Improve samplesheet generation

[grst]

Description of feature

It's great that the pipeline can generate samplesheets for downstream pipelines automatically.
I suggest the following improvements:

• always generate all types of samplesheets. It's computationally cheap, it saves one pipeline parameter, and sometimes the data can be used for multiple pipelines (e.g. rnafusion + rnaseq)
• allow to explicitly set the strandedness for rnaseq samplesheets. It's currently always 'auto', but I'd like to set it explicitly to reverse for Illumina libraries.

Something we could also discuss at some later point is to carry on information from the demux samplesheet to the generated samplesheets, e.g. additional columns from the [Data] section could be added to the generated samplesheets.

@apeltzer @atrigila @nschcolnicov

[grst]

Another issue: the output paths are not correct.

They currently are:

s3://xxxxxxxxxxxxx-dev/demultiplex/241003_A01664_XXXXXXXXXXXXXXX/101_d_xxx_S1_L001/101_d_xxx_S1_L001_R1_001.fastq.gz

while it should be

s3://xxxxxxxxxxxxx-dev/demultiplex/241003_A01664_XXXXXXXXXXXXXXX/241003_A01664_XXXXXXXXXXXXXXX/101_d_xxx_S1_L001_R1_001.fastq.gz

The path until s3://xxxxxxxxxxxxx-dev/demultiplex/241003_A01664_XXXXXXXXXXXXXXX is the output folder specified by the user, so it's not taking into account the publishDir subdirectory.

[grst]

I just tested this with the dev version, and it seems the output paths are still incorrect as described in this comment: #272 (comment)

[nschcolnicov]

@grst Ok, I'll take a look

[nschcolnicov]

PR merged, closing this ticket