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

[addresses] Adding ability to determine unit numbers using a known floor number

Browse Source
This commit is contained in:
Al
2016-06-29 15:56:45 -04:00
parent 6fc18b9adb
commit 9efc2d4d79
3 changed files with 30 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
scripts/geodata/addresses/components.py
View File

@@ -811,18 +811,14 @@ class AddressComponents(object):
def generate_sub_building_component(self, component, address_components, language, country=None, **kw):
existing = address_components.get(component, None)
component_class = self.sub_building_component_class_map[component]
if existing is None:
generated_type = self.generated_type(component, address_components, language, country=country)
if generated_type == self.ALPHANUMERIC_PHRASE:
num = component_class.random(language, country=country, **kw)
address_components[component] = num
return num
return True
elif generated_type == self.STANDALONE_PHRASE:
return None
return False
return None
return False
def add_sub_building_phrase(self, component, address_components, language, country, generated_components=None, **kw):
num = address_components.get(component)
@@ -845,16 +841,23 @@ class AddressComponents(object):
def add_sub_building_components(self, address_components, language, country=None, num_floors=None, num_basements=None, zone=None):
generated_components = set()
floor = None
if self.generate_sub_building_component(AddressFormatter.ENTRANCE, address_components, language, country=country):
address_components[AddressFormatter.ENTRANCE] = Entrance.random(language, country=country)
generated_components.add(AddressFormatter.ENTRANCE)
if self.generate_sub_building_component(AddressFormatter.STAIRCASE, address_components, language, country=country):
address_components[AddressFormatter.STAIRCASE] = Staircase.random(language, country=country)
generated_components.add(AddressFormatter.STAIRCASE)
if self.generate_sub_building_component(AddressFormatter.LEVEL, address_components, language, country=country, num_floors=num_floors, num_basements=num_basements):
floor = Floor.random_int(language, country=country, num_floors=num_floors, num_basements=num_basements)
address_components[AddressFormatter.LEVEL] = Floor.random_from_int(floor, language, country=country)
generated_components.add(AddressFormatter.LEVEL)
if self.generate_sub_building_component(AddressFormatter.UNIT, address_components, language, country=country, num_floors=num_floors, num_basements=num_basements):
address_components[AddressFormatter.UNIT] = Unit.random(language, country=country, num_floors=num_floors, num_basements=num_basements, floor=floor)
generated_components.add(AddressFormatter.UNIT)
# Combine fields like unit/house_number here

21
scripts/geodata/addresses/floors.py
View File

@@ -37,16 +37,20 @@ class Floor(NumberedComponent):
return random.randint(1, (num_floors - 1) if num_floors > 1 else 1)
@classmethod
def random(cls, language, country=None, num_floors=None, num_basements=None):
num_type, num_type_props = cls.choose_alphanumeric_type('levels.alphanumeric', language, country=country)
if num_type is None:
return None
def random_int(cls, language, country=None, num_floors=None, num_basements=None):
if num_floors is not None:
number = cls.sample_floors(num_floors, num_basements or 0)
else:
number = weighted_choice(cls.numbered_floors, cls.floor_probs_cdf)
return number
@classmethod
def random_from_int(cls, number, language, country=None):
num_type, num_type_props = cls.choose_alphanumeric_type('levels.alphanumeric', language, country=country)
if num_type is None:
return None
numbering_starts_at = int(address_config.get_property('levels.numbering_starts_at', language, country=country, default=0))
if number >= 0:
@@ -73,6 +77,13 @@ class Floor(NumberedComponent):
elif num_type == cls.NUMERIC_PLUS_ALPHA:
return six.u('{}{}').format(number, letter)
return None
@classmethod
def random(cls, language, country=None, num_floors=None, num_basements=None):
number = cls.random_int(language, country=country, num_floors=num_floors, num_basements=num_basements)
return cls.random_from_int(number, language, country=country)
@classmethod
def phrase(cls, floor, language, country=None, num_floors=None):
if floor is None:

7
scripts/geodata/addresses/units.py
View File

@@ -51,17 +51,18 @@ class Unit(NumberedComponent):
return six.u('{}{}').format(floor_number, safe_decode(unit).zfill(num_digits))
@classmethod
def random(cls, language, country=None, num_floors=None, num_basements=None):
def random(cls, language, country=None, num_floors=None, num_basements=None, floor=None):
num_type, num_type_props = cls.choose_alphanumeric_type('units.alphanumeric', language, country=country)
if num_type is None:
return None
use_floor_prob = address_config.get_property('units.alphanumeric.use_floor_probability', language, country=country, default=0.0)
if num_floors is None or random.random() >= use_floor_prob:
if (num_floors is None and floor is None) or random.random() >= use_floor_prob:
number = weighted_choice(cls.numbered_units, cls.unit_probs_cdf)
else:
floor = Floor.random(language, country=country, num_floors=num_floors, num_basements=num_basements)
if floor is None:
floor = Floor.random_int(language, country=country, num_floors=num_floors, num_basements=num_basements)
floor_numbering_starts_at = address_config.get_property('levels.numbering_starts_at', language, country=country, default=0)
ground_floor_starts_at = address_config.get_property('units.alphanumeric.use_floor_ground_starts_at', language, country=country, default=None)