search

carpentries-incubator / workflows-nextflow

Workflow management with Nextflow and nf-core
https://carpentries-incubator.github.io/workflows-nextflow/
Other
19 stars 29 forks source link

Amend Getting started episode #19

Closed mahesh-panchal closed 3 years ago

mahesh-panchal commented 3 years ago

I started proof reading the getting started section, and then realised it became way too long. I think everything after the first script section should be put somewhere else. Do you have opinions on when the rest should be presented?

I also modified the DSL2 section to follow the Nextflow DSL2 conventions. I think this is something we should encourage as it helps with readability.

Lastly the code example has an issue. It performs a line count on a compressed file. How should we address this? Modify the command to use zcat or use an uncompressed file?

ggrimes commented 3 years ago

I think zcat is the best option to save space zcat file.gz | wc -l

ggrimes commented 3 years ago

I am writing a reporting episode and everything in the log section could be incorporated there. The work directory could be a brief aside or covered in more detail in the log episode

mahesh-panchal commented 3 years ago

How should we best take care of the split off stuff here then? Shall we leave it in the spare file in this commit and move parts to appropriate places in other pull requests?

mahesh-panchal commented 3 years ago

Also are you happy with the episode content reordering?

ggrimes commented 3 years ago

I would like to include the Pipeline parameters section and the Process identification callout.

mahesh-panchal commented 3 years ago

I would like to include the Pipeline parameters section and the Process identification callout.

Where?

ggrimes commented 3 years ago

It would appear after the "Run a Nextflow script" and exercise 1 in episode 1

mahesh-panchal commented 3 years ago

I've added back the part on pipeline parameters and process identification. I still think these should be removed though from getting started.

I think it would be good to have a full episode on workflow parameterisation.

Process identification I think should be put elsewhere. It should be talked about when we go over the work folder. The question is where exactly to put this. Should we make a troubleshooting episode, where the work dir is explained, we add this callout on process id, and how to use the .command.sh and .command.run files?

I've opened this up to merge if you're happy with the current state. I've shorted the parameter challenge because 18 files is unnecessary.