Hi,
I hope this forum is still active.
My group is using Ruffus to run pipelines, and I definitely have troubles in understanding how ruffus deals with names.
It seems that depending on the decorator (@transform versus @subdivide), the name expected is different, and that is very surprising.
It seems if my stages 1, 2 and 3 have :
1 : input : _appl.vcf output : _flqg.vcf
2 : input : .vcf output : .avinput
3 : input : .avinput output : _1000G_filter
If stage 2 is @subdivide (because my stage 2 is basically giving 2 output for 1 input) the file name expected is : *.appl_flqg.avinput and when I use @transform it's *.appl.vcf_filqg.vcf.avinput
That's a huge difficulty then to be able to run a pipeline if the names are not expected to be the same.
My stage 2 uses a software that changes the filename (let's say it adds a part of my samples name in the output file), I have to hack it, by doing a glob on the files by myself, but it seems that ruffus loads all the var in the function at the beginning, so it's pretty hard to re-load the filelist when launching the script. (Basically I have to relaunch the script from the beginning each time)
The documentation is for me quite unclear, can you give me clear examples about how to deal with that?
Best,