pelias / parser

natural language classification engine for geocoding
https://parser.demo.geocode.earth
MIT License
55 stars 27 forks source link

Addresses on Bowery in NYC are not properly handled #177

Open Andrew-Dickinson opened 6 months ago

Andrew-Dickinson commented 6 months ago

Describe the bug

Addresses such as

186 Bowery, New York, NY 10012

are not properly parsed

Steps to Reproduce

Spin up interactive webserver. Type 186 Bowery, New York, NY 10012 in

Expected behavior

Housenumber: 186 Street: Bowery City: New York State: NY Zip: 10012

Environment (please complete the following information):

N/A

Pastebin/Screenshots

Screenshot 2024-02-13 at 21 19 38

Additional context

References

https://en.wikipedia.org/wiki/Bowery

missinglink commented 6 months ago

Thanks for the bug report, most streets in the USA contain a suffix such as st/rd but of course there are notable exceptions such as Broadway and this example.

Ideally we could infer that the token between the house number and the city was the street, particularly when a comma is supplied.