Port the read-only functionality of FastaDir in the seqrepo Python code.

We currently don't have issue or PR templates. We should have them.

Full FASTA sequences are read by the noodles crate.

The fix has already been merged into the head branch. However, a release with this patch is currently pending.

• zaeleus/noodles#146

In the case that seqrepo is to be used as a library, people don't want to have the cli.

We do not want it to give us lower code coverage.

Describe the bug
The find() method doesn't use the Query seqid property to search against the seq_id column, but rather, looks it up as an alias.

To Reproduce
A Query object like the below won't return any alias records:

    let sr = SeqRepo::new("/usr/local/share/seqrepo", "latest").unwrap();
    let alias_db = sr.alias_db();
    let query = Query{
        namespace: Some(Namespace::new("%")),
        alias: Some("%".to_string()),
        seqid: Some("yVj-SzT0jvX2CBlpieemKnZe9FTd1zQl".to_string()),
        current_only: false,
    };
    let _ = alias_db.find(&query, |record| {
        println!("{:?}", record.unwrap());
    });

The following, however, will work -- because alias entries prefix GA4GH IDs with GS_:

    let sr = SeqRepo::new("/usr/local/share/seqrepo", "latest").unwrap();
    let alias_db = sr.alias_db();
    let query = Query{
        namespace: Some(Namespace::new("%")),
        alias: Some("%".to_string()),
        seqid: Some("GS_yVj-SzT0jvX2CBlpieemKnZe9FTd1zQl".to_string()),
        current_only: false,
    };
    let _ = alias_db.find(&query, |record| {
        println!("{:?}", record.unwrap());
    });

Expected behavior
The above snippet should print this record when the givenseqid arg is "yVj-SzT0jvX2CBlpieemKnZe9FTd1zQl":

AliasDbRecord { seqalias_id: 9349420, seqid: "yVj-SzT0jvX2CBlpieemKnZe9FTd1zQl", alias: "GS_yVj-SzT0jvX2CBlpieemKnZe9FTd1zQl", added: 2017-07-19T03:33:58, is_current: true, namespace: Namespace { value: "VMC" } }

Additional context
I'm guessing this was just a typo?

We should split the interface and the sqlite-based implementation into separate features. This enables client code to depend on the interface only.

Rather than providing the license.txt file, specify license directly.

seqrepo-cli export should work similar to the Python app.

Put the original apache footers to the files and state clearly how/that the code was originally derived from the bio Commons python package.

The aim is to write out a FASTA file with the requested sequences/regions. This is useful in CI situations of dependents.

We will need to provide a trait Interface so the types can be used interchangeably.