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

[osm] Include CLDR country even if the place didn't match simplified OSM polygons

Browse Source
This commit is contained in:
Al
2016-07-28 14:11:31 -04:00
parent ea1226082e
commit bebb33fe64
2 changed files with 7 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

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

@@ -14,7 +14,7 @@ intersections:
places: places:
drop_postcode_probability: 0.6 drop_postcode_probability: 0.6
cldr_country_probability: 0.5 cldr_country_probability: 0.35
streets: streets:
abbreviate_probability: 0.3 abbreviate_probability: 0.3

20
scripts/geodata/osm/formatter.py
View File

@@ -479,11 +479,6 @@ class OSMAddressFormatter(object):
language_suffix=language_suffix, language_suffix=language_suffix,
drop_duplicate_city_names=False) drop_duplicate_city_names=False)
if random.random() < cldr_country_prob and AddressFormatter.COUNTRY in address_components:
address_country = self.components.cldr_country_name(country, None)
if address_country:
address_components[AddressFormatter.COUNTRY] = address_country
place_tags.append((address_components, None, True)) place_tags.append((address_components, None, True))
for language, is_default in local_languages: for language, is_default in local_languages:
@@ -508,10 +503,6 @@ class OSMAddressFormatter(object):
random_key=is_default, random_key=is_default,
language_suffix=language_suffix, language_suffix=language_suffix,
drop_duplicate_city_names=False) drop_duplicate_city_names=False)
if random.random() < cldr_country_prob and AddressFormatter.COUNTRY in address_components:
address_country = self.components.cldr_country_name(country, language)
if address_country:
address_components[AddressFormatter.COUNTRY] = address_country
place_tags.append((address_components, language, is_default)) place_tags.append((address_components, language, is_default))
@@ -537,11 +528,6 @@ class OSMAddressFormatter(object):
language_suffix=language_suffix, language_suffix=language_suffix,
drop_duplicate_city_names=False) drop_duplicate_city_names=False)
if random.random() < cldr_country_prob and AddressFormatter.COUNTRY in address_components:
address_country = self.components.cldr_country_name(country, language)
if address_country:
address_components[AddressFormatter.COUNTRY] = address_country
place_tags.append((address_components, language, False)) place_tags.append((address_components, language, False))
if postal_codes: if postal_codes:
@@ -552,8 +538,14 @@ class OSMAddressFormatter(object):
for address_components, language, is_default in place_tags: for address_components, language, is_default in place_tags:
new_address_components = place_config.dropout_components(address_components, osm_components, country=country) new_address_components = place_config.dropout_components(address_components, osm_components, country=country)
new_address_components[component_name] = address_components[component_name] new_address_components[component_name] = address_components[component_name]
self.components.drop_invalid_components(new_address_components) self.components.drop_invalid_components(new_address_components)
if new_address_components: if new_address_components:
if (AddressFormatter.COUNTRY in new_address_components or place_config.include_component(AddressFormatter.COUNTRY, containing_ids, country=country)) and random.random() < cldr_country_prob:
address_country = self.components.cldr_country_name(country, language)
if address_country:
new_address_components[AddressFormatter.COUNTRY] = address_country
revised_place_tags.append((new_address_components, language, is_default)) revised_place_tags.append((new_address_components, language, is_default))
return revised_place_tags, country return revised_place_tags, country