Update shapeit test and input tuple

modules/nf-core/shapeit5/ligate/main.nf

@@ -22,7 +22,7 @@ process SHAPEIT5_LIGATE {
     def prefix = task.ext.prefix ?: "${meta.id}"
     def suffix = task.ext.suffix ?: "vcf.gz"
     """
-    printf "%s\\n" $input_list | tr -d '[],' > all_files.txt
+    printf "%s\\n" $input_list | tr -d '[],' | sort -V > all_files.txt
 
     SHAPEIT5_ligate \\
         $args \\

modules/nf-core/shapeit5/ligate/tests/main.nf.test

@@ -4,6 +4,8 @@ nextflow_process {
     script "../main.nf"
     process "SHAPEIT5_LIGATE"
 
+    config "./nextflow.config"
+
     tag "modules"
     tag "modules_nfcore"
     tag "shapeit5"
@@ -13,36 +15,17 @@ nextflow_process {
     test("homo sapiens - map, [vcf], [tbi]") {
         config "./nextflow.config"
         setup {
-            run("BCFTOOLS_VIEW", alias: "BCFTOOLS_VIEW_1") {
-                script "../../../bcftools/view"
-                params {
-                    bcftools_args = "--regions chr22:16570000-16600000 -Ob --write-index=csi -e 'GT=\"./.\"||GT=\".\"'"
-                }
-                process {
-                    """
-                    input[0] = [
-                        [ id:'NA12878_1' ], // meta map
-                        file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz", checkIfExists: true),
-                        file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz.csi", checkIfExists: true),
-                    ]
-                    input[1] = []
-                    input[2] = []
-                    input[3] = []
-                    """
-                }
-            }
-            run("BCFTOOLS_VIEW", alias: "BCFTOOLS_VIEW_2") {
+            run("BCFTOOLS_VIEW") {
                 script "../../../bcftools/view"
-                params {
-                    bcftools_args = "--regions chr22:16580000-16610000 -Ob --write-index=csi -e 'GT=\"./.\"||GT=\".\"'"
-                }
                 process {
                     """
-                    input[0] = [
-                        [ id:'NA12878_2' ], // meta map
-                        file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz", checkIfExists: true),
-                        file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz.csi", checkIfExists: true),
-                    ]
+                    input[0] = channel.of(
+                        [ id:'NA12878', region: "chr22:16570000-16600000" ],
+                        [ id:'NA12878', region: "chr22:16580000-16610000" ]
+                    ).combine(channel.of([
+                        file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr21_22.vcf.gz", checkIfExists: true),
+                        file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr21_22.vcf.gz.csi", checkIfExists: true),
+                    ]))
                     input[1] = []
                     input[2] = []
                     input[3] = []
@@ -54,9 +37,8 @@ nextflow_process {
         when {
             process {
                 """
-                input[0] = BCFTOOLS_VIEW_1.out.vcf.join(BCFTOOLS_VIEW_1.out.csi)
-                    .mix(BCFTOOLS_VIEW_2.out.vcf.join(BCFTOOLS_VIEW_2.out.csi))
-                    .map { meta, vcf, csi -> [ [id : "NA12878"], vcf, csi ] }
+                input[0] = BCFTOOLS_VIEW.out.vcf.join(BCFTOOLS_VIEW.out.csi)
+                    .map{ meta, vcf, index -> [meta.subMap("id"), vcf, index]}
                     .groupTuple()
                 """
             }
@@ -66,8 +48,15 @@ nextflow_process {
             assertAll(
                 { assert process.success },
                 { assert snapshot(
-                    process.out.merged_variants.collect {meta, vcf -> [ meta, path(vcf).vcf.summary ]},
-                    process.out.versions
+                    process.out.merged_variants.collect {
+                        meta, vcf -> [
+                            meta,
+                            file(vcf).name,
+                            path(vcf).vcf.summary,
+                            path(vcf).vcf.variantsMD5
+                        ]
+                    },
+                    ["versions": process.out.versions]
                 ).match() }
             )
         }
@@ -81,8 +70,8 @@ nextflow_process {
                 """
                 input[0] = [
                     [id: 'NA12878'],
-                    [file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz", checkIfExists: true)],
-                    [file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz.csi", checkIfExists: true)]
+                    [file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr21_22.vcf.gz", checkIfExists: true)],
+                    [file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr21_22.vcf.gz.csi", checkIfExists: true)]
                 ]
                 """
             }

