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

[addresses] fix for Colombian house number formatting if the second regex group is not found

Browse Source
This commit is contained in:
Al
2017-02-25 23:24:06 -05:00
parent 64d0783e73
commit f507f2bb3e
1 changed files with 16 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
scripts/geodata/openaddresses/formatter.py
View File

@@ -48,7 +48,7 @@ number_space_letter_regex = re.compile('^[\d]+\s+[a-z]$', re.I)
number_slash_number_regex = re.compile('^(?:[\d]+|[a-z]|[\d]+[a-z]|[a-z][\d]+)[\s]*/[\s]*(?:[\d]+|[a-z]|[\d]+[a-z]|[a-z][\d]+)$', re.I) number_slash_number_regex = re.compile('^(?:[\d]+|[a-z]|[\d]+[a-z]|[a-z][\d]+)[\s]*/[\s]*(?:[\d]+|[a-z]|[\d]+[a-z]|[a-z][\d]+)$', re.I)
number_fraction_regex = re.compile('^(?:[\d]+\s+)?(?:1[\s]*/[\s]*[234]|2[\s]*/[\s]*3)$') number_fraction_regex = re.compile('^(?:[\d]+\s+)?(?:1[\s]*/[\s]*[234]|2[\s]*/[\s]*3)$')
colombian_standard_house_number_regex = re.compile('^(\d+[\s]*[a-z]?)(?: ([a-z]?[\d]+[\s]*[a-z]?))?', re.I) colombian_standard_house_number_regex = re.compile('^(\d+[\s]*[a-z]?)\s+([a-z]?[\d]+[\s]*[a-z]?)?', re.I)
dutch_house_number_regex = re.compile('([\d]+)( [a-z])?( [\d]+)?', re.I) dutch_house_number_regex = re.compile('([\d]+)( [a-z])?( [\d]+)?', re.I)
@@ -202,22 +202,31 @@ class OpenAddressesFormatter(object):
@classmethod @classmethod
def format_colombian_house_number(cls, house_number): def format_colombian_house_number(cls, house_number):
house_number = house_number.strip()
match = colombian_standard_house_number_regex.match(house_number) match = colombian_standard_house_number_regex.match(house_number)
if match: if match:
separator = random.choice((u'-', u' - ', u' ')) separator = random.choice((u'-', u' - ', u' '))
cross_street, building_number = match.groups() cross_street, building_number = match.groups()
if u' ' in cross_street and random.choice((True, False)): numbers = []
if cross_street and u' ' in cross_street and random.choice((True, False)):
cross_street = cross_street.replace(u' ', u'') cross_street = cross_street.replace(u' ', u'')
if u' ' in building_number and random.choice((True, False)): if cross_street:
numbers.append(cross_street)
if building_number and u' ' in building_number and random.choice((True, False)):
building_number = building_number.replace(u' ', u'') building_number = building_number.replace(u' ', u'')
house_number = separator.join([cross_street, building_number]) if building_number:
house_number_prefixes = (u'#', u'no.', u'no', u'nº') numbers.append(building_number)
if random.choice((True, False)) and not any((house_number.lower().startswith(p) for p in house_number_prefixes)):
house_number = u' '.join([random.choice(house_number_prefixes), house_number]) if numbers:
house_number = separator.join(numbers)
house_number_prefixes = (u'#', u'no.', u'no', u'nº')
if random.choice((True, False)) and not any((house_number.lower().startswith(p) for p in house_number_prefixes)):
house_number = u' '.join([random.choice(house_number_prefixes), house_number])
return house_number return house_number