def test_nbest(self):
# Less-naive "N-best" algorithm, much faster (if len(data) is big
# enough <wink>) than sorting all of data. However, if we had a max
# heap instead of a min heap, it could go faster still via
# heapify'ing all of data (linear time), then doing 10 heappops
# (10 log-time steps).
data = [random.randrange(2000) for _ in xrange(1000)]
heap = self.makeHeap(data[:10])
key = self.key or (lambda x:x)
for item in data[10:]:
if key(item) > key(heap[0]): # this gets rarer the longer we run
heap.replace(item)
assert list(heap.iterpop()) == self.sorted(data)[-10:]
def test_replace(self):
heap = self.makeHeap([16,12,18,15])
m = self.min(heap); assert heap.replace(13) == m; assert 13 in heap
m = self.min(heap); assert heap.pushpop(17) == m; assert 17 in heap
assert self.makeHeap().pushpop(4) == 4
py.test.raises(TypeError, heap.replace, None, None)
# replace fails on an empty heap
py.test.raises(IndexError, self.makeHeap().replace, 1)
