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

[osm/formatting] replacing keys with the highest priority so addr:* tags take precedence over is_in:* tags

Browse Source
This commit is contained in:
Al
2015-11-22 22:25:41 -05:00
parent 2695b5dd26
commit f1b6620369
1 changed files with 13 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
scripts/geodata/address_formatting/formatter.py
View File

@@ -6,7 +6,7 @@ import subprocess
import yaml import yaml
from postal.text.tokenize import tokenize, tokenize_raw, token_types from postal.text.tokenize import tokenize, tokenize_raw, token_types
from collections import OrderedDict from collections import OrderedDict, defaultdict
from itertools import ifilter from itertools import ifilter
FORMATTER_GIT_REPO = 'https://github.com/openvenues/address-formatting' FORMATTER_GIT_REPO = 'https://github.com/openvenues/address-formatting'
@@ -155,11 +155,22 @@ class AddressFormatter(object):
for a in c.get('aliases', []): for a in c.get('aliases', []):
self.aliases[a] = name self.aliases[a] = name
def key_priority(self, key):
return self.prioritized_aliases.get(key, len(self.prioritized_aliases))
def replace_aliases(self, components): def replace_aliases(self, components):
replacements = defaultdict(list)
values = {}
for k in components.keys(): for k in components.keys():
new_key = self.aliases.get(k) new_key = self.aliases.get(k)
if new_key and new_key not in components: if new_key and new_key not in components:
components[new_key] = components.pop(k) value = components.pop(k)
values[k] = value
replacements[new_key].append(k)
for key, source_keys in replacements.iteritems():
source_keys.sort(key=self.key_priority)
value = values[source_keys[0]]
components[key] = value
def country_template(self, c): def country_template(self, c):
return self.config.get(c, self.config['default']) return self.config.get(c, self.config['default'])