remove compress input

mirpedrol · Jul 5, 2024 · 9a2ca2e · 9a2ca2e
1 parent 2844bf1
commit 9a2ca2e
Show file tree

Hide file tree

Showing 29 changed files with 85 additions and 400 deletions.
diff --git a/modules/class-modules/clustalo/align/main.nf b/modules/class-modules/clustalo/align/main.nf
@@ -10,19 +10,17 @@ process CLUSTALO_ALIGN {
     input:
     tuple val(meta) , path(fasta)
     tuple val(meta2), path(tree)
-    val(compress)
 
     output:
-    tuple val(meta), path("*.aln{.gz,}"), emit: alignment
-    path "versions.yml"                 , emit: versions
+    tuple val(meta), path("*.aln.gz"), emit: alignment
+    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 write_output = compress ? "--force -o >(pigz -cp ${task.cpus} > ${prefix}.aln.gz)" : "> ${prefix}.aln"
     // using >() is necessary to preserve the return value,
     // so nextflow knows to display an error when it failed
     // the --force -o is necessary, as clustalo expands the commandline input,
@@ -33,7 +31,7 @@ process CLUSTALO_ALIGN {
         -i ${fasta} \
         --threads=${task.cpus} \
         $args \
-        $write_output
+        --force -o >(pigz -cp ${task.cpus} > ${prefix}.aln.gz)
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
@@ -46,7 +44,7 @@ process CLUSTALO_ALIGN {
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
     """
-    touch ${prefix}.aln${compress ? '.gz' : ''}
+    touch ${prefix}.aln.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":

diff --git a/modules/class-modules/clustalo/align/meta.yml b/modules/class-modules/clustalo/align/meta.yml
@@ -39,22 +39,17 @@ input:
         type: file
         description: Input guide tree in Newick format
         pattern: "*.{dnd}"
-  - - compress:
-        type: boolean
-        description: Flag representing whether the output MSA should be compressed.
-          Set to true to enable/false to disable compression. Compression is done using
-          pigz, and is multithreaded.
 output:
   - alignment:
       - meta:
           type: map
           description: |
             Groovy Map containing sample information
             e.g. `[ id:'test']`
-      - "*.aln{.gz,}":
+      - "*.aln.gz":
           type: file
           description: Alignment file, in gzipped fasta format
-          pattern: "*.aln{.gz,}"
+          pattern: "*.aln.gz"
           ontologies:
             - edam: http://edamontology.org/format_1984
   - versions:

diff --git a/modules/class-modules/clustalo/align/tests/main.nf.test b/modules/class-modules/clustalo/align/tests/main.nf.test
@@ -11,31 +11,8 @@ nextflow_process {
     tag "clustalo/align"
     tag "clustalo/guidetree"
 
-    test("sarscov2 - contigs-fasta - uncompressed") {
 
-        when {
-            process {
-                """
-                input[0] = [ [ id:'test' ], // meta map
-                             file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
-                           ]
-                input[1] = [[:],[]]
-                input[2] = false
-                """
-            }
-        }
-
-        then {
-            assertAll(
-                { assert process.success },
-                { assert snapshot(process.out.alignment).match("alignment - uncompressed")},
-                { assert snapshot(process.out.versions).match("versions0") }
-            )
-        }
-
-    }
-
-    test("sarscov2 - contigs-fasta - compressed") {
+    test("sarscov2 - contigs-fasta") {
 
         when {
             process {
@@ -44,15 +21,14 @@ nextflow_process {
                              file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
                            ]
                 input[1] = [[:],[]]
-                input[2] = true
                 """
             }
         }
 
         then {
             assertAll(
                 { assert process.success },
-                { assert snapshot(process.out.alignment).match("alignment - compressed")},
+                { assert snapshot(process.out.alignment).match("alignment ")},
                 { assert snapshot(process.out.versions).match("versions1") }
             )
         }
@@ -82,7 +58,6 @@ nextflow_process {
                              file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
                            ]
                 input[1] = CLUSTALO_GUIDETREE.out.tree.collect{ meta, tree -> tree }.map{ tree -> [[ id: 'test_summary'], tree]}
-                input[2] = true
                 """
             }
         }

