pelias / parser

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

improved support for streets with no suffix, such as "broadway" #141

Closed missinglink closed 3 years ago

missinglink commented 3 years ago

this PR resolves the issue discussed in https://github.com/pelias/parser/issues/140

it works by:

no regressions, cc @kochis resolves https://github.com/pelias/parser/issues/140

missinglink commented 3 years ago

The build is was failing because Circle-CI has been removed from this project but it still thinks it's active. I've just removed the webhooks for CircleCI and TravisCI, so it's fixed for next time, this one will require a manual admin override.

missinglink commented 3 years ago

@joxit are "Esplanade/Esplanades" common street types in French? If not we could remove these too:

git grep -i esplanade
...
resources/libpostal/dictionaries/fr/street_types.txt:esplanade|esp
resources/libpostal/dictionaries/fr/street_types.txt:esplanades|esps

another one we might consider removing?

resources/whosonfirst/dictionaries/locality/name:fra_x_preferred.txt:broadway
Joxit commented 3 years ago

Yes, Eslplanade is a common street/place type in French, always in the singular (without s) Some usage examples:

Esplanade de la Liberté
Esplanade du Géneral de Gaulle
Esplanade Méditerranée

Broadway can be safely removed :smile:

missinglink commented 3 years ago

Removed locality/name:fra_x_preferred = Broadway via rebase.

missinglink commented 3 years ago

Removed libpostal/dictionaries/fr/street_types.txt = esplanades|esps via rebase.

missinglink commented 3 years ago

I think this is good to go @Joxit? It comes with a fair few test cases.