Aggregate from Path will fail in certain aggregation process when certain samples have non-complete subworkflow

[gongyixiao]

Below lines of code/script will fail when doing aggregate from path, and have samples haven't finished corresponding subworkflows.
The goal of aggregate from path is to leverage whatever have already completed samples/subworkflows to construct a cohort to the best of it can. It was designed to tolerate missing samples/subworkflows and aggregate whatever there is.
Since below scripts specifically naming files, instead of using wildcard like the examples at the end here shows, the process will fail since aggrgate will create invalid link based on the path, which will be ignored by using wildcard but error will raise when specifying file names.

tempo/modules/process/Aggregate/GermlineAggregateSv.nf

Line 14 in e29f5be

awk '\$1 ~ /^#/ && FNR < NR {next;}{print}' ${bedpeFiles} > sv_germline.bedpe

tempo/modules/process/Aggregate/SomaticAggregateHRDetect.nf

Line 13 in e29f5be

awk 'FNR==1 && NR!=1{next;}{print}' ${hrdetectFiles} > hrdetect.tsv

tempo/modules/process/Aggregate/SomaticAggregateSv.nf

Line 14 in e29f5be

awk '\$1 ~ /^#/ && FNR < NR {next;}{print}' ${bedpeFiles} > sv_somatic.bedpe

And for loops in the script block of this process
https://github.com/mskcc/tempo/blob/develop/modules/process/Aggregate/CohortRunMultiQC.nf

Successful examples:

tempo/modules/process/Aggregate/SomaticAggregateMaf.nf

Line 21 in e29f5be

for i in mut/*.maf ; do

tempo/modules/process/Aggregate/SomaticAggregateLOHHLA.nf

Line 20 in e29f5be

awk 'FNR==1 && NR!=1{next;}{print}' lohhla/*HLAlossPrediction_CI.txt > HLAlossPrediction_CI.txt