def test_gets(self):
a = BitIndex()
pos = [random.randint(0, 1000) for i in xrange(100)]
a.sets(pos)
self.assertEqual(a.size, max(pos) + 1)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertTrue(a.get(pos[0]))
self.assertTrue(all(a.gets(pos)))
self.assertEqual(list(a.positions()), sorted(set(pos)))
def test_gets(self):
a = BitIndex()
pos = [random.randint(0, 1000) for i in xrange(100)]
a.sets(pos)
self.assertEqual(a.size, max(pos) + 1)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertTrue(a.get(pos[0]))
self.assertTrue(all(a.gets(pos)))
self.assertEqual(list(a.positions()), sorted(set(pos)))
def add(self, value, position):
if self.index_type == BITMAP_INDEX and len(self.index) > MAX_BIT_INDEX_SIZE:
index = self.index
self.index_type = BLOOMFILTER_INDEX
self.index = Bloomfilter(BF_SIZE, BF_HASHES)
for v in index:
for pos in index[v].positions():
self.index.add([(v, pos)])
if self.index_type == BITMAP_INDEX:
if value not in self.index:
self.index[value] = BitIndex()
index = self.index[value]
index.set(position)
if self.index_type == BLOOMFILTER_INDEX:
self.index.add([(value, position)])
assert (value, position) in self.index
def test_sets(self):
a = BitIndex()
self.assertEqual(a.size, 0)
self.assertEqual(len(a.array), 0)
self.assertFalse(bool(a))
self.assertEqual(str(a), '')
a.set(100, True)
self.assertEqual(a.size, 101)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(
a.array, b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x10')
self.assertTrue(bool(a))
self.assertEqual(str(a), ''.join(['0'] * 100 + ['1']))
a.sets([1, 3, 16])
self.assertEqual(a.size, 101)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(
a.array, b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x10')
a.append()
self.assertEqual(a.size, 102)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(
a.array, b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x30')
a.append(0)
self.assertEqual(a.size, 103)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(
a.array, b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x30')
a.appends([1, 0, 1, True, 1, None])
self.assertEqual(a.size, 109)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(
a.array,
b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\xb0\x0e')
def test_operations(self):
a = BitIndex()
pos_a = set(random.randint(0, 1000) for i in xrange(100))
a.sets(pos_a)
b = BitIndex()
pos_b = set(random.randint(0, 1000) for i in xrange(100))
b.sets(pos_b)
c = BitIndex()
pos_c = set(random.randint(0, 1000) for i in xrange(100))
c.sets(pos_c)
self.assertEqual(list(a.intersect(b)), sorted(pos_a & pos_b))
self.assertEqual(list(a.intersect(b, c)),
sorted(pos_a & pos_b & pos_c))
self.assertEqual(list(a.union(b)), sorted(pos_a | pos_b))
self.assertEqual(list(a.union(b, c)), sorted(pos_a | pos_b | pos_c))
self.assertEqual(list(a.excepts(b)), sorted(pos_a ^ (pos_a & pos_b)))
self.assertEqual(list(a.excepts(b, c)),
sorted(pos_a ^ (pos_a & (pos_b | pos_c))))
self.assertEqual(list(a.xor(b)), sorted(pos_a ^ pos_b))
def test_sets(self):
a = BitIndex()
self.assertEqual(a.size, 0)
self.assertEqual(len(a.array), 0)
self.assertFalse(bool(a))
self.assertEqual(str(a), '')
a.set(100, True)
self.assertEqual(a.size, 101)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(a.array, b'\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x10')
self.assertTrue(bool(a))
self.assertEqual(str(a), ''.join(['0']*100 + ['1']))
a.sets([1,3,16])
self.assertEqual(a.size, 101)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(a.array, b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x10')
a.append()
self.assertEqual(a.size, 102)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(a.array, b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x30')
a.append(0)
self.assertEqual(a.size, 103)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(a.array, b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x30')
a.appends([1,0,1,True,1, None])
self.assertEqual(a.size, 109)
self.assertEqual(len(a.array), (a.size + 7) >> 3)
self.assertEqual(a.array, b'\x0a\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\xb0\x0e')
def test_operations(self):
a = BitIndex()
pos_a = set(random.randint(0, 1000) for i in xrange(100))
a.sets(pos_a)
b = BitIndex()
pos_b = set(random.randint(0, 1000) for i in xrange(100))
b.sets(pos_b)
c = BitIndex()
pos_c = set(random.randint(0, 1000) for i in xrange(100))
c.sets(pos_c)
self.assertEqual(list(a.intersect(b)), sorted(pos_a & pos_b))
self.assertEqual(list(a.intersect(b, c)), sorted(pos_a & pos_b & pos_c))
self.assertEqual(list(a.union(b)), sorted(pos_a | pos_b))
self.assertEqual(list(a.union(b, c)), sorted(pos_a | pos_b | pos_c))
self.assertEqual(list(a.excepts(b)), sorted(pos_a ^ (pos_a & pos_b)))
self.assertEqual(list(a.excepts(b, c)), sorted(pos_a ^ (pos_a & (pos_b | pos_c))))
self.assertEqual(list(a.xor(b)), sorted(pos_a ^ pos_b))
