Resolved bug when prefetching large studies

Prefetching large studies resulted in error stated in issue

nf-core#236

Resolved this by handing over text file with filepaths instead
of list of filepaths. Rewrote sra_merge_samplesheet accordingly.

Tested with study with ~25k samples.

modules/local/sra_merge_samplesheet/main.nf

@@ -6,26 +6,28 @@ process SRA_MERGE_SAMPLESHEET {
         'nf-core/ubuntu:20.04' }"
 
     input:
-    path ('samplesheets/*')
-    path ('mappings/*')
-
+    // path ('samplesheets/*')
+    // path ('mappings/*')
+    path 'samplesheets.txt'
+    path 'mappings.txt'
+
     output:
     path "samplesheet.csv", emit: samplesheet
     path "id_mappings.csv", emit: mappings
     path "versions.yml"   , emit: versions
 
     script:
     """
-    head -n 1 `ls ./samplesheets/* | head -n 1` > samplesheet.csv
-    for fileid in `ls ./samplesheets/*`; do
+    head -n 1 `head -n 1 samplesheets.txt` > samplesheet.csv
+    while read fileid; do
         awk 'NR>1' \$fileid >> samplesheet.csv
-    done
-
-    head -n 1 `ls ./mappings/* | head -n 1` > id_mappings.csv
-    for fileid in `ls ./mappings/*`; do
+    done < samplesheets.txt
+    
+    head -n 1 `head -n 1 mappings.txt` > id_mappings.csv
+    while read fileid; do
         awk 'NR>1' \$fileid >> id_mappings.csv
-    done
-
+    done < mappings.txt
+    
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
         sed: \$(echo \$(sed --version 2>&1) | sed 's/^.*GNU sed) //; s/ .*\$//')

workflows/sra/main.nf

@@ -118,14 +118,24 @@ workflow SRA {
         params.nf_core_rnaseq_strandedness ?: 'auto',
         params.sample_mapping_fields
     )
+
+
+    // Combine all sample sheet/mapping paths into two files
+    ch_samplesheets=SRA_TO_SAMPLESHEET.out.samplesheet.map{it[1].toString()}.collectFile(name: 'samplesheets.txt', newLine: true)
+    ch_mappings=SRA_TO_SAMPLESHEET.out.mappings.map{it[1].toString()}.collectFile(name: 'mappings.txt', newLine: true)
 
     //
     // MODULE: Create a merged samplesheet across all samples for the pipeline
     //
+    // SRA_MERGE_SAMPLESHEET (
+    //     SRA_TO_SAMPLESHEET.out.samplesheet.collect{it[1]},
+    //     SRA_TO_SAMPLESHEET.out.mappings.collect{it[1]}
+    // )
     SRA_MERGE_SAMPLESHEET (
-        SRA_TO_SAMPLESHEET.out.samplesheet.collect{it[1]},
-        SRA_TO_SAMPLESHEET.out.mappings.collect{it[1]}
+        ch_samplesheets,
+        ch_mappings
     )
+
     ch_versions = ch_versions.mix(SRA_MERGE_SAMPLESHEET.out.versions)
 
     //