Improvement for pileup memory usage

[athos]

cljam.algo.pileup tends to consume a very large amount of memory and in the worst case it ends up with OOME.

This PR mitigates the situation by improving memory efficiency for the pileup data structure. The primary changes are as follows:

• Use more space-efficient data structures (eg. primitive arrays, dedicated deftypes) to represent a pileup result
• Drop unnecessary intermediate data (:seqs-at-ref / :quals-at-ref ) from the pileup result
• Unchunk a pileup result (which was a chunked sequence) and return a lazy sequence instead to make each element easier to consume and deallocate one by one

In my small examination, the changes reduced the memory usage by ~45% at peak time with almost no impact on time efficiency:

See here for the detailed raw profiling data for the above result.

[codecov]

Codecov Report

Merging #188 into master will increase coverage by <.01%.
The diff coverage is 96%.

@@            Coverage Diff             @@
##           master     #188      +/-   ##
==========================================
+ Coverage   86.67%   86.68%   +<.01%     
==========================================
  Files          76       76              
  Lines        6051     6055       +4     
  Branches      501      501              
==========================================
+ Hits         5245     5249       +4     
  Misses        305      305              
  Partials      501      501

Impacted Files	Coverage Δ
src/cljam/algo/pileup.clj	96.66% <96%> (+0.06%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e98b889...ea0d300. Read the comment docs.

[athos]

Rebased the patch onto the master branch.

[alumi]

Thank you for the investigation and PR! Added a suggestion comment.

[alumi]

LGTM 👍

[yito88]

Thank you for improving that!
LGTM 👍

[athos]

Thanks for the review! 🙌