My workflow process the following optional inputs params.yaml

1. samples id, alignment (aln) file OR
2. Sample id, variants (vcf) file OR
3. Sample id, both alignment (aln) file and variants (vcf) file.

How do I give the list of output files which are varying in numbers (sample id + only vcf (2 output files) or sample id + only aln (5 output files) or sample id + vcf and aln (7 output files) based on the the above given optional inputs for each sample as a common input sets to a process ga4gh_metrics. Appreciate any help.

params.yaml

samples:
-
 biosample_id: NA12878 
 aln: /data/NA12878.bam
 vcf: /data/NA12878.hard-filtered.vcf.gz
-
 biosample_id: NA12877
 aln: /data/NA12877.bam
 vcf: /data/NA12877.hard-filtered.vcf.gz

OR

samples:
-
 biosample_id: NA12878 
 aln: /data/NA12878.bam
-
 biosample_id: NA12877
 aln: /data/NA12877.bam

OR

samples:
-
 biosample_id: NA12878 
 vcf: /data/NA12878.hard-filtered.vcf.gz
-
 biosample_id: NA12877
 vcf: /data/NA12877.hard-filtered.vcf.gz

…

Channel
    samples.map { it.biosample_id }
    .set { sample_ids }

…

// channel for samplelist vcf input file processed outputs
  Channel
    .empty()
    sample_ids
    .join( count_variants.out )
    .join( bcftools_stats.out )
    .set { vcf_qc }

..

// aln can be either bam or cram format
// channel for samplelist input file type bam processed outputs
  Channel
    .empty()
    sample_ids
    .join( samtools_stats_bam.out.stats )
    .join( picard_collect_multiple_metrics_bam.out.insert_size )
    .join( picard_collect_multiple_metrics_bam.out.quality )
    .join( picard_collect_wgs_metrics_bam.out.wgs_coverage )
    .join( verifybamid2_bam.out.freemix, remainder: true )
    .set { ch_bam }

// channel for samplelist input file type cram processed outputs
  Channel
    .empty()
    sample_ids
    .join( samtools_stats_cram.out.stats )
    .join( picard_collect_multiple_metrics_cram.out.insert_size )
    .join( picard_collect_multiple_metrics_cram.out.quality )
    .join( picard_collect_wgs_metrics_cram.out.wgs_coverage )
    .join( verifybamid2_cram.out.freemix, remainder: true )
    .set { ch_cram }

…

// channel to mix the bam/cram process outputs and/or variants process outputs 
  ch_bam.mix(ch_cram) // .ifEmpty([]) The channel contain sample id, will not be fully empty even if the aln (bam/cram) input are not given 
    //.combine(vcf_qc,by:0)
    .join(vcf_qc, remainder: true)
    .map { file -> file - null } //.map { it.minus(null) }
    //.map { sample, stats, insertsize, quality, wgs_coverage, freemix, count_variants, bcftools_stats -> [ sample, stats ?: [], insertsize ?: [], quality ?: [], wgs_coverage ?: [], freemix ?: [], count_variants ?: [], bcftools_stats ?: [] ] }
    .view()
    //.flatten()
    .set { ga4gh_metrics_in }

  ga4gh_metrics( ga4gh_metrics_in )

I get the following lists…
If the input samples list (params.yaml) contain..

1. only vcf file, then the processed output to channel ga4gh_metrics are… (sample id + 2 files in the list)

[NA12878, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.variant_counts.json, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.pass.stats]

2. only aln file, then the processed output to channel ga4gh_metrics are… (sample id + 5 files in the list)

[NA12878, ../NPM-sample-qc/tests/NA12878_1000genomes-dragen-3.7.6/work/../NA12878.stats, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.insert_size_metrics.txt, /../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.quality_yield_metrics.txt, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878_wgs_metrics.txt, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.selfSM]

3. both aln and vcf files, the processed output to channel ga4gh_metrics are…(sample id + 7 files in the list)

[NA12878, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.stats, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.insert_size_metrics.txt, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.quality_yield_metrics.txt, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878_wgs_metrics.txt, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.selfSM, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.variant_counts.json, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.pass.stats]

I filter the null (.map { file -> file – null }) from the list if the input contain only alignment (aln) or variants (vcf) files.
eg. only vcf input option

[NA12878, null, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.variant_counts.json, ../NA12878_1000genomes-dragen-3.7.6/work/../NA12878.pass.stats]

process ga4gh_metrics

process ga4gh_metrics {

  tag { sample }

  input:
  ??????

  //tuple val(sample), path('*')

  //tuple val(sample), path(stats), path(picard_insert_size), path(picard_quality), path(picard_wgs_coverage), path(verifybamid_freemix), path(count_variants), path(bcftools_stats) // both aln and vcf processed outputs

  //tuple val(sample), path(variantscounts), path(stats) // only vcf processed output 

  //tuple val(sample), path(stats), path(picard_insert_size), path(picard_quality), path(picard_wgs_coverage), path(verifybamid_freemix) // only aln processed output