Closed AndreaGuarracino closed 1 year ago
Thanks for this, @AndreaGuarracino! I can recreate the fact that it now (correctly) does nothing for the graph LPA.gfa
. However, it now does the same fishy thing for note5.gfa
.
When I run flip on note5, the only meaningful change is that
L 3 - 4 + 0M
becomes
L 4 - 3 + 0M
.
The paths, segments, and other links remain the same.
If I understand flip
correctly, the correct behavior for this graph is to do nothing.
This PR introduces a different way to construct the flipped graph:
In the first tests, this avoids introducing edges that are not used in the flipped paths and avoids losing edges that are present in the input graph.