def name_fuzzy_partial_score(product1, product2):
product1 = _filter_name(product1)
product2 = _filter_name(product2)
s1 = utils.full_process(product1)
s2 = utils.full_process(product2)
return partial_ratio(s1, s2)
def WRatio(s1, s2):
p1 = utils.full_process(s1)
p2 = utils.full_process(s2)
if not utils.validate_string(p1): return 0
if not utils.validate_string(p2): return 0
# should we look at partials?
try_partial = True
unbase_scale = .95
partial_scale = .90
base = ratio(p1, p2)
len_ratio = float(max(len(p1), len(p2))) / min(len(p1), len(p2))
# if strings are similar length, don't use partials
if len_ratio < 1.5: try_partial = False
# if one string is much much shorter than the other
if len_ratio > 8: partial_scale = .6
if try_partial:
partial = partial_ratio(p1, p2) * partial_scale
ptsor = partial_token_sort_ratio(p1, p2) * unbase_scale * partial_scale
ptser = partial_token_set_ratio(p1, p2) * unbase_scale * partial_scale
return int(max(base, partial, ptsor, ptser))
else:
tsor = token_sort_ratio(p1, p2) * unbase_scale
tser = token_set_ratio(p1, p2) * unbase_scale
return int(max(base, tsor, tser))
def QRatio(s1, s2):
if not utils.validate_string(s1): return 0
if not utils.validate_string(s2): return 0
p1 = utils.full_process(s1)
p2 = utils.full_process(s2)
return ratio(p1, p2)
def brand_fuzzy_match(product_name, brand):
from product_spiders.fuzzywuzzy.fuzz import ratio, partial_ratio
from product_spiders.fuzzywuzzy import utils
s1 = utils.full_process(product_name)
s2 = utils.full_process(brand)
if ratio(s1, s2) > 80:
return True
if partial_ratio(s1, s2) > 80:
return True
return False
def testCaseInsensitive(self):
self.assertNotEqual(ratio(self.s1, self.s2),100)
self.assertEqual(ratio(utils.full_process(self.s1), utils.full_process(self.s2)),100)
def test_fullProcess(self):
for s in self.mixed_strings:
utils.full_process(s)