Add mafft guidetree (nf-core#7097)

* move mafft to mafft_align * add mafft guidetree * update mafft align * add mafft guidetree tests * fixes * up * update other mafft tests * update mafft * fix * don't lint deleted files * filter out main.nf.test and main.nf.snapshot * try different strategy * fix other * fix other * add tests * remove pytest * up * update nftest * fix --------- Co-authored-by: mashehu <[email protected]>
LouisLeNezet · Dec 4, 2024 · a0657ec · a0657ec
1 parent a35192c
commit a0657ec
Show file tree

Hide file tree

Showing 24 changed files with 502 additions and 91 deletions.
diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml
@@ -65,11 +65,11 @@ jobs:
         with:
           filters: |
             modules:
-              - 'modules/nf-core/**'
-              - 'tests/modules/nf-core/**'
+              - added|modified: 'modules/nf-core/**'
+              - added|modified: 'tests/modules/nf-core/**'
             subworkflows:
-              - 'subworkflows/nf-core/**'
-              - 'tests/subworkflows/nf-core/**'
+              - added|modified: 'subworkflows/nf-core/**'
+              - added|modified: 'tests/subworkflows/nf-core/**'
           token: ""
           list-files: "json"
 
@@ -84,7 +84,7 @@ jobs:
                 .replace('modules/nf-core/', '')
                 .split('/')
                 .slice(0, 2)
-                .filter(x => x !== 'main.nf' && x !== 'tests' && x !== 'meta.yml' && x !== 'environment.yml')
+                .filter(x => !x.startsWith('main.nf') && x !== 'tests' && x !== 'meta.yml' && x !== 'environment.yml')
                 .join('/'))
               )
             ];

