Closed d4straub closed 1 year ago
Currently (2.5.0), empty input files can be ignored with --ignore_empty_input_files, or samples after trimming with --ignore_failed_trimming. "empty input files" is checked based on compressed fastq file sizes (< 1.KB) using file.size(), see subworkflows/local/parse_input.nf and subworkflows/local/cutadapt_workflow.nf. A better solution might be file.countFastq(), example from here:
file.size()
channel .fromPath( 'data/yeast/reads/*.fq.gz' ) .map ({ file -> [file, file.countFastq()] }) .filter({ file, numreads -> numreads > 25000}) .view ({ file, numreads -> "file $file contains $numreads reads" })
This way an exact read count threshold could be defined and even modified if desired. The disadvantage might be the computational overhead to check file lines.
Description of feature
Currently (2.5.0), empty input files can be ignored with --ignore_empty_input_files, or samples after trimming with --ignore_failed_trimming. "empty input files" is checked based on compressed fastq file sizes (< 1.KB) using
file.size(), see subworkflows/local/parse_input.nf and subworkflows/local/cutadapt_workflow.nf. A better solution might be file.countFastq(), example from here:This way an exact read count threshold could be defined and even modified if desired. The disadvantage might be the computational overhead to check file lines.