Error using multi_easy_bin with --no-recluster

[Ralwoss]

Hi,

I'm working with @bebatut on updating the Galaxy Wrapper for SemiBin. There was a problem using SemiBin multi_easy_bin with the parameter --no-recluster.

SemiBin multi_easy_bin \
--separator ':' \
--taxonomy-annotation-table test-data/taxonomy*.tsv \
--input-fasta test-data/input_multi.fasta.gz \
--input-bam test-data/input_multi*.bam \
--output 'output' \
--cannot-name 'cannot'  \
--ratio 0.05   \
--orf-finder 'fraggenescan' \
--random-seed 0 \
--epoches 20 \
--batch-size 2048 \
--max-node 1.0 \
--max-edges 200 \
--minfasta-kbs 200 \
--no-recluster

The error is :

Traceback (most recent call last):
  File "/home/ralwoss/anaconda3/envs/SemiBin/bin/SemiBin", line 10, in <module>
    sys.exit(main1())
             ^^^^^^^
  File "/home/ralwoss/anaconda3/envs/SemiBin/lib/python3.11/site-packages/SemiBin/main.py", line 1482, in main1
    main2(args, is_semibin2=False)
  File "/home/ralwoss/anaconda3/envs/SemiBin/lib/python3.11/site-packages/SemiBin/main.py", line 1468, in main2
    multi_easy_binning(
  File "/home/ralwoss/anaconda3/envs/SemiBin/lib/python3.11/site-packages/SemiBin/main.py", line 1270, in multi_easy_binning
    for bf in os.listdir(os.path.join(output, 'samples', sample, bin_dir_name)):
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: 'output/samples/S1/output_prerecluster_bins'

Without --no-recluster everything works. With --no-recluster there is no output/samples/S1/output_prerecluster_bins directory, only an output/samples/S1/output directory.

It would be great if you could take a look

[psj1997]

Sorry for the late reply. I have fixed this issue.

Thanks for more feedbacks from you.

Sincerely
Shaojun

[bebatut]

Thanks for fixing it.

Would it be possible to have a release with the fix so a new conda package will be available and can be used in Galaxy wrapper? Thanks

[luispedro]

We released a new version (v1.5.1). Conda sometimes takes a few hours for the auto-update to kick in, but it will be done later today or tomorrow. HTH, Luis

[luispedro]

Maybe their bot is better because it was actually immediate: bioconda/bioconda-recipes#39775

Once those tests pass, it can be merged, so there will be a new conda version in an hour or so

[Ralwoss]

Hi,
this issue is back in SemiBin2.
multi_easy_bin --write-pre-reclustering-bins can not find the directory output_bins because there are output_pre_recluster_bins and output_after_recluster_bins instead of output_bins.

It would be great if you could take a look

[luispedro]

Thanks for the report, I can reproduce

[luispedro]

FYI, we released a new version (2.0.2) to fix this issue. At least this particular combination of arguments is now part of our test suite and future regressions will be caught

[Ralwoss]

Thanks a lot