Update chopper module to take optional fasta file (#6913)

* Update chopper module to take optional fasta file * linting issues * Upgrade chopper biocontainer, add stub, and use only data in test-data for testing * bump chopper version in conda
nf-core · Nov 1, 2024 · 2273783 · 2273783
1 parent 6c32d37
commit 2273783
Show file tree

Hide file tree

Showing 5 changed files with 143 additions and 15 deletions.
diff --git a/modules/nf-core/chopper/environment.yml b/modules/nf-core/chopper/environment.yml
@@ -2,4 +2,4 @@ channels:
   - conda-forge
   - bioconda
 dependencies:
-  - bioconda::chopper=0.3.0
+  - bioconda::chopper=0.9.0
diff --git a/modules/nf-core/chopper/main.nf b/modules/nf-core/chopper/main.nf
@@ -4,11 +4,12 @@ process CHOPPER {
 
     conda "${moduleDir}/environment.yml"
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/chopper:0.3.0--hd03093a_0':
-        'biocontainers/chopper:0.3.0--hd03093a_0' }"
+        'https://depot.galaxyproject.org/singularity/chopper:0.9.0--hdcf5f25_0':
+        'biocontainers/chopper:0.9.0--hdcf5f25_0' }"
 
     input:
     tuple val(meta), path(fastq)
+    path  fasta
 
     output:
     tuple val(meta), path("*.fastq.gz") , emit: fastq
@@ -22,6 +23,7 @@ process CHOPPER {
     def args2  = task.ext.args2  ?: ''
     def args3  = task.ext.args3  ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
+    def fasta_filtering = fasta ? "--contam ${fasta}" : ""
 
     if ("$fastq" == "${prefix}.fastq.gz") error "Input and output names are the same, set prefix in module configuration to disambiguate!"
     """
@@ -30,6 +32,7 @@ process CHOPPER {
         $fastq | \\
     chopper \\
         --threads $task.cpus \\
+        $fasta_filtering \\
         $args2 | \\
     gzip \\
         $args3 > ${prefix}.fastq.gz
@@ -39,4 +42,15 @@ process CHOPPER {
         chopper: \$(chopper --version 2>&1 | cut -d ' ' -f 2)
     END_VERSIONS
     """
+
+    stub:
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    echo | gzip > ${prefix}.fastq.gz
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        chopper: \$(chopper --version 2>&1 | cut -d ' ' -f 2)
+    END_VERSIONS
+    """
 }
diff --git a/modules/nf-core/chopper/meta.yml b/modules/nf-core/chopper/meta.yml
@@ -38,6 +38,10 @@ input:
         type: file
         description: FastQ with reads from long read sequencing e.g. PacBio or ONT
         pattern: "*.{fastq.gz}"
+  - - fasta:
+        type: file
+        description: An optional reference fasta file against which to remove reads that align to it.
+        pattern: "*.fasta"
 output:
   - fastq:
       - meta:

diff --git a/modules/nf-core/chopper/tests/main.nf.test b/modules/nf-core/chopper/tests/main.nf.test
@@ -7,7 +7,43 @@ nextflow_process {
     tag "modules"
     tag "modules_nfcore"
 
-    test("Should run without failures") {
+    test("test with lambda reference") {
+
+        when {
+            params {
+                outdir = "$outputDir"
+            }
+            process {
+                """
+                input[0] =  [
+                                [id:'test_out' ], // meta map
+                                file(params.modules_testdata_base_path + 'genomics/sarscov2/nanopore/fastq/test_2.fastq.gz', checkIfExists: true)
+                            ]
+                input[1] = file(params.modules_testdata_base_path + 'genomics/sarscov2/genome/genome.fasta.gz', checkIfExists: true)
+                """
+            }
+        }
+
+        then {
+
+            def fastq_content = path(process.out.fastq.get(0).get(1)).linesGzip
+
+            assertAll(
+                { assert process.success },
+                // original pytest checks
+                { assert process.out.fastq.get(0).get(1) ==~ ".*/test_out.fastq.gz" },
+                { assert !fastq_content.contains("@a52a642e-88d0-4584-babd-414ea84db484 runid=71c83ae0021f873e29b130c6562a4c27185f93b8 read=2768 ch=489 start_time=2021-08-11T12:07:39Z flow_cell_id=FAQ57606 protocol_group_id=210811_47CoV_SA sample_id=CS5 barcode=barcode04 barcode_alias=barcode04")},
+                // additional nf-test checks
+                // Order of reads is not deterministic, so only assess whether the number of reads is correct
+                { assert snapshot(
+                    fastq_content.size(),
+                    process.out.versions
+                    ).match() }
+            )
+        }
+    }
+
+        test("test without lambda reference") {
 
