hifiasm version bump to 0.18.5

[abueg]

draft PR, new hifiasm update today -- biocondabot has merged the PR for it a couple hours earlier

[abueg]

i think the conda update is live!

[bgruening]

The new linter complains about <has_size value="0"/> value 0 is not allowed anymore. Do you recall why we test files with 0 size?

[abueg]

hmm no i didn't write that test, but looking at that test case it seems it would just be a case where everything is in the primary gfa so the alternate gfa is made but empty; should i edit the test to just check if the alternate gfa exists?

[bgruening]

Yes I think this would be nice. @gallardoalba any other idea?

[gallardoalba]

hmm no i didn't write that test, but looking at that test case it seems it would just be a case where everything is in the primary gfa so the alternate gfa is made but empty; should i edit the test to just check if the alternate gfa exists?

The test data was generated from E.coli if I remember correctly, so I guess that hifiasm is not able to generate an alternative assembly because of not haplotypes present in the samples.

[abueg]

working on this, i've incorporated Cristo's changes but the tests won't run without a file to actually check the alt against, but since these cases will just output an empty file, i'm just going to remove the line about the alternate in the outputs -- that being said, the tests now run but some of the actual content seems to have changed with the significant version update, so i'll work on regenerating the test outputs and updating them 👍🏼

[bgruening]

Do you know the planemo test --update_test_data trick?

[abueg]

.... now i do 😅

[abueg]

works fine locally -- looking at the tool tests output artifact on here, it seems test 3 ran with 2 threads but my test was looking for "-t 1" flag in the log -- let me edit it to search for the command text that doesn't include the thread parameter in case that can change

[bgruening]

Thanks a bunch! Automatic update will deploy it this weekend :)