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

[test] Checking probability distributions as part of the address config tests

Browse Source
This commit is contained in:
Al
2016-07-31 22:29:21 -04:00
parent 3ead069b1b
commit 973ac42a97
1 changed files with 16 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
scripts/geodata/tests/test_address_configs.py
View File

@@ -11,6 +11,8 @@ from geodata.addresses.staircases import *
from geodata.addresses.units import * from geodata.addresses.units import *
from geodata.categories.query import * from geodata.categories.query import *
from geodata.math.floats import isclose
invalid_phrase_re = re.compile(r'\b(None|False|True)\b') invalid_phrase_re = re.compile(r'\b(None|False|True)\b')
@@ -19,6 +21,18 @@ class TestAddressConfigs(unittest.TestCase):
def valid_phrase(self, phrase): def valid_phrase(self, phrase):
return phrase is None or not invalid_phrase_re.search(phrase) return phrase is None or not invalid_phrase_re.search(phrase)
def check_components(self, language, country):
conf = address_config.get_property('components', language, country=country)
for component, value in six.iteritems(conf):
if component == 'combinations':
continue
total_prob = 0.0
for k, v in six.iteritems(value):
if k.endswith('probability'):
total_prob += v
self.assertTrue(isclose(total_prob, 1.0), six.u('language: {}, country: {}, component: {}'.format(language, country, component)))
def check_entrance_phrases(self, language, country=None): def check_entrance_phrases(self, language, country=None):
for i in xrange(1000): for i in xrange(1000):
phrase = Entrance.phrase(Entrance.random(language, country=country), language, country=country) phrase = Entrance.phrase(Entrance.random(language, country=country), language, country=country)
@@ -68,6 +82,8 @@ class TestAddressConfigs(unittest.TestCase):
def check_config(self, language, country=None): def check_config(self, language, country=None):
print('Doing lang={}, country={}'.format(language, country)) print('Doing lang={}, country={}'.format(language, country))
print('Checking components')
self.check_components(language, country=country)
print('Checking entrances') print('Checking entrances')
self.check_entrance_phrases(language, country=country) self.check_entrance_phrases(language, country=country)
print('Checking staircases') print('Checking staircases')