Added a check for input assembly size

CHANGELOG.md

@@ -26,7 +26,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 17. `eggnogmapper_tax_scope` is now set to 1 (root div) by default
 18. Added a `test` profile based on public data
 19. Added parameter `add_attrs_to_proteins_fasta` to enable/disable addition of decoded gff attributes to proteins fasta [#58](https://github.com/plant-food-research-open/genepal/issues/58)
-20. Updated modules and sub-workflows
+20. Added a check for input assemblies. If an assembly is smaller than 1 MB (or 300KB in zipped format), the pipeline errors out before starting the downstream processes [#47](https://github.com/plant-food-research-open/genepal/issues/47)
+21. Updated modules and sub-workflows
 
 ### `Fixed`
 

docs/output.md

@@ -45,7 +45,6 @@ A repeat library is created with either [REPEATMODELER](https://github.com/Dfam-
 
 - `repeatmasker/`
   - `*.masked`: Masked assembly
-  - `*.gff`: Repeatmasker output in GFF3 format
 
 </details>
 

subworkflows/local/utils_nfcore_genepal_pipeline/main.nf

@@ -85,7 +85,16 @@ workflow PIPELINE_INITIALISATION {
                                     def tag         = it[0]
                                     def fasta       = it[1]
 
-                                    [ [ id: tag ], file(fasta, checkIfExists: true) ]
+                                    def fasta_file  = file(fasta, checkIfExists: true)
+                                    def is_zipped   = fasta.endsWith('.gz')
+                                    def sz_thresh   = is_zipped ? 300_000 : 1_000_000
+                                    def fasta_size  = fasta_file.size()
+
+                                    if ( fasta_size < sz_thresh ) { // < 1 MB
+                                        error "The assembly represented by tag '$tag' is only $fasta_size bytes. The minimum allowed size is 1 MB!"
+                                    }
+
+                                    [ [ id: tag ], fasta_file ]
                                 }
 
     ch_tar_assm_str             = ch_input

tests/short/assemblysheet.csv

@@ -0,0 +1,2 @@
+tag,fasta,is_masked
+sarscov2,https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/sarscov2/genome/genome.fasta.gz,no

tests/short/params.json

@@ -0,0 +1,5 @@
+{
+    "input": "tests/short/assemblysheet.csv",
+    "protein_evidence": "https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/genomics/sarscov2/genome/proteome.fasta.gz",
+    "busco_skip": true
+}