Synchronize screed and khmer read record details

[ctb]

khmer.ReadParser returns read objects that differ from screed-created read objects in silly ways; this should be fixed.

[ctb]

One question - is it OK for a legitimate read to have an empty quality attribute?

[mr-c]

Note: ReadParser reads still refer to 'accuracy'

[ctb]

See also #645

[ctb]

Needs to be verified.

[mr-c]

ReadParser reads do correctly refer to 'quality' instead of 'accuracy'

Reads from FASTA files still have an 'accuracy' attribute (an empty UTF string).

In [6]: o = khmer.ReadParser('tests/test-data/paired.fa')

In [7]: n = o.next()

In [8]: n.
n.annotations  n.name         n.quality      n.sequence  

In [9]: n.quality
Out[9]: u''

[betatim]

Has the opinion on exposing the Read object "fully" changed since #645? Motivation for doing so would be to make it easy to test that screed.Record and khmer.Read look and feel the same.

[ctb]

On Mon, Oct 24, 2016 at 07:02:46AM -0700, Tim Head wrote:

Has the opinion on exposing the Read object "fully" changed since #645? Motivation for doing so would be to make it easy to test that screed.Record and khmer.Read look and feel the same.

Yes, I think it's a good idea for that reason.