modules/nf-core/shapeit5/ligate/tests/main.nf.test.snap

@@ -27,10 +27,10 @@
             }
         ],
         "meta": {
-            "nf-test": "0.9.2",
-            "nextflow": "24.10.1"
+            "nf-test": "0.9.3",
+            "nextflow": "25.10.0"
         },
-        "timestamp": "2024-11-22T17:22:34.235879155"
+        "timestamp": "2025-11-26T17:58:05.130390005"
     },
     "homo sapiens - map, [vcf], [tbi]": {
         "content": [
@@ -39,17 +39,21 @@
                     {
                         "id": "NA12878"
                     },
-                    "VcfFile [chromosomes=[chr22], sampleCount=1, variantCount=60, phased=false, phasedAutodetect=false]"
+                    "NA12878.vcf.gz",
+                    "VcfFile [chromosomes=[chr22], sampleCount=1, variantCount=81, phased=false, phasedAutodetect=false]",
+                    "e0d8b01df09f70f7ccd010f42a7e07c5"
                 ]
             ],
-            [
-                "versions.yml:md5,4b66c655a4fd210ca7bc47b6fe35230b"
-            ]
+            {
+                "versions": [
+                    "versions.yml:md5,4b66c655a4fd210ca7bc47b6fe35230b"
+                ]
+            }
         ],
         "meta": {
-            "nf-test": "0.9.2",
-            "nextflow": "24.10.1"
+            "nf-test": "0.9.3",
+            "nextflow": "25.10.0"
         },
-        "timestamp": "2024-12-04T10:07:53.185018802"
+        "timestamp": "2025-11-26T18:01:39.348167832"
     }
 }

modules/nf-core/shapeit5/ligate/tests/nextflow.config

@@ -1,6 +1,6 @@
 process {
-    withName: "BCFTOOLS_VIEW_*" {
-        ext.args = params.bcftools_args
-        ext.prefix = { "${meta.id}" }
+    withName: "BCFTOOLS_VIEW" {
+        ext.args   = { "--regions ${meta.region} -Ob --write-index=csi -e 'GT=\"./.\"||GT=\".\"'" }
+        ext.prefix = { "${meta.id}_${meta.region}" }
     }
 }

modules/nf-core/shapeit5/phasecommon/main.nf

@@ -8,24 +8,25 @@ process SHAPEIT5_PHASECOMMON {
         'biocontainers/shapeit5:5.1.1--hb60d31d_0'}"
 
     input:
-        tuple val(meta) , path(input), path(input_index), path(pedigree), val(region)
-        tuple val(meta2), path(reference), path(reference_index)
-        tuple val(meta3), path(scaffold), path(scaffold_index)
-        tuple val(meta4), path(map)
+        tuple val(meta), path(input), path(input_index), path(pedigree), val(region), path(reference), path(reference_index), path(scaffold), path(scaffold_index), path(map)
 
     output:
-        tuple val(meta), path("*.{vcf,bcf,vcf.gz,bcf.gz}"), emit: phased_variant
-        path "versions.yml"                               , emit: versions
+        tuple val(meta), path("*.{bcf,graph,bh}"), emit: phased_variant
+        path "versions.yml"                      , emit: versions
 
     when:
         task.ext.when == null || task.ext.when
 
     script:
     def args   = task.ext.args   ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
-    def suffix = task.ext.suffix ?: "bcf"
 
