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

[openaddresses] adding random formatting to Colombian house numbers that match the {calle}-{building number} format

Browse Source
This commit is contained in:
Al
2017-02-18 11:28:14 -08:00
parent 146412f4f8
commit 7cab675809
1 changed files with 22 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
scripts/geodata/openaddresses/formatter.py
View File

@@ -48,6 +48,8 @@ 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)
dutch_house_number_regex = re.compile('([\d]+)( [a-z])?( [\d]+)?', re.I) dutch_house_number_regex = re.compile('([\d]+)( [a-z])?( [\d]+)?', re.I)
SPANISH = 'es' SPANISH = 'es'
@@ -198,6 +200,23 @@ class OpenAddressesFormatter(object):
return house_number return house_number
return cls.chinese_annex_regex.sub(u'\\1号', house_number) return cls.chinese_annex_regex.sub(u'\\1号', house_number)
@classmethod
def format_colombian_house_number(cls, house_number):
match = colombian_standard_house_number_regex.match(house_number)
if match:
separator = random.choice((u'-', u' - ', u' '))
cross_street, building_number = match.groups()
if u' ' in cross_street and random.choice((True, False)):
cross_street = cross_street.replace(u' ', u'')
if u' ' in building_number and random.choice((True, False)):
building_number = building_number.replace(u' ', u'')
return separator.join([cross_street, building_number])
return house_number
def get_property(self, key, *configs): def get_property(self, key, *configs):
for config in configs: for config in configs:
value = config.get(key, None) value = config.get(key, None)
@@ -424,6 +443,9 @@ class OpenAddressesFormatter(object):
if language == CHINESE: if language == CHINESE:
house_number = self.format_chinese_house_number(house_number) house_number = self.format_chinese_house_number(house_number)
if country_dir == Countries.COLOMBIA:
house_number = self.format_colombian_house_number(house_number)
if house_number is not None: if house_number is not None:
components[AddressFormatter.HOUSE_NUMBER] = house_number components[AddressFormatter.HOUSE_NUMBER] = house_number