To fix bug #55

[DLBPointon]

During testing today we found a bug caused by the CSV_PULL module, I has recently begun mixing it's input channels with another instance of the module, e.g, the module was being run twice with the correct input values and yet the module would pull a variable from another instance. This resulted in the pipeline failing due to incorrect output files.

This has now been corrected, by removing 1 of the inputs which is not needed for the running of the pipeline.

The pipeline passes the local tests and s3 tests will be fixed in the future.

nf-core lint passes

prettier is failing due to local node-14 requirements

[muffato]

What is the purpose of this csv_pull module in the first place ? It seems it's just a pass-through module, the output being the same as the input. If the goal is to make sure that the input file is downloaded to disk just once (in case it's a remote file) and used throughout, I think that's not needed because (as far as I understand !), Nextflow will anyway download the file just once under work/stage and symlink that into all the jobs that need it.

[DLBPointon]

It converts the string to a path object. Without it, it will not stage anything from S3 only from local.

With the function is can do both, for local it is simply a pass-through. If there is a cleaner method then I'll happily switch over.

[muffato]

Ah, I see. It feels like something that could be fixed by adding a path or file somewhere, but my first attempt at it didn't work. Let's merge this pull-request, and if I find a solution I'll let you know.