diff --git a/modules/class-modules/famsa/align/main.nf b/modules/class-modules/famsa/align/main.nf
@@ -12,27 +12,25 @@ process FAMSA_ALIGN {
     input:
     tuple val(meta) , path(fasta)
     tuple val(meta2), path(tree)
-    val(compress)
 
     output:
-    tuple val(meta), path("*.aln{.gz,}"), emit: alignment
-    path "versions.yml"                 , emit: versions
+    tuple val(meta), path("*.aln.gz"), emit: alignment
+    path "versions.yml"              , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
 
     script:
     def args = task.ext.args ?: ''
-    def compress_args = compress ? '-gz' : ''
     def prefix = task.ext.prefix ?: "${meta.id}"
     def options_tree = tree ? "-gt import $tree" : ""
     """
     famsa $options_tree \\
-        $compress_args \\
+        -gz # compress output \\
         $args \\
         -t ${task.cpus} \\
         ${fasta} \\
-        ${prefix}.aln${compress ? '.gz':''}
+        ${prefix}.aln$.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
@@ -43,7 +41,7 @@ process FAMSA_ALIGN {
     stub:
     def prefix = task.ext.prefix ?: "${meta.id}"
     """
-    touch ${prefix}.aln${compress ? '.gz' : ''}
+    touch ${prefix}.aln.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":

diff --git a/modules/class-modules/famsa/align/meta.yml b/modules/class-modules/famsa/align/meta.yml
@@ -35,23 +35,17 @@ input:
         type: file
         description: Input guide tree in Newick format
         pattern: "*.{dnd}"
-  - - compress:
-        type: boolean
-        description: Flag representing whether the output MSA should be compressed.
-          Set to true to enable/false to disable compression. Compression is handled
-          by passing '-gz' to FAMSA along with any other options specified in task.ext.args.
 output:
   - alignment:
       - meta:
           type: map
           description: |
             Groovy Map containing sample information
             e.g. `[ id:'test']`
-      - "*.aln{.gz,}":
+      - "*.aln.gz":
           type: file
-          description: Alignment file, in FASTA format. May be gzipped or uncompressed,
-            depending on if compress is set to true or false
-          pattern: "*.aln{.gz,}"
+          description: Alignment file, in FASTA format.
+          pattern: "*.aln.gz"
           ontologies:
             - edam: http://edamontology.org/format_1984
   - versions:

diff --git a/modules/class-modules/famsa/align/tests/main.nf.test b/modules/class-modules/famsa/align/tests/main.nf.test
@@ -10,31 +10,8 @@ nextflow_process {
     tag "famsa/align"
     tag "famsa/guidetree"
 
-    test("sarscov2 - fasta - uncompressed") {
 
-        when {
-            process {
-                """
-                input[0] = [ [ id:'test' ], // meta map
-                             file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
-                           ]
-                input[1] = [[:],[]]
-                input[2] = false
-                """
-            }
-        }
-
-        then {
-            assertAll(
-                { assert process.success },
-                { assert snapshot(process.out.alignment).match("alignment_uncompressed")},
-                { assert snapshot(process.out.versions).match("versions0") }
-            )
-        }
-
-    }
-
-    test("sarscov2 - fasta - compressed") {
+    test("sarscov2 - fasta") {
 
         when {
             process {
@@ -43,15 +20,14 @@ nextflow_process {
                              file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
                            ]
                 input[1] = [[:],[]]
-                input[2] = true
                 """
             }
         }
 
         then {
             assertAll(
                 { assert process.success },
-                { assert snapshot(process.out.alignment).match("alignment_compressed")},
+                { assert snapshot(process.out.alignment).match("alignment")},
                 { assert snapshot(process.out.versions).match("versions1") }
             )
         }
@@ -80,7 +56,6 @@ nextflow_process {
                              file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
                            ]
                 input[1] = FAMSA_GUIDETREE.out.tree.collect{ meta, tree -> tree }.map{ tree -> [[ id: 'test_summary'], tree]}
-                input[2] = true
                 """
             }
         }

diff --git a/modules/class-modules/kalign/align/main.nf b/modules/class-modules/kalign/align/main.nf
@@ -9,24 +9,22 @@ process KALIGN_ALIGN {
 
