def characterize(bd, # this document
bj, # in this document collection
boost=param_dict(default=batch.BOOST),
k1=batch.K1,
b=param_dict(default=batch.B)):
words = set()
for bow in bd.values():
for w in bow.keys():
words.add(w)
pr = prf_result()
for w in words:
pr[w] = bm25f({w: 1}, bd, bj, boost, k1, b)
return pr
def test_b(self):
b = param_dict(default=0.75)
b['title'] = 0.50
b['body'] = 1.00
self.assertEqual((1 * 1.0) / ((1 - 0.50) + 0.50 * 2 / (4 / 4)) +
(1 * 1.0) / ((1 - 1.00) + 1.00 * 1 / (3 / 4)) +
(0 * 1.0) / ((1 - 0.75) + 0.75 * 1 / (1 / 4)),
weight('テスト', self.bd0, self.bj, b=b))
def test_boost(self):
boost = param_dict(default=1.0)
boost['title'] = 100
boost['body'] = 0.1
self.assertEqual((1 * 100) / ((1 - 0.75) + 0.75 * 2 / (4 / 4)) +
(1 * 0.1) / ((1 - 0.75) + 0.75 * 1 / (3 / 4)) +
(0 * 1.0) / ((1 - 0.75) + 0.75 * 1 / (1 / 4)),
weight('テスト', self.bd0, self.bj, boost=boost))
def test_param_dict_omit_default(self):
d = {'title': 10}
pd = param_dict(d=d)
self.assertEqual(10, pd['title'])
self.assertEqual(None, pd['body'])
def test_param_dict_omit_d(self):
pd = param_dict(default=1)
self.assertEqual(1, pd['title'])
self.assertEqual(1, pd['body'])
def test_param_dict(self):
d = {'title': 10}
pd = param_dict(d=d, default=1)
self.assertEqual(10, pd['title'])
self.assertEqual(1, pd['body'])