tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[openaddresses] abbreviating boundary components for OpenAddresses

Browse Source
This commit is contained in:
Al
2016-09-10 00:04:11 -04:00
parent 604e898d65
commit ae02b0769d
2 changed files with 10 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
resources/parser/data_sets/openaddresses.yaml
View File

@@ -9,6 +9,8 @@ global:
abbreviate_unit_probability: 0.3 abbreviate_unit_probability: 0.3
separate_unit_probability: 0.2 separate_unit_probability: 0.2
abbreviate_toponym_probability: 0.3
place_only_probability: 0.2 place_only_probability: 0.2
place_and_postcode_probability: 0.1 place_and_postcode_probability: 0.1

8
scripts/geodata/openaddresses/formatter.py
View File

@@ -216,6 +216,7 @@ class OpenAddressesFormatter(object):
separate_street_prob = float(self.get_property('separate_street_probability', *configs) or 0.0) separate_street_prob = float(self.get_property('separate_street_probability', *configs) or 0.0)
abbreviate_unit_prob = float(self.get_property('abbreviate_unit_probability', *configs)) abbreviate_unit_prob = float(self.get_property('abbreviate_unit_probability', *configs))
separate_unit_prob = float(self.get_property('separate_unit_probability', *configs) or 0.0) separate_unit_prob = float(self.get_property('separate_unit_probability', *configs) or 0.0)
abbreviate_toponym_prob = float(self.get_property('abbreviate_toponym_probability', *configs))
add_osm_boundaries = bool(self.get_property('add_osm_boundaries', *configs) or False) add_osm_boundaries = bool(self.get_property('add_osm_boundaries', *configs) or False)
add_osm_neighborhoods = bool(self.get_property('add_osm_neighborhoods', *configs) or False) add_osm_neighborhoods = bool(self.get_property('add_osm_neighborhoods', *configs) or False)
@@ -374,6 +375,13 @@ class OpenAddressesFormatter(object):
components.pop(AddressFormatter.UNIT) components.pop(AddressFormatter.UNIT)
unit = None unit = None
for component_key in AddressFormatter.BOUNDARY_COMPONENTS:
component = components.get(component_key, None)
if component is not None:
component = abbreviate(toponym_gazetteer, component, language,
abbreviate_prob=abbreviate_toponym_prob)
components[AddressFormatter.ROAD] = component
# CLDR country name # CLDR country name
country_name = self.cldr_country_name(country, language, configs) country_name = self.cldr_country_name(country, language, configs)
if country_name: if country_name: