New simulated dataset, new "kevlar localize" command

[standage]

This PR addresses two issues that have captured my focus recently.

• First, we're in dire need of a data set for testing where 1) we know the "correct" answers (variant calls) and 2) we can run and re-run the kevlar workflow quickly. I've simulated various data sets for testing & continuous integration purposes in the past: there are on the side of "too trivial". I've also simulated some larger data sets recently, but these are on the side of "take too long to run".

• Second, up until recently I've been doing all variant calling by examining alignments manually. It's past time to automate this! But this underscored the need for the first point, so there's been a bit of yak shaving going on.

I'm happy to present notebook/human-sim-pico and kevlar localize.

• The directory notebook/human-sim-pico has the complete record for how I produced a 2.5 Mb random human-like genome from scratch. This includes a Jupyter notebook, some data files, several commands, and a bit of commentary.

• The command kevlar localize takes a kevlar assemble-generated Fasta file, invokes BWA to localize the k-mers in the reference genome, and (assuming it maps to a single region) extracts the genomic interval associated with the assembled contig(s) plus a bit. The assembled contig(s) will then be aligned to this genomic region with a dynamic programming solution to be implemented soon by Fereydoun.

This PR still needs a bit of cleanup (mostly documentation and tests) before it is merged.

[codecov-io]

Codecov Report

Merging #83 into master will increase coverage by 0.33%.
The diff coverage is 87.27%.

@@            Coverage Diff             @@
##           master      #83      +/-   ##
==========================================
+ Coverage   82.72%   83.05%   +0.33%     
==========================================
  Files          27       29       +2     
  Lines        1430     1523      +93     
  Branches      220      239      +19     
==========================================
+ Hits         1183     1265      +82     
- Misses        205      211       +6     
- Partials       42       47       +5

Impacted Files	Coverage Δ
kevlar/overlap.py	70.5% <0%> (ø)	⬆️
kevlar/novel.py	75.96% <100%> (ø)	⬆️
kevlar/cli/localize.py	100% <100%> (ø)
kevlar/reaugment.py	91.66% <100%> (ø)	⬆️
kevlar/seqio.py	93.06% <100%> (+0.08%)	⬆️
kevlar/__init__.py	90.27% <100%> (+0.13%)	⬆️
kevlar/cli/__init__.py	100% <100%> (ø)	⬆️
kevlar/assemble.py	84.82% <50%> (ø)	⬆️
kevlar/filter.py	63.11% <50%> (ø)	⬆️
kevlar/localize.py	86.41% <86.41%> (ø)
... and 2 more

---

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 586c55e...df991ed. Read the comment docs.