These are chat archives for nextflow-io/nextflow

6th
Feb 2019
Jose Espinosa-Carrasco
@JoseEspinosa
Feb 06 08:19
Hi there!
I am sending some jobs to be run on SLURM. My account on SLURM is my user name jespinosa which is the account where NF launches my jobs by default. The way SLURM is configured in our system is that if we want to send jobs to a high priority queue we have to not only send in to the high priority partition sbatch -p highprio(queue NF directive) but also send it to a share account `sbatch -A highprio. I don't know how to the latter in NF, any tip?
I realized that clusterOptionsmay do the trick, sorry
Maxime Garcia
@MaxUlysse
Feb 06 09:44
Rubber duck rules!
Paolo Di Tommaso
@pditommaso
Feb 06 10:22
:+1:
Harshil Patel
@drpatelh
Feb 06 13:31
@pditommaso Is there a way to provide the path to a pre-built conda environment per process in a config file?
e.g.
Paolo Di Tommaso
@pditommaso
Feb 06 13:31
yes, provide the the path to a pre-built conda environment .. ;)
Harshil Patel
@drpatelh
Feb 06 13:32
Is this the wrong syntax?
process {
  withName:imctools {
    module = ['Anaconda2/5.1.0']
    conda = '/path/to/my/environment/'
  }
}
Paolo Di Tommaso
@pditommaso
Feb 06 13:32
look good
Harshil Patel
@drpatelh
Feb 06 13:33
Hmmm. Not working for me.
Paolo Di Tommaso
@pditommaso
Feb 06 13:33
is that path listed by
conda env list
Harshil Patel
@drpatelh
Feb 06 13:35
Yep.
Paolo Di Tommaso
@pditommaso
Feb 06 13:35
then check the conda command in the .command.run script
Harshil Patel
@drpatelh
Feb 06 13:36
N E X T F L O W  ~  version 0.32.0
Launching `main.nf` [kickass_mirzakhani] - revision: 75cab20507
Unknown configuration profile: 'crick_custom'
N E X T F L O W  ~  version 0.32.0
Launching `main.nf` [grave_crick] - revision: 75cab20507
ERROR ~ Unable to parse config file: 'crick_custom.config'

  No signature of method: nextflow.config.ConfigParser$_parse_closure5.conda() is applicable for argument types: (String) values: [conda/envs/nf-core-imcyto-1.0dev/]
  Possible solutions: find(), find(), find(groovy.lang.Closure), find(groovy.lang.Closure), any(), any()

 -- Check '.nextflow.log' file for details
Maybe Im missing something.
Paolo Di Tommaso
@pditommaso
Feb 06 13:36
I would start updating to a more recent version
Harshil Patel
@drpatelh
Feb 06 13:47
Thanks @pditommaso Ill try and work out if I can fix it. If not Ill update and let you know if I have any problems.
Paolo Di Tommaso
@pditommaso
Feb 06 13:47
:+1:
Harshil Patel
@drpatelh
Feb 06 15:02
It would help to use the config you have actually specified on the command-line! Sorry @pditommaso .Lack of sleep. Works with 0.32.0.
Paolo Di Tommaso
@pditommaso
Feb 06 15:02
ah-ah :v:
Paolo Di Tommaso
@pditommaso
Feb 06 16:05
Alexey Dushen
@blacky0x0
Feb 06 16:12
Is it possible to get rid of newline character while splitting text file? splitter_out_channel.splitText(newLine: false)
Paolo Di Tommaso
@pditommaso
Feb 06 16:14
channel.splitText() { it.trim() }
Alexey Dushen
@blacky0x0
Feb 06 16:21
got java.lang.StackOverflowError: null = (
micans
@micans
Feb 06 16:22
.splitText().map { it.trim() } ?
is what I have in one of my pipelines
Paolo Di Tommaso
@pditommaso
Feb 06 16:24
weird, do instead
Channel.from('foo\nbar\n').splitText().map { it.trim() }
Alexey Dushen
@blacky0x0
Feb 06 16:24
yep, it works with map, thanks
Tobias "Tobi" Schraink
@tobsecret
Feb 06 17:33
@stevekm @pditommaso thanks for sorting that out, I can also talk with our cluster admins to make sure we have a temporary server we can log into.
Anthony Ferrari
@af8
Feb 06 17:43
Sorry for the lazyness,intervals_count = intervals_count_ch.count()
how to make intervals_count an actual integer, to use it as size parameter in groupTuple ?
Paolo Di Tommaso
@pditommaso
Feb 06 17:45
what are referring ?
Anthony Ferrari
@af8
Feb 06 19:27
count1 = channel1.count() I've got this channel of count1 elements.
Then I have another channel for which I need to use groupTuple with a fixed size of count1.
channel2 = channel2.groupTuple(by: 0, size: count1, remainder: false)
And I get this error :
ERROR ~ Value 'DataflowVariable(value=10)' cannot be used in in parameter 'size' for operator 'groupTuple' -- Value don't match: class java.lang.Integer
Thanks
Anthony Ferrari
@af8
Feb 06 19:57
count1 = channel1.count().get() works but don't know if it is the right way to do it.