diff --git a/modules/nf-core/epang/split/main.nf b/modules/nf-core/epang/split/main.nf
@@ -34,4 +34,18 @@ process EPANG_SPLIT {
         epang: \$(echo \$(epa-ng --version 2>&1) | sed 's/^EPA-ng v//')
     END_VERSIONS
     """
+
+    stub:
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}_query.fasta
+    gzip ${prefix}_query.fasta
+    touch ${prefix}_reference.fasta
+    gzip ${prefix}_reference.fasta
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        epang: \$(echo \$(epa-ng --version 2>&1) | sed 's/^EPA-ng v//')
+    END_VERSIONS
+    """
 }
diff --git a/modules/nf-core/epang/split/tests/main.nf.test b/modules/nf-core/epang/split/tests/main.nf.test
@@ -0,0 +1,83 @@
+nextflow_process {
+
+    name "Test Process EPANG_SPLIT"
+    script "../main.nf"
+    process "EPANG_SPLIT"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "epang"
+    tag "epang/split"
+    tag "mafft/align"
+
+    setup {
+        run("MAFFT_ALIGN") {
+            script "../../../mafft/align/main.nf"
+            process {
+                """
+                input[0] = [ [ id:'test', single_end:false ], // meta map
+                            file('https://raw.githubusercontent.com/nf-core/test-datasets/phyloplace/testdata/PF14720_seed.alnfaa', checkIfExists: true)
+                        ]
+                input[1] = [[id:"test2"],
+                            file('https://raw.githubusercontent.com/nf-core/test-datasets/phyloplace/testdata/PF14720_3_sequences.faa', checkIfExists: true)
+                            ]
+                input[2] = [[:],[]]
+                input[3] = [[:],[]]
+                input[4] = [[:],[]]
+                input[5] = [[:],[]]
+                input[6] = false
+                """
+            }
+        }
+    }
+
+
+    test("PF14720") {
+
+        when {
+            process {
+                """
+                input[0] =  MAFFT_ALIGN.out.fas.map {
+                                [
+                                    [ id:'test'],
+                                    file('https://raw.githubusercontent.com/nf-core/test-datasets/phyloplace/testdata/PF14720_seed.alnfaa', checkIfExists: true),
+                                    it[1]
+                                ]
+                            }
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+    }
+
+
+    test("PF14720 - stub") {
+        options "-stub"
+        when {
+            process {
+                """
+                input[0] =  MAFFT_ALIGN.out.fas.map {
+                                [
+                                    [ id:'test'],
+                                    file('https://raw.githubusercontent.com/nf-core/test-datasets/phyloplace/testdata/PF14720_seed.alnfaa', checkIfExists: true),
+                                    it[1]
+                                ]
+                            }
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+    }
+}
diff --git a/modules/nf-core/epang/split/tests/main.nf.test.snap b/modules/nf-core/epang/split/tests/main.nf.test.snap
@@ -0,0 +1,100 @@
+{
+    "PF14720 - stub": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test_query.fasta.gz:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "1": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test_reference.fasta.gz:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "2": [
+                    "versions.yml:md5,83ff91cc742cb443b1793f2b28dbf21a"
+                ],
+                "query": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test_query.fasta.gz:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "reference": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test_reference.fasta.gz:md5,d41d8cd98f00b204e9800998ecf8427e"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,83ff91cc742cb443b1793f2b28dbf21a"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.1"
+        },
+        "timestamp": "2024-11-27T15:15:38.13000903"
+    },
+    "PF14720": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.query.fasta.gz:md5,ff1159ff8a5d2587fb20309a81da3c29"
+                    ]
+                ],
+                "1": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.reference.fasta.gz:md5,e7753871101c48c42fc417ce0e9b398a"
+                    ]
+                ],
+                "2": [
+                    "versions.yml:md5,83ff91cc742cb443b1793f2b28dbf21a"
+                ],
+                "query": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.query.fasta.gz:md5,ff1159ff8a5d2587fb20309a81da3c29"
+                    ]
+                ],
+                "reference": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "test.reference.fasta.gz:md5,e7753871101c48c42fc417ce0e9b398a"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,83ff91cc742cb443b1793f2b28dbf21a"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.1"
+        },
+        "timestamp": "2024-11-27T14:47:33.857838942"
+    }
+}
diff --git a/modules/nf-core/epang/split/tests/tags.yml b/modules/nf-core/epang/split/tests/tags.yml
@@ -0,0 +1,2 @@
+epang/split:
+  - "modules/nf-core/epang/split/**"
diff --git a/modules/nf-core/mafft/environment.yml → modules/nf-core/mafft/align/environment.yml b/modules/nf-core/mafft/environment.yml → modules/nf-core/mafft/align/environment.yml
diff --git a/modules/nf-core/mafft/main.nf → modules/nf-core/mafft/align/main.nf b/modules/nf-core/mafft/main.nf → modules/nf-core/mafft/align/main.nf
@@ -1,4 +1,4 @@
-process MAFFT {
+process MAFFT_ALIGN {
     tag "$meta.id"
     label 'process_high'
 

diff --git a/modules/nf-core/mafft/meta.yml → modules/nf-core/mafft/align/meta.yml b/modules/nf-core/mafft/meta.yml → modules/nf-core/mafft/align/meta.yml
@@ -1,4 +1,4 @@
-name: mafft
+name: mafft_align
 description: Multiple sequence alignment using MAFFT
 keywords:
   - fasta

diff --git a/modules/nf-core/mafft/tests/main.nf.test → ...es/nf-core/mafft/align/tests/main.nf.test b/modules/nf-core/mafft/tests/main.nf.test → ...es/nf-core/mafft/align/tests/main.nf.test
@@ -1,11 +1,12 @@
 nextflow_process {
 
-    name "Test Process MAFFT"
+    name "Test Process MAFFT_ALIGN"
     script "../main.nf"
-    process "MAFFT"
+    process "MAFFT_ALIGN"
     tag "modules"
     tag "modules_nfcore"
     tag "mafft"
+    tag "mafft/align"
 
     test("SARS-CoV-2 scaffolds fasta - uncompressed") {
         when {
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,2 @@
		epang/split:
		- "modules/nf-core/epang/split/**"