On Tue, 27 Jul, 2021, 9:02 pm Steffen Möller, <steffen_moel...@gmx.de> wrote:
> Hello, > > This package is one of the closest that we have on our radar to suit > virologists, to I went for CATCH from the broad institute at > https://github.com/broadinstitute/catch . Conda has it already, does not > surface on bio.tools, though. > > Testing takes quite a while (~15mins) and the few that fail seem to > depend on viral sequence data that cannot be downloaded. So, I consider > the package functional, but don't really know what to do with the test > that fail because they have to fail. > > Are there suggestions from your sides what to do about them? > > On another note, when I just wanted to inject it all to salsa, it failed > to create the pristine-tar: > > [...] > That file is small > > $ du -sh catch/datasets/data/achimota_rubulavirus_1.fasta.gz > 8.5K catch/datasets/data/achimota_rubulavirus_1.fasta.gz > > so there is no need for git-lsf in the first place, right? > Yes, no need. It looks weird on first look though Any idea? > If you have the sources in a non-git repository could you do something like: $ dpkg-source -b . $ cd .. $ gbp import-dsc --pristine-tar <your .dsc file> And see what you get? > <snip> > Huge test log > </snip> Some of these don't work due to a missing .tsv file, others because they try to access internet w/ urllib, few others due to a missing fasta file. But it's difficult to debug till we have a repository in salsa to look stuff up and fix, so if you could commit, it'd be easy to move forward Nilesh