-    if ("$input" == "${prefix}.${suffix}") error "Input and output names are the same, set prefix in module configuration to disambiguate!"
+    def extension = args.contains("--output-format bcf")   ? "bcf"   :
+                    args.contains("--output-format graph") ? "graph" :
+                    args.contains("--output-format bh")    ? "bh"    :
+                    "bcf"
+
+    if ("$input" == "${prefix}.${extension}") error "Input and output names are the same, set prefix in module configuration to disambiguate!"
 
     def map_command       = map       ? "--map $map"             : ""
     def reference_command = reference ? "--reference $reference" : ""
@@ -42,24 +43,27 @@ process SHAPEIT5_PHASECOMMON {
         $pedigree_command \\
         --region $region \\
         --thread $task.cpus \\
-        --output ${prefix}.${suffix}
+        --output ${prefix}.${extension}
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
-        shapeit5: "\$(SHAPEIT5_phase_common | sed -nr '/Version/p' | grep -o -E '([0-9]+.){1,2}[0-9]' | head -1)"
+        shapeit5: "\$(SHAPEIT5_phase_common | sed -nr '/Version/p' | grep -o -E '([0-9]+.){1,2}[0-9]' | head -n 1)"
     END_VERSIONS
     """
 
     stub:
-    def prefix     = task.ext.prefix        ?: "${meta.id}"
-    def suffix     = task.ext.suffix        ?: "bcf"
-    def create_cmd = suffix.endsWith(".gz") ? "echo '' | gzip >" : "touch"
+    def args      = task.ext.args   ?: ''
+    def prefix    = task.ext.prefix ?: "${meta.id}"
+    def extension = args.contains("--output-format bcf")   ? "bcf"   :
+                    args.contains("--output-format graph") ? "graph" :
+                    args.contains("--output-format bh")    ? "bh"    :
+                    "bcf"
     """
-    ${create_cmd} ${prefix}.${suffix}
+    touch ${prefix}.${extension}
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
-        shapeit5: "\$(SHAPEIT5_phase_common | sed -nr '/Version/p' | grep -o -E '([0-9]+.){1,2}[0-9]' | head -1)"
+        shapeit5: "\$(SHAPEIT5_phase_common | sed -nr '/Version/p' | grep -o -E '([0-9]+.){1,2}[0-9]' | head -n 1)"
     END_VERSIONS
     """
 }

modules/nf-core/shapeit5/phasecommon/meta.yml

@@ -45,11 +45,6 @@ input:
           Target region, usually a full chromosome (e.g. chr20:1000000-2000000 or chr20).
           For chrX, please treat PAR and non-PAR regions as different choromosome in order to avoid mixing ploidy.
         pattern: "chrXX:leftBufferPosition-rightBufferPosition"
-  - - meta2:
-        type: map
-        description: |
-          Groovy Map containing sample information
-          e.g. [ id:'test', single_end:false ]
     - reference:
         type: file
         description: Reference panel of haplotypes in VCF/BCF format.
@@ -60,11 +55,6 @@ input:
         description: Index file of the Reference panel file.
         pattern: "*.{vcf.gz.csi,bcf.gz.csi}"
         ontologies: []
-  - - meta3:
-        type: map
-        description: |
-          Groovy Map containing sample information
-          e.g. [ id:'test', single_end:false ]
     - scaffold:
         type: file
         description: Scaffold of haplotypes in VCF/BCF format.
@@ -75,14 +65,9 @@ input:
         description: Index file of the scaffold file.
         pattern: "*.{vcf.gz.csi,bcf.gz.csi}"
         ontologies: []
-  - - meta4:
-        type: map
-        description: |
-          Groovy Map containing sample information
-          e.g. [ id:'test', single_end:false ]
     - map:
         type: file
-        description: File containing the genetic map.
+        description: File containing the genetic map in Glimpse format.
         pattern: "*.gmap"
         ontologies: []
 output:
@@ -92,10 +77,10 @@ output:
           description: |
             Groovy Map containing sample information
             e.g. [ id:'test', single_end:false ]
-      - "*.{vcf,bcf,vcf.gz,bcf.gz}":
+      - "*.{bcf,graph,bh}":
           type: file
-          description: Phased variant dataset in VCF/BCF format.
-          pattern: "*.{vcf,bcf,vcf.gz,bcf.gz}"
+          description: Phased variant dataset in BCF, GRAPH or XCF binary format.
+          pattern: "*.{bcf,graph,bh}"
           ontologies: []
   versions:
     - versions.yml:

