Replace memoizing cache for parsed CIGARs with core.memoize's LU cache

[athos]

cljam has a memoizing cache for parsed CIGAR strings, and it tends to limitlessly increase in size after a long-term process such as pileup. The following chart shows how large the whole memory usage got during a certain run of variant call using cljam's pileup. Note that the bottom line of the used heap size is steadily getting higher as time goes along:

This PR replaces the memoizing cache with core.memoize's LU cache, limiting the cache size to an appropriate upper bound. Also, it defines a new system property named cljam.sam.cigar.cache-size to specify the upper bound of cache size when launching the JVM.

After applying the change, memory usage would grow as follows. In this case, the bottom line keeps as low as the beginning throughout the whole run.

See here for the detailed data for the above charts.

[codecov]

Codecov Report

Merging #189 into master will decrease coverage by 0.01%.
The diff coverage is 66.66%.

@@            Coverage Diff             @@
##           master     #189      +/-   ##
==========================================
- Coverage   86.68%   86.67%   -0.02%     
==========================================
  Files          76       76              
  Lines        6055     6057       +2     
  Branches      501      502       +1     
==========================================
+ Hits         5249     5250       +1     
  Misses        305      305              
- Partials      501      502       +1

Impacted Files	Coverage Δ
src/cljam/io/sam/util/cigar.clj	73.41% <66.66%> (-0.61%)	⬇️

---

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 c320af1...d1c7312. Read the comment docs.

[alumi]

LGTM 👍

[r6eve]

LGTM