replace manual vector initialization

[bmesuere]

This pull request replaces manual vector initialization with vec![value; count] for alpha and path.

Benchmark on my laptop:

Benchmark: Short reads (NC_000913-454.fna)
Benchmark 1: /Users/bart/Code/FragGeneScanRs/target/release/FragGeneScanRs -s /Users/bart/Code/FragGeneScanRs/example/NC_000913-454.fna -t 454_10 -w 0 -o /var/folders/j3/38fskpy159v07np8syk3p_2m0000gn/T/tmp.Bw0QpkNK1h/NC_000913-454
  Time (mean ± σ):     450.7 ms ±   7.4 ms    [User: 437.3 ms, System: 10.1 ms]
  Range (min … max):   441.8 ms … 476.7 ms    20 runs


Benchmark: Complete genome (NC_000913.fna)
Benchmark 1: /Users/bart/Code/FragGeneScanRs/target/release/FragGeneScanRs -s /Users/bart/Code/FragGeneScanRs/example/NC_000913.fna -t complete -w 1 -o /var/folders/j3/38fskpy159v07np8syk3p_2m0000gn/T/tmp.Bw0QpkNK1h/NC_000913
  Time (mean ± σ):     626.8 ms ±   3.9 ms    [User: 509.3 ms, System: 111.1 ms]
  Range (min … max):   620.5 ms … 635.1 ms    20 runs


Benchmark: Long reads (contigs.fna)
Benchmark 1: /Users/bart/Code/FragGeneScanRs/target/release/FragGeneScanRs -s /Users/bart/Code/FragGeneScanRs/example/contigs.fna -t complete -w 1 -o /var/folders/j3/38fskpy159v07np8syk3p_2m0000gn/T/tmp.Bw0QpkNK1h/contigs
  Time (mean ± σ):      3.937 s ±  0.009 s    [User: 3.880 s, System: 0.050 s]
  Range (min … max):    3.928 s …  3.952 s    10 runs

short reads: 450.7 ms ± 7.4 ms
complete genome: 626.8 ms ± 3.9 ms
long reads: 3.937 s ± 0.009 s

See #17 for a comparison.

[Copilot]

Pull request overview

Replaces manual Vec::with_capacity + push loops with vec![value; count] initialization for the alpha and path buffers in the Viterbi forward pass.

Changes:

• Pre-initialize alpha as vec![[0.0; State::COUNT]; seq.len()].
• Pre-initialize path as vec![[Some(State::S); State::COUNT]; seq.len()].
• Remove the manual for _ in 0..seq.len() initialization loop.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

forward() allocates alpha/path using seq.len(), but the function later unconditionally indexes alpha[0] and seq[0..=2] (and writes to alpha[1]/alpha[2]). Since the caller currently only guards is_empty(), inputs with length 1–2 will panic. Consider adding an early return/guard for seq.len() < 3 (either here or in viterbi()/caller) so short records are handled safely.

[bmesuere]

This is a valid comment, but was also valid for the original code. I fixed it at the caller side in FragGeneScanRs.rs.

[ninewise]

Doesn't seem to make a time difference. Whichever way you find more readable.

Benchmark 1: ./benchmark-1 -s example/NC_000913.fna -t complete -w 1 -o /tmp/asdf/NC_000913
  Time (mean ± σ):      1.469 s ±  0.026 s    [User: 0.918 s, System: 0.543 s]
  Range (min … max):    1.432 s …  1.509 s    10 runs
 
Benchmark 2: ./benchmark-2 -s example/NC_000913.fna -t complete -w 1 -o /tmp/asdf/NC_000913
  Time (mean ± σ):      1.487 s ±  0.103 s    [User: 0.927 s, System: 0.553 s]
  Range (min … max):    1.351 s …  1.587 s    10 runs

Summary
  ./benchmark-1 -s example/NC_000913.fna -t complete -w 1 -o /tmp/asdf/NC_000913 ran
    1.01 ± 0.07 times faster than ./benchmark-2 -s example/NC_000913.fna -t complete -w 1 -o /tmp/asdf/NC_000913