Performance problems with large insertions

[martijnvermaat]

As you can see, comparing AAAA... with A is instant, but comparing A with AAAA... takes a lot of time:

In [5]: %timeit extractor.describe_dna('A' * 10000, 'A')
10000 loops, best of 3: 129 µs per loop

In [6]: %timeit extractor.describe_dna('A', 'A' * 10000)
1 loops, best of 3: 1.13 s per loop

Perhaps more importantly, memory usage also sky rockets. I couldn't run this test with a 50 Kbp sample sequence on a machine with 4G memory, completely freezing my machine for half a minute. I would like to prevent this from happening on the server.

I didn't look into this further, but I suspect it tries to find the inserted sequence in the original sequence, which of course is not possible. Could this be an easy case to optimize?