Closed pb-jchin closed 9 years ago
(I said I'd take a look at this.)
We could refactor this a bit, but the code is already pretty readable to me. (That's an advantage of pypeflow, and of the script-oriented workflow.) We could add some tests for the complicated bits, but I don't want to slow you down.
Yes. I hope the current code capture the right (or at least, a working) structure for the data flow. Based upon this, we can start to work to make it more smooth for satisfying various future requirements.
significant workflow change. Instead of doing phasing and haplotig-assembly contig-by-contig, we do phasing for all contig and then haplotig-assembly for all contigs at once. We only have to scan the overlap files (
*.las
files) once this way.