output_fmt crashing pipeline on barcoded samples

[marchoeppner]

Operating System

Other Linux (please specify below)

Other Linux

AlmaLinux 9.4

Workflow Version

v1.4.2

Workflow Execution

Command line (Cluster)

Other workflow execution

No response

EPI2ME Version

NA

CLI command run

nextflow run /home/mhoeppner/git/wf-basecalling/main.nf -profile standard --input /var/lib/minknow/data/./Listerien_Test/L02/20241206_1246_MN47840_FBA71669_e99225ff --out_dir results --basecaller_cfg [email protected] --barcode_kit SQK-RBK114-24 --output_fmt fastq

Workflow Execution - CLI Execution Profile

None

What happened?

(yes, I hacked the pipeline - the issue is not related to that tho ;))

I noticed that when running the pipeline "standalone", requesting the demuxed data to be exported to fastq, I am getting a crash with a null value in the input channel of split_calls

My best guess is that this is related to the way this bit is constructed, here:

wf-basecalling/lib/signal/ingress.nf

Line 442 in 9c0f9db

barcode_bams = split_calls(pass, margs.input_cache)

Basically, the input to split_calls ('pass') when requesting fastq output is not a cram/crai pair, but just the previously generated fastq file.

This probably should not happen, right? Maybe something got switched around here?

Relevant log output

DataflowVariable(value=null)
WARN: Input tuple does not match tuple declaration in process `wf_dorado:split_calls` -- offending value: /home/mhoeppner/test/work/c1/ddc92659aad1fd109a183936f1dbba/SAMPLE.pass.fq.gz
WARN: Environment variable `REF_PATH` evaluates to an empty value
ERROR ~ Error executing process > 'wf_dorado:split_calls'

Caused by:
  Path value cannot be null



Tip: you can replicate the issue by changing to the process work dir and entering the command `bash .command.run`

 -- Check '.nextflow.log' file for details

Application activity log entry

No response

Were you able to successfully run the latest version of the workflow with the demo data?

yes

Other demo data information

No response