Random access a region in fasta file by fai index

[wjwei-handsome]

Hi,

The fai index file records name/seq_len/byte_offset/line_width/line_byte infomations.

I' working for add an implementation for FastaReader to seek or fetch function to random access the region(chr:start-end).

But when I add a std::io::Seek trait for parse_fastx_reader, I noticed that the std::io::Chain struct didn't satisfied Seek trait.

So I would love to hear your team thoughts on this and if could give me some advices or help.

[audy]

Hi @wjwei-handsome. Can you submit a PR of what you have so far? Supporting fai files is an interesting use case but I'm not sure if it will be compatible with current functionality as this library assumes sequential, rather than random, access to data (for example, it's impossible to seek if you're reading from /dev/stdin).

[wjwei-handsome]

As you say, sequential data is necessary, actually, only plain FASTA file support.

I'm trying to imitate the IndexedReader struct, I still need some time to have a try.

Thanks for your advice!

[wjwei-handsome]

In addition, I noticed that needletail has a perfect byte positioning system, which theoretically can achieve random access based on fai.

[wjwei-handsome]

@audy Hi, I just finished this function, please check PR :)