Memory testing

[Adoni5]

Blocked, waiting on jguhlin/minimap2-rs#40 to be merged.

This PR rewrites the thread pool to be a permanently waiting pool, rather than spawning a new pool each map_batch call.
An improvement to be made would be to park the threads, rather than sleep them but the effect is the same.
There is some memory weirdness around using Py<PyAny>> to store the Dictionary on the side

Waiting until the minimap2-rs crate is updated as currently references a local version that only exists on my laptop.

[Adoni5]

This release switches from spawning threads every time map_batch is called to having a long running thread pool on the Aligner class, which is created if enable_threading is called. This is a naive implementation, but it does work.

adds some complexity in passing reads into the threadpool via an ArrayQueue and then transferring them back to map_batch, where they are passed on again to the AlignmentBatchResultIter.

Some unused methods are added for catching memory leaks, which I intend to implement in PyTest. These are map_no_op on the PyMethods Aligner struct and no_op_map on the impl Aligner. Respectively, these take in a sequence and return a Pr e created Alignment and create the pre created alignment to be returned.