         when {
             params {
@@ -19,6 +55,7 @@ nextflow_process {
                                 [id:'test_out' ], // meta map
                                 file(params.modules_testdata_base_path + 'genomics/sarscov2/nanopore/fastq/test.fastq.gz', checkIfExists: true)
                             ]
+                input[1] = []
                 """
             }
         }
@@ -34,12 +71,37 @@ nextflow_process {
                 { assert fastq_content.contains("@2109d790-67ec-4fd1-8931-6c7e61908ff3 runid=97ca62ca093ff43533aa34c38a10b1d6325e7e7b read=52274 ch=243 start_time=2021-02-05T23:27:30Z flow_cell_id=FAP51364 protocol_group_id=data sample_id=RN20097 barcode=barcode01 barcode_alias=barcode01")},
                 // additional nf-test checks
                 // Order of reads is not deterministic, so only assess whether the number of reads is correct
-                { assert snapshot(fastq_content.size()).match("number_of_lines") },
-                { assert snapshot(process.out.versions).match("versions") }
-
+                { assert snapshot(
+                    fastq_content.size(),
+                    process.out.versions
+                ).match() }
             )
         }
-
     }
 
-}
+    test("test-chopper-stub") {
+        options '-stub'
+
+        when {
+            process {
+                """
+                input[0] =  [
+                                [id:'test_out' ], // meta map
+                                file(params.modules_testdata_base_path + 'genomics/sarscov2/nanopore/fastq/test.fastq.gz', checkIfExists: true)
+                            ]
+                input[1] = []
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(
+                    process.out,
+                    process.out.versions
+                ).match() }
+            )
+        }
+    }
+}
diff --git a/modules/nf-core/chopper/tests/main.nf.test.snap b/modules/nf-core/chopper/tests/main.nf.test.snap
@@ -1,16 +1,64 @@
 {
-    "versions": {
+    "test without lambda reference": {
         "content": [
+            400,
             [
-                "versions.yml:md5,5fe28ea455482c9fe88603ddcc461881"
+                "versions.yml:md5,74a27493c09d0c481f6e52b517e12023"
             ]
         ],
-        "timestamp": "2023-10-20T08:27:24.592662298"
+        "meta": {
+            "nf-test": "0.9.0",
+            "nextflow": "24.04.4"
+        },
+        "timestamp": "2024-10-31T15:14:36.37897815"
     },
-    "number_of_lines": {
+    "test with lambda reference": {
         "content": [
-            400
+            15984,
+            [
+                "versions.yml:md5,74a27493c09d0c481f6e52b517e12023"
+            ]
+        ],
+        "meta": {
+            "nf-test": "0.9.0",
+            "nextflow": "24.04.4"
+        },
+        "timestamp": "2024-10-31T15:14:31.324993049"
+    },
+    "test-chopper-stub": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test_out"
+                        },
+                        "test_out.fastq.gz:md5,68b329da9893e34099c7d8ad5cb9c940"
+                    ]
+                ],
+                "1": [
+                    "versions.yml:md5,74a27493c09d0c481f6e52b517e12023"
+                ],
+                "fastq": [
+                    [
+                        {
+                            "id": "test_out"
+                        },
+                        "test_out.fastq.gz:md5,68b329da9893e34099c7d8ad5cb9c940"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,74a27493c09d0c481f6e52b517e12023"
+                ]
+            },
+            [
+                "versions.yml:md5,74a27493c09d0c481f6e52b517e12023"
+            ]
         ],
-        "timestamp": "2023-10-20T08:27:24.581289647"
+        "meta": {
+            "nf-test": "0.9.0",
+            "nextflow": "24.04.4"
+        },
+        "timestamp": "2024-10-31T15:29:08.715579423"
     }
 }