3, limit=60)
#    [{'origin': 'Warsaw', 'index': 1, 'token': 'warsaw', 
#      'distance': 3, 'score': 100000.0, 'should_score': 0.0}]
 
streets.search("Nowy", ["świat"], max_distance=2, constraint=1)
#    [{'origin': 'Nowy Świat', 'index': 2, 'token': 'nowy', 
#      'distance': 1, 'score': 1.4999924898147583, 
#      'should_score': 7.499962329864502}]

streets.search("Nowy", ["świat"], constraint=2)
#    []