ENH: refactor `orient-seqs` revcomp mode to use vsearch `--fastx_revcomp`

[nbokulich]

Problem

The reverse complement mode of orient-seqs action (to revcomp sequences if no orientation reference is passed as input) uses scikit-bio to loop through a FASTA and reverse complement each sequence.

This might be faster / less maintenance to use vsearch instead.

Solution

These lines:

RESCRIPt/rescript/orient.py

Lines 88 to 91 in 009611c

	oriented = DNAFASTAFormat()
	with oriented.open() as out_fasta:
	for seq in sequences.view(DNAIterator):
	seq.reverse_complement().write(out_fasta)

Should be replaced with a vsearch subprocess like this:

RESCRIPt/rescript/orient.py

Lines 96 to 100 in 009611c

	def _vsearch_revcomp_fastq(seqs_fp, out_fp):
	cmd = [
	'vsearch',
	'--fastx_revcomp', str(seqs_fp),
	'--fastqout', str(out_fp),

but note that --fastaout should be used instead of --fastqout (the _vsearch_revcomp_fastq function should just be refactored to make this more generalizable, using the appropriate output format depending on the input type).

Question

I think the main reason to do this would be for speed — it would be useful to check if vsearch --fastx_revcomp is any faster with large FASTA inputs.