[Enhancement] Support xz and/or zstd in addition to gzip

[corneliusroemer]

For SARS-CoV-2 sequences gzip isn't really a good compression method. Everyone uses xz but zstd also works well.

It'd be great if seqkit supported one or both of these compression formats for input/output compression as well.

[shenwei356]

We use the xopen for I/O, which recognize gzip data according to the magic number,
the golang packages for zstd and xz might be easy to be integrated.

[corneliusroemer]

That would be amazing! I can imagine this being useful for many people.

[shenwei356]

Sorry, it's late, but I've made it. 🎆

• seqkit_linux_amd64.tar.gz
• seqkit_darwin_amd64.tar.gz

$ seqkit stats ../tests/hairpin.* -e
file                     format  type  num_seqs    sum_len  min_len  avg_len  max_len
../tests/hairpin.fa      FASTA   RNA     28,645  2,949,871       39      103    2,354
../tests/hairpin.fa.gz   FASTA   RNA     28,645  2,949,871       39      103    2,354
../tests/hairpin.fa.xz   FASTA   RNA     28,645  2,949,871       39      103    2,354
../tests/hairpin.fa.zst  FASTA   RNA     28,645  2,949,871       39      103    2,354