unexpected F1 score running somatic snv/indel variants benchmarking

Description of the bug

Hi,

I am running the following
nextflow run nf-core/variantbenchmarking
-r 1.3.0
-c $inDIR/yale.config
-work-dir $workDIR
--input $inDIR/${smplshtFILE}.csv
--outdir $outPATH/$outDIR
--analysis $analysisTYPE
--variant_type snv
--method sompy \
--truth_id ${smpID}
--truth_vcf $refPATH/${refFILE}.vcf.gz
--preprocess split_multiallelic,normalize,deduplicate,prepy
--fasta $GENOME
--fai $GENOMEIDX

I am testing a VCF derived from RNAseq and one from WGS, against the same WGS. Again, this is to have a sort of control

The resulting sompy.summary.csv is below
Tool,File,Type,TP_base,TP,FN,TP_call,FP,UNK,Recall,Precision,recall_lower,recall_upper,recall2,precision_lower,precision_upper,na,ambiguous,fp.region.size,F1
RNAsentieonGATK,RNAsentieonGATK.661T.GATK.stats.csv,SNVs,5,1,4,67,66,0,0.2,0.0149253731343283,0.0,0.6286264006319943,0.2,0.0,0.0676321414491473,0.0,0.0,675359842.0,0.0977256802900045
WGSsentieonGATK,WGSsentieonGATK.661T.GATK.stats.csv,SNVs,5,5,0,5,0,0,1.0,1.0,0.4781762498950185,1.0,1.0,0.4781762498950185,1.0,0.0,0.0,675359842.0,0.0

The second row VCF is the same as the reference. Beyond the
recall_lower,recall_upper,recall2,precision_lower,precision_upper which I will have to read about, I have F1=0, but Recall and Precision are both 1.0

Am I doing something. wrong ?

Thanks

Command used and terminal output

Relevant files

No response

System information

No response

Reproduce the supplied Nextflow command and inspect the generated sompy.summary.csv, starting with how F1 is calculated from TP, FP, FN, Recall, and Precision. No relevant files or tests are identified; done means the discrepancy is explained and the benchmark reports a consistent F1 value or documents why it is 0.