Release PR - #115
Release PR#115
Conversation
sguizard
left a comment
There was a problem hiding this comment.
Congratulation for the pipeline.
Some corrections (in comments) need to be applied before approving the PR.
|
@sguizard, thank you very much for the review! I'll incorporate the feedback. |
erikrikarddaniel
left a comment
There was a problem hiding this comment.
I left reviewing to Claude since I don't know the field. Since it found real bugs, I'm not approving yet. Ping me when you've addressed the below. Here's the verdict, plus a few inline comments:
Two things break real use and are cheap to fix. Details are in the inline comments.
- Relative paths are rejected. Every example in docs/usage.md uses relative paths. But rna_matrix/hto_matrix in the samplesheet, and --demultiplexing_result, --cell_genotype, --vireo_filtered_variants, --gt_donors and --common_variants, reach a path input as plain strings. Nextflow only accepts absolute strings there. The user sees only Execution aborted due to an unexpected error; the real cause (Not a valid path value: 'hto.tar.gz') is only in .nextflow.log. Reproduced with -profile test_hashing,docker -stub and a local samplesheet, and with -profile test_donor_match,docker -stub and local files for the four params.
- -stub fails in donor_match mode. The FIND_VARIANTS stub writes file names that do not match its output globs: Missing output file(s) *_all_representative_variants.csv. Reproduced with -profile test_donor_match,docker -stub.
|
@sguizard and @erikrikarddaniel, thank you very much! I incorporated all your feedback. Again, I learned a lot, using MultiQC, reviewing with Claude,... Ready for the first release? 🚀 |
|
Approved! 🍾🎉 |
First release of nf-core/hadge! 🚀