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

[fix] normalize place names before adding OSM components, modify components in place, delete keys and use the boundary components if the component is ambiguous

Browse Source
This commit is contained in:
Al
2016-05-05 13:47:44 -04:00
parent cec0d6f6df
commit 1e2894a665
1 changed files with 13 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
scripts/geodata/addresses/components.py
View File

@@ -272,16 +272,16 @@ class AddressExpander(object):
return name return name
def normalize_place_names(self, address_components, osm_components, country=None, languages=None): def normalize_place_names(self, address_components, osm_components, country=None, languages=None):
components = {}
for key in list(address_components): for key in list(address_components):
name = address_components[key] name = address_components[key]
if key in self.BOUNDARY_COMPONENTS: if key in set(self.BOUNDARY_COMPONENTS):
name = self.normalized_place_name(name, key, osm_components, name = self.normalized_place_name(name, key, osm_components,
country=country, languages=languages) country=country, languages=languages)
components[key] = name if name is not None:
return components address_components[key] = name
else:
address_components.pop(key)
def normalize_address_components(self, components): def normalize_address_components(self, components):
address_components = {k: v for k, v in components.iteritems() address_components = {k: v for k, v in components.iteritems()
@@ -699,6 +699,10 @@ class AddressExpander(object):
osm_suffix = self.tag_suffix(language, non_local_language, more_than_one_official_language) osm_suffix = self.tag_suffix(language, non_local_language, more_than_one_official_language)
osm_components = self.osm_reverse_geocoded_components(latitude, longitude) osm_components = self.osm_reverse_geocoded_components(latitude, longitude)
neighborhoods = self.neighborhood_components(latitude, longitude)
all_osm_components = osm_components + neighborhoods
self.normalize_place_names(address_components, all_osm_components, country=country, languages=all_languages)
self.add_admin_boundaries(address_components, osm_components, country, language, self.add_admin_boundaries(address_components, osm_components, country, language,
non_local_language=non_local_language, non_local_language=non_local_language,
@@ -708,7 +712,6 @@ class AddressExpander(object):
if city: if city:
address_components[AddressFormatter.CITY] = city address_components[AddressFormatter.CITY] = city
neighborhoods = self.neighborhood_components(latitude, longitude)
self.add_neighborhoods(address_components, neighborhoods, self.add_neighborhoods(address_components, neighborhoods,
osm_suffix=osm_suffix) osm_suffix=osm_suffix)
@@ -717,9 +720,6 @@ class AddressExpander(object):
all_languages = set([l['lang'] for l in candidate_languages]) all_languages = set([l['lang'] for l in candidate_languages])
all_osm_components = osm_components + neighborhoods
self.normalize_place_names(address_components, all_osm_components, country=country, languages=all_languages)
self.replace_name_affixes(address_components) self.replace_name_affixes(address_components)
self.replace_names(address_components) self.replace_names(address_components)
@@ -770,6 +770,10 @@ class AddressExpander(object):
osm_suffix = self.tag_suffix(language, non_local_language, more_than_one_official_language) osm_suffix = self.tag_suffix(language, non_local_language, more_than_one_official_language)
osm_components = self.osm_reverse_geocoded_components(latitude, longitude) osm_components = self.osm_reverse_geocoded_components(latitude, longitude)
neighborhoods = self.neighborhood_components(latitude, longitude)
all_osm_components = osm_components + neighborhoods
self.normalize_place_names(address_components, all_osm_components, country=country, languages=all_languages)
self.add_admin_boundaries(address_components, osm_components, country, language, self.add_admin_boundaries(address_components, osm_components, country, language,
osm_suffix=osm_suffix, osm_suffix=osm_suffix,