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

[openaddresses] adding new config option to OA config for aliasing fields based on a regex

Browse Source
This commit is contained in:
Al
2016-12-18 01:50:58 -05:00
parent d357f0f37c
commit 86a8315b9d
1 changed files with 8 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

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

@@ -246,6 +246,9 @@ class OpenAddressesFormatter(object):
ignore_fields_containing = {field: re.compile(six.u('|').join([six.u('(?:{})').format(safe_decode(v)) for v in value]), re.I | re.UNICODE) ignore_fields_containing = {field: re.compile(six.u('|').join([six.u('(?:{})').format(safe_decode(v)) for v in value]), re.I | re.UNICODE)
for field, value in six.iteritems(dict(self.get_property('ignore_fields_containing', *configs) or {}))} for field, value in six.iteritems(dict(self.get_property('ignore_fields_containing', *configs) or {}))}
alias_fields_containing = {field: [(re.compile(v['pattern']), re.I | re.UNICODE), v) for v in value]
for field, value in six.iteritems(dict(self.get_property('alias_fields_containing', *configs) or {}))}
language = self.get_property('language', *configs) language = self.get_property('language', *configs)
add_components = self.get_property('add', *configs) add_components = self.get_property('add', *configs)
@@ -320,6 +323,11 @@ class OpenAddressesFormatter(object):
if key in ignore_fields_containing and ignore_fields_containing[key].search(value): if key in ignore_fields_containing and ignore_fields_containing[key].search(value):
continue continue
for (pattern, alias) in alias_fields_containing.get(key, []):
if pattern.search(value):
if 'component' in alias:
key = alias['component']
if value: if value:
components[key] = value components[key] = value