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

[osm] Add more English + non-local language names for places in OSM

Browse Source
This commit is contained in:
Al
2016-07-29 10:31:26 -04:00
parent ed0b867c13
commit 9dc52ea3c4
1 changed files with 14 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
scripts/geodata/osm/formatter.py
View File

@@ -455,9 +455,11 @@ class OSMAddressFormatter(object):
except (ValueError, TypeError): except (ValueError, TypeError):
population = 0 population = 0
num_references = population / 10000 + 5 # Calculate how many records to produce for this place given its population
if num_references > 1000: population_divisor = 10000 # Add one record for every 10k in population
num_references = 1000 min_references = 5 # Every place gets at least 5 reference to account for variations
max_references = 1000 # Cap the number of references e.g. for India and China country nodes
num_references = min(population / population_divisor + min_references, max_references)
cldr_country_prob = float(nested_get(self.config, ('places', 'cldr_country_probability'), default=0.0)) cldr_country_prob = float(nested_get(self.config, ('places', 'cldr_country_probability'), default=0.0))
@@ -521,14 +523,16 @@ class OSMAddressFormatter(object):
elif six.u(',') in name: elif six.u(',') in name:
name = name.split(six.u(','), 1)[0] name = name.split(six.u(','), 1)[0]
address_components = {component_name: name.strip()} # Add half as many English records as the local language, every other language gets min_referenes / 2
self.components.add_admin_boundaries(address_components, osm_components, country, language, for i in xrange(num_references / 2 if language == ENGLISH else min_references / 2):
random_key=False, address_components = {component_name: name.strip()}
non_local_language=language, self.components.add_admin_boundaries(address_components, osm_components, country, language,
language_suffix=language_suffix, random_key=False,
drop_duplicate_city_names=False) non_local_language=language,
language_suffix=language_suffix,
drop_duplicate_city_names=False)
place_tags.append((address_components, language, False)) place_tags.append((address_components, language, False))
if postal_codes: if postal_codes:
for address_components, language, is_default in place_tags: for address_components, language, is_default in place_tags: