From e222d3b8068152221a818842d21f0e94500dc72e Mon Sep 17 00:00:00 2001
From: Matthieu Muffato <mm49@sanger.ac.uk>
Date: Tue, 31 Oct 2023 12:57:15 +0000
Subject: [PATCH] Collect the size of the genome (file size is a good proxy)

---
 subworkflows/local/prepare_genome.nf | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/subworkflows/local/prepare_genome.nf b/subworkflows/local/prepare_genome.nf
index b5f23fd..5264569 100644
--- a/subworkflows/local/prepare_genome.nf
+++ b/subworkflows/local/prepare_genome.nf
@@ -19,12 +19,15 @@ workflow PREPARE_GENOME {
 
     // Uncompress genome fasta file if required
     if ( params.fasta.endsWith('.gz') ) {
-        ch_fasta    = GUNZIP ( fasta ).gunzip
+        ch_unzipped = GUNZIP ( fasta ).gunzip
         ch_versions = ch_versions.mix ( GUNZIP.out.versions )
     } else {
-        ch_fasta    = fasta
+        ch_unzipped = fasta
     }
 
+    ch_unzipped
+    | map { meta, fa -> [ meta + [id: fa.baseName, genome_size: fa.size()], fa] }
+    | set { ch_fasta }
 
     // Unmask genome fasta
     UNMASK ( ch_fasta )