I'm working on an implementation of a fastqc wrapper written in CWL, and am wondering if there is a better way to ensure that the log outputs of my script are unique (instead of having to name them manually).
outputs:
output_qc_report_file
type: File
outputBinding:
glob: "*_fastqc.zip"
console_log:
type: stdout
error_log:
type: stderr
Instead of fastqc_con.txt for console log, it would be something like "sample_id-tool_ran.txt" or "22_TO_04_S38_R2_001-fastqc_con.txt.
See https://www.commonwl.org/user_guide/05-stdout/index.html
You can choose any name you want for the file created from stdout. Using a javascript expression, you can easily derive the name from other parameters. For your case (using an inputs section i made up) this might look like
inputs:
reads:
type: File
inputBinding:
position: 2
outputs:
output_qc_report_file
type: File
outputBinding:
glob: "*_fastqc.zip"
console_log:
type: stdout
stdout: $((inputs.reads.nameroot)+"_console_log.txt")
Just in case it is useful to you: Here is a fastqc tool wrapper i use. However, it does not capture stdout.
Thank you for the response! It has helped a bit, I now get a complete run of my script with fastqc and log output. I ended using something similar to the solution you posted:
stdout: $((inputs.seqfile.basename)+"_console_log.txt")
However, the logs themselves are named undefined_console_log.txt and undefined_error_log.txt respectively.
Since the code is too long to write out, and copy and pasting is giving me formatting errors, I am going to post screenshots of it
https://photos.app.goo.gl/5PERRYeA9m11qT126
Thank you for your help, I really appreciate it!
Hi again!
inputs.seqfile is an array. While the elements in the array probably have a basename, the array itself does not. Would taking the name of the first file in the array be sufficient for your purposes?
inputs.seqfile[0].basename
Also note that basename includes the file extension, so the output would be something like "mysequence.fastq_console_log.txt". You can use nameroot to get the name without the extension.
Cheers, Tom
A simpler variation of the stdout:
stdout: $(inputs.reads.nameroot)_console_log.txt
Log in to answer this question.
Don't know about the exact capabilities of CWL, but adding
.logto the name of your sample is one of common solutions. Adding a time stamp would be another.