     input:
     tuple val(meta), path(fasta)
-    val(compress)
 
     output:
-    tuple val(meta), path("*.aln{.gz,}"), emit: alignment
-    path "versions.yml"                 , emit: versions
+    tuple val(meta), path("*.aln.gz"), emit: alignment
+    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 write_output = compress ? ">(pigz -cp ${task.cpus} > ${prefix}.aln.gz)" : "${prefix}.aln"
     """
     unpigz -cdf $fasta | \\
     kalign \\
         $args \\
-        -o ${write_output}
+        -o >(pigz -cp ${task.cpus} > ${prefix}.aln.gz)
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
@@ -39,7 +37,7 @@ process KALIGN_ALIGN {
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
     """
-    touch ${prefix}.aln${compress ? '.gz' : ''}
+    touch ${prefix}.aln.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":

diff --git a/modules/class-modules/kalign/align/meta.yml b/modules/class-modules/kalign/align/meta.yml
@@ -25,23 +25,17 @@ input:
         pattern: "*.{fa,fasta}{.gz,}"
         ontologies:
           - edam: http://edamontology.org/format_1929
-  - - compress:
-        type: boolean
-        description: Flag representing whether the output MSA should be compressed.
-          Set to true to enable/false to disable compression. Compression is done using
-          pigz, and is multithreaded.
 output:
   - alignment:
       - meta:
           type: map
           description: |
             Groovy Map containing sample information
             e.g. `[ id:'test']`
-      - "*.aln{.gz,}":
+      - "*.aln.gz":
           type: file
-          description: Alignment file. May be gzipped or uncompressed, depending on if
-            `compress` is set to `true` or `false`.
-          pattern: "*.{aln}{.gz,}"
+          description: Alignment file.
+          pattern: "*.aln.gz"
           ontologies:
             - edam: http://edamontology.org/format_1984
   - versions:

diff --git a/modules/class-modules/kalign/align/tests/main.nf.test b/modules/class-modules/kalign/align/tests/main.nf.test
@@ -10,44 +10,23 @@ nextflow_process {
     tag "kalign"
     tag "kalign/align"
 
-    test("sarscov2 - fasta - uncompressed") {
 
-        when {
-            process {
-                """
-                input[0] = [ [ id:'test' ], // meta map
-                          file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
-                        ]
-                input[1] = false
-                """
-            }
-        }
-
-        then {
-            assertAll(
-                { assert process.success },
-                { assert snapshot(process.out).match("SARS-CoV-2 scaffolds fasta - uncompressed")},
-            )
-        }
-    }
-
-    test("sarscov2 - fasta - compressed") {
+    test("sarscov2 - fasta") {
 
         when {
             process {
                 """
                 input[0] = [ [ id:'test' ], // meta map
                           file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
                         ]
-                input[1] = true
                 """
             }
         }
 
         then {
             assertAll(
                 { assert process.success },
-                { assert snapshot(process.out).match("SARS-CoV-2 scaffolds fasta - compressed")},
+                { assert snapshot(process.out).match("SARS-CoV-2 scaffolds fasta")},
             )
         }
     }

diff --git a/modules/class-modules/learnmsa/align/main.nf b/modules/class-modules/learnmsa/align/main.nf
@@ -9,24 +9,22 @@ process LEARNMSA_ALIGN {
 
     input:
     tuple val(meta), path(fasta)
-    val(compress)
 
     output:
-    tuple val(meta), path("*.aln{.gz,}"), emit: alignment
-    path "versions.yml"                 , emit: versions
+    tuple val(meta), path("*.aln.gz"), emit: alignment
+    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 write_output = compress ? ">(pigz -cp ${task.cpus} > ${prefix}.aln.gz)" : "${prefix}.aln"
     """
     learnMSA \\
         $args \\
         -i <(unpigz -cdf $fasta) \\
-        -o $write_output
+        -o >(pigz -cp ${task.cpus} > ${prefix}.aln.gz)
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
@@ -39,7 +37,7 @@ process LEARNMSA_ALIGN {
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
     """
-    touch ${prefix}.aln${compress ? '.gz' : ''}
+    touch ${prefix}.aln.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":