From d29e4f3b2eaeee0c98eae86e1f179030dabdbbe6 Mon Sep 17 00:00:00 2001 From: Al Date: Fri, 26 Aug 2016 23:46:19 -0400 Subject: [PATCH] [openaddresses] Adding optional hyphen between unit number --- scripts/geodata/openaddresses/formatter.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/geodata/openaddresses/formatter.py b/scripts/geodata/openaddresses/formatter.py index 085b75bd..bdb2b734 100644 --- a/scripts/geodata/openaddresses/formatter.py +++ b/scripts/geodata/openaddresses/formatter.py @@ -43,7 +43,7 @@ class OpenAddressesFormatter(object): for (lang, dictionary_type), values in six.iteritems(address_phrase_dictionaries.phrases): if dictionary_type == 'unit_types_numbered': unit_phrases = itertools.chain(*[safe_encode(p) for p in values if len(p) > 2]) - pattern = re.compile(r'\b(?:{})\s+(?:#?\s*)(?:[\d]+|[a-z]|[a-z][\d]+|[\d]+[a-z])\s*$'.format(six.u('|').join(unit_phrases)), + pattern = re.compile(r'\b(?:{})\s+(?:#?\s*)(?:[\d]+|[a-z]|[a-z]\-?[\d]+|[\d]+\-?[a-z])\s*$'.format(six.u('|').join(unit_phrases)), re.I | re.UNICODE) unit_type_regexes[lang] = pattern