Add tag uniqueness checks

Plant-Food-Research-Open · Feb 28, 2024 · 66de124 · 66de124
1 parent 39f6854
commit 66de124
Show file tree

Hide file tree

Showing 5 changed files with 47 additions and 4 deletions.
diff --git a/assets/schema_xref_assemblies.json b/assets/schema_xref_assemblies.json
@@ -1,6 +1,6 @@
 {
     "$schema": "http://json-schema.org/draft-07/schema",
-    "$id": "https://raw.githubusercontent.com/plant-food-research-open/assemblyqc/master/assets/schema_input.json",
+    "$id": "https://raw.githubusercontent.com/plant-food-research-open/assemblyqc/master/assets/schema_xref_assemblies.json",
     "title": "plant-food-research-open/assemblyqc pipeline - params.synteny_xref_assemblies schema",
     "description": "Schema for the file provided with params.synteny_xref_assemblies",
     "type": "array",

diff --git a/lib/WorkflowAssemblyqc.groovy b/lib/WorkflowAssemblyqc.groovy
@@ -48,6 +48,40 @@ class WorkflowAssemblyqc {
         }
     }
 
+    public static ArrayList validateInput(input) {
+        def inputFields = 5
+        def assemblyTags = input[(0..input.size()-1).step(inputFields)]
+
+        def tagCounts = [:]
+        assemblyTags.each { tag ->
+            tagCounts[tag] = tagCounts.containsKey(tag) ? tagCounts[tag] + 1 : 1
+        }
+        def repeatedTags = tagCounts.findAll { key, count -> count > 1 }.collect { key, count -> key }
+
+        if (repeatedTags.size() > 0) {
+            Nextflow.error("Please check input assemblysheet -> Multiple assemblies have the same tags!: ${repeatedTags}")
+        }
+
+        return input
+    }
+
+    public static ArrayList validateXrefAssemblies(xref) {
+        def xrefFields = 3
+        def xrefTags = xref[(0..xref.size()-1).step(xrefFields)]
+
+        def tagCounts = [:]
+        xrefTags.each { tag ->
+            tagCounts[tag] = tagCounts.containsKey(tag) ? tagCounts[tag] + 1 : 1
+        }
+        def repeatedTags = tagCounts.findAll { key, count -> count > 1 }.collect { key, count -> key }
+
+        if (repeatedTags.size() > 0) {
+            Nextflow.error("Please check synteny_xref_assemblies -> Multiple xref assemblies have the same tags!: ${repeatedTags}")
+        }
+
+        return xref
+    }
+
     public static String jsonifyParams(params) {
         return JsonOutput.toJson(params).toString()
     }

diff --git a/nextflow.config b/nextflow.config
@@ -145,10 +145,11 @@ profiles {
         shifter.enabled        = false
         charliecloud.enabled   = false
         apptainer.enabled      = false
-        docker.runOptions      = '-u $(id -u):$(id -g)'
+        // docker.runOptions      = '-u $(id -u):$(id -g)'
     }
     arm {
-        docker.runOptions      = '-u $(id -u):$(id -g) --platform=linux/amd64'
+        // docker.runOptions      = '-u $(id -u):$(id -g) --platform=linux/amd64'
+        docker.runOptions      = '--platform=linux/amd64'
     }
     singularity {
         singularity.enabled    = true

diff --git a/tests/stub/stub.config b/tests/stub/stub.config
@@ -25,7 +25,7 @@ params {
 
     hic                         = 'tests/stub/hic/Dummy_hic.R{1,2}.fq.gz'
 
-    synteny_skip                = false
+    synteny_skip                = true // GitHub action runner runs out of memory
     synteny_xref_assemblies     = 'https://raw.githubusercontent.com/plant-food-research-open/assemblyqc/dev/assets/xrefsheet.csv'
 
     // Limit resources so that this can run on GitHub Actions

diff --git a/workflows/assemblyqc.nf b/workflows/assemblyqc.nf
@@ -65,12 +65,17 @@ include { CUSTOM_DUMPSOFTWAREVERSIONS       } from '../modules/nf-core/custom/du
 
 // Info required for completion email and summary
 def assemblyqc_report                       = []
+def input_assembly_sheet_fields             = 5
+def synteny_xref_assemblies_fields          = 3
 
 workflow ASSEMBLYQC {
 
     // Input channels
     ch_versions                             = Channel.empty()
     ch_input                                = Channel.fromSamplesheet('input')
+                                            | collect
+                                            | flatMap { WorkflowAssemblyqc.validateInput(it) }
+                                            | buffer(size: input_assembly_sheet_fields)
 
     ch_target_assemby_branch                = ch_input
                                             | map { tag, fasta, gff, mono_ids, labels ->
@@ -126,6 +131,9 @@ workflow ASSEMBLYQC {
     ch_xref_assembly                        = params.synteny_skip || ! params.synteny_xref_assemblies
                                             ? Channel.empty()
                                             : Channel.fromSamplesheet('synteny_xref_assemblies')
+                                            | collect
+                                            | flatMap { WorkflowAssemblyqc.validateXrefAssemblies(it) }
+                                            | buffer(size: synteny_xref_assemblies_fields)
                                             | map { tag, fa, labels ->
                                                 [ tag, file(fa, checkIfExists: true), file(labels, checkIfExists: true) ]
                                             }