Parse a VCF file into a data structure

[figueroakl]

Description

As a developer, I'd like to know what the user's VCF file version is to get us started on what parsing a VCF file into a data structure entails.
The VCF version can be found within the VCF headers. The headers are the first lines of the file and they start with "##".

One thing to think about is that reading the heders and variant records (or rows) into 2 different structures. Both them can be streaming parsers meaning that they do not read the whole file into memory.

This component is responsible for reading a VCF file, which contains genetic variation data, and converting it into a data structure that can be used by the tool. For invetigation or inspiration we can look to tools like rust-htslib or rust-vcf to help with this.

This story leads to 2 other stories, linked below:
#14 - headers data structure
#15 - variant records data structure

Acceptance criteria

Tasks

[stevieing]

Example file VCF4 with description and structure https://www.internationalgenome.org/wiki/Analysis/vcf4.0/

[sb10]

Someone specifically mentioned the desire to support https://samtools.github.io/hts-specs/VCFv4.4.pdf which the current libraries don't support fully (?).

[sb10]

I also don't think this issue should be taken at face-value. It sounds like it might involve pulling all the information from a VCF file in to memory. Since VCF files can be very large, I think we wouldn't want to do that.

How about a starter user story and UAT:

"As a user, I'd like to know the version of my VCF file, so that I can know if its compatible with a certain piece of software"
UAT: Given a valid VCF file of any version, the parser returns the correct version found in the first line.

[jeff-k]

We can read the first lines of the filestream while they start with the header symols, ##, into memory to construct a static header struct.

After that the VCF file object can implement an iterator that yields parsed Row structs.

So, if a users wants to load a huge VCF from stdin or a file, only the relevant lines are actually read if they only want to access the version field of the header

Maybe something like:

struct Header { ... } // includes a mapping between header keys and values

struct Row { ... } // the stuff like chromosome, position, ref, alt, etc

struct VCF {
    header: Header,
    buf: BufReader, // or however we store a reference to the mutable, open file handle
}

impl Iterator for VCF {
// yields Row structs
}