modules/nf-core/shapeit5/phasecommon/tests/main.nf.test

@@ -4,25 +4,30 @@ nextflow_process {
     script "../main.nf"
     process "SHAPEIT5_PHASECOMMON"
 
+    config "./nextflow.config"
+
     tag "modules"
     tag "modules_nfcore"
     tag "shapeit5"
     tag "shapeit5/phasecommon"
 
-    test("homo sapiens - vcf, [], [], []") {
+    test("homo sapiens - vcf, no reference, no scaffold, no map") {
         when {
+            params {
+                shapeit5_phasecommon_args = "--seed 1"
+            }
             process {
                 """
                 input[0] = Channel.of([
                     [ id:'input', single_end:false ], // meta map
                     file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.vcf.gz", checkIfExists: true),
                     file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.vcf.gz.csi", checkIfExists: true),
                     [],
-                    "chr22"
+                    "chr22",
+                    [],[], // reference
+                    [],[], // scaffold
+                    [] // map
                 ])
-                input[1] = Channel.of([[],[],[]])
-                input[2] = Channel.of([[],[],[]])
-                input[3] = Channel.of([[],[]])
                 """
             }
         }
@@ -31,30 +36,65 @@ nextflow_process {
             assertAll(
                 { assert process.success },
                 { assert snapshot(
-                    process.out.phased_variant.collect{ meta, vcf -> [meta, file(vcf).name]},
-                    process.out.versions
+                    process.out.phased_variant.collect{
+                        meta, vcf -> [ meta, file(vcf).name ] // Only file name tested due to bcf format
+                    },
+                    ["versions": process.out.versions]
                 ).match() }
             )
         }
 
     }
 
-    test("homo sapiens - vcf, [], [], [] - stub") {
-        options '-stub'
+    test("homo sapiens - vcf, no reference, no scaffold, with map") {
         when {
+            params {
+                shapeit5_phasecommon_args = "--seed 1"
+            }
             process {
                 """
                 input[0] = Channel.of([
                     [ id:'input', single_end:false ], // meta map
                     file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.vcf.gz", checkIfExists: true),
                     file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.vcf.gz.csi", checkIfExists: true),
                     [],
-                    "chr22"
+                    "chr22",
+                    [],[], // reference
+                    [],[], // scaffold
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/genome/genetic_map/genome.GRCh38.chr22.glimpse.map", checkIfExists: true)
                 ])
+                """
+            }
+        }
 
-                input[1] = Channel.of([[],[],[]])
-                input[2] = Channel.of([[],[],[]])
-                input[3] = Channel.of([[],[]])
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(
+                    process.out.phased_variant.collect{
+                        meta, vcf -> [ meta, file(vcf).name ] // Only file name tested due to bcf format
+                    },
+                    ["versions": process.out.versions]
+                ).match() }
+            )
+        }
+
+    }
+
+    test("homo sapiens - vcf, [], [], [] - stub") {
+        options '-stub'
+        when {
+            params {
+                shapeit5_phasecommon_args = "--seed 1 --output-format graph"
+            }
+            process {
+                """
+                input[0] = Channel.of([
+                    [ id:'input', single_end:false ], // meta map
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.vcf.gz", checkIfExists: true),
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.vcf.gz.csi", checkIfExists: true),
+                    [], "chr22", [], [], [], [], []
+                ])
                 """
             }
         }