Support separate odgi depth -b for arbitrary interval depth

[sampsyo]

In #259, we added an end-to-end fgfa window-depth command that both calculates evenly-sized windows and then computes their average depth. The original script we're working from does this in two commands, though, so I was interested in implementing them separately too. What we needed was a drop-in replacement for odgi depth -b something.bed, which I have now added.

I refactored the stuff in window_depth a bit so that we could share almost all of the implementation across the "fixed-function" window version and the arbitrary-interval version.

I then added odgi depth -b to flash. Following the benchmarks in #259, we can now compare odgi and fgfa against a slightly adapted shell script. Here's that three-way comparison:

$ hyperfine -w1 -r3 \
    'sh odgi_wdepth.sh > odgi_result.bed' \
    'fgfa -i chr8.flatgfa window-depth "chm13#chr8" 5000 > fgfa_result.bed' \
    'flash flash_wdepth.sh > flash_result.bed'

Benchmark 1: sh odgi_wdepth.sh > odgi_result.bed
  Time (mean ± σ):     15.203 s ±  0.358 s    [User: 13.980 s, System: 3.043 s]
  Range (min … max):   14.810 s … 15.509 s    3 runs

Benchmark 2: fgfa -i chr8.flatgfa window-depth "chm13#chr8" 5000 > fgfa_result.bed
  Time (mean ± σ):     989.9 ms ±   3.6 ms    [User: 849.1 ms, System: 138.1 ms]
  Range (min … max):   987.3 ms … 994.1 ms    3 runs

Benchmark 3: flash flash_wdepth.sh > flash_result.bed
  Time (mean ± σ):      1.322 s ±  0.003 s    [User: 1.083 s, System: 0.234 s]
  Range (min … max):    1.318 s …  1.324 s    3 runs

  Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet system without any interferences from other programs. It might help to use the '--warmup' or '--prepare' options.

Summary
  fgfa -i chr8.flatgfa window-depth "chm13#chr8" 5000 > fgfa_result.bed ran
    1.34 ± 0.01 times faster than flash flash_wdepth.sh > flash_result.bed
   15.36 ± 0.37 times faster than sh odgi_wdepth.sh > odgi_result.bed

As expected, flash pays some overhead w/r/t the all-in-one fgfa invocation in order to write that BED file and then parse it back, but it's not too much (1.34x ain't bad). The goal is to mostly eliminate this through optimizations.

---

As a reminder, here's the script for odgi:

odgi depth -i chr8.pan.og -r chm13#chr8 | \
    bedtools makewindows -b /dev/stdin -w 5000 > chm13.chr8.w5kbps.bed
odgi depth -i chr8.pan.og -b chm13.chr8.w5kbps.bed --threads 2 | \
    bedtools sort

Here's the flash script:

odgi depth -i chr8.flatgfa -r "chm13#chr8" | bedtools makewindows -b /dev/stdin -w 5000 > chm13.chr8.w5kbps.bed
odgi depth -i chr8.flatgfa -b chm13.chr8.w5kbps.bed --threads 2

Some notes on the shell syntax:

• I'm doing this without optimizations now, hence the hard-coded FlatGFA input file. With -O, you could use the .og input file and not pay for it.
• Our shell syntax parser sadly doesn't support escaped newlines (Skip escaped newlines raphamorim/flash#13).
• --threads 2 doesn't do anything; it's silently ignored.
• We're skipping bedtools sort, which is a little unfair.