Centrifuge db build (nf-core#3008)

* Add inputs and outputs for db build * Pytest modules for db build * Add test scripts * Correct problem with cf output files * Use same version in singularity and docker containers Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Correct biocontainer name Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Rename taxonomy to taxonomy_tree Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Update modules/nf-core/centrifuge/build/main.nf Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove files form metamix test Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove files form metamix test Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove comment Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove metamix lines Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove reference to kraken Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Change taxonomy to taxonomy_tree Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Update fasta description Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Alternative extension for fasta file Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Update conversion table description Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Update taxonomy tree description Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Update modules/nf-core/centrifuge/build/meta.yml Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Update cf description Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove reference to kraken Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Update modules/nf-core/centrifuge/build/main.nf Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove . Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove reference to kraken Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove empty line Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove empty line Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove empty line Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove empty line Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove empty line Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Remove empty line Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> * Change build to build_test to avoid conflict with pytest internal behaviour * Change to build_test * Remove trailing whitespace * Symlink to build folder + prettier * Correct symlink * Remove unnecessary change to test * Correct nextflow typo * Tidy commas * Tidying and update descriptions * Correct indentation * Change symlink * Test again with synlink * Symlink test * Same centrifuge version as kreport * Correct typo * Change description * Migrate to nf-test * Post linting cleaning and remove final pytest leftovers * Harshil align and add missing input file * Clean up after reviews --------- Co-authored-by: Sofia Stamouli <91951607+sofstam@users.noreply.github.com> Co-authored-by: sofstam <sofia.stamouli@scilifelab.se> Co-authored-by: James A. Fellows Yates <jfy133@gmail.com>
asp8200 · Feb 13, 2024 · 64163cd · 64163cd
1 parent 8d8f0ae
commit 64163cd
Show file tree

Hide file tree

Showing 6 changed files with 282 additions and 0 deletions.
diff --git a/modules/nf-core/centrifuge/build/environment.yml b/modules/nf-core/centrifuge/build/environment.yml
@@ -0,0 +1,7 @@
+name: centrifuge_build
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - bioconda::centrifuge=1.0.4.1
diff --git a/modules/nf-core/centrifuge/build/main.nf b/modules/nf-core/centrifuge/build/main.nf
@@ -0,0 +1,58 @@
+process CENTRIFUGE_BUILD {
+    tag "$meta.id"
+    label 'process_high'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/centrifuge:1.0.4.1--hdcf5f25_1' :
+        'biocontainers/centrifuge:1.0.4.1--hdcf5f25_1' }"
+
+    input:
+    tuple val(meta), path(fasta)
+    path conversion_table
+    path taxonomy_tree
+    path name_table
+    path size_table
+
+    output:
+    tuple val(meta), path("*.cf") , emit: cf
+    path "versions.yml"           , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def size_table_cmd = size_table ? "--size_table ${size_table}" : ""
+    """
+    centrifuge-build \\
+        -p $task.cpus \\
+        $fasta \\
+        ${prefix} \\
+        --conversion-table $conversion_table \\
+        --taxonomy-tree $taxonomy_tree \\
+        --name-table $name_table \\
+        ${size_table_cmd}
+
+    cat <<-END_VERSIONS > versions.yml
+        "${task.process}":
+        centrifuge: \$( centrifuge --version | sed -n 1p | sed 's/^.*centrifuge-class version //')
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}.1.cf
+    touch ${prefix}.2.cf
+    touch ${prefix}.3.cf
+    touch ${prefix}.4.cf
+
+    cat <<-END_VERSIONS > versions.yml
+        "${task.process}":
+        centrifuge: \$( centrifuge --version | sed -n 1p | sed 's/^.*centrifuge-class version //')
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/centrifuge/build/meta.yml b/modules/nf-core/centrifuge/build/meta.yml
@@ -0,0 +1,68 @@
+name: centrifuge_build
+description: Build centrifuge database for taxonomic profiling
+keywords:
+  - database
+  - metagenomics
+  - build
+  - db
+  - fasta
+tools:
+  - centrifuge:
+      description: Classifier for metagenomic sequences
+      homepage: https://ccb.jhu.edu/software/centrifuge/
+      documentation: https://ccb.jhu.edu/software/centrifuge/manual.shtml
+      doi: 10.1101/gr.210641.116
+      licence: ["GPL v3"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+
+  - fasta:
+      type: file
+      description: Fasta file containing sequences to be used in centrifuge database.
+      pattern: "*.{fasta,fna}"
+
+  - conversion_table:
+      type: file
+      description: A tab-separated file with sequence ID to taxonomy ID mapping
+      pattern: "*.{map}"
+
+  - taxonomy_tree:
+      type: file
+      description: A \t|\t-separated file mapping taxonomy. Typically nodes.dmp from the NCBI taxonomy dump. Links taxonomy IDs to their parents
+      pattern: "*.{dmp}"
+
+  - name_table:
+      type: file
+      description: A '|'-separated file mapping taxonomy IDs to a name. Typically names.dmp from the NCBI taxonomy dump. Links taxonomy IDs to their scientific name
+      pattern: "*.{dmp}"
+
+  - size_table:
+      type: file
+      description: Optional list of taxonomic IDs and lengths of the sequences belonging to the same taxonomic IDs.
+      pattern: "*"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+
+  - cf:
+      type: file
+      description: Index files for the centrifuge database
+      pattern: "*.{cf}"
+
+authors:
+  - "@sarah-buddle"
+  - "@jfy133"
diff --git a/modules/nf-core/centrifuge/build/tests/main.nf.test b/modules/nf-core/centrifuge/build/tests/main.nf.test
@@ -0,0 +1,59 @@
+nextflow_process {
+
+    name "Test Process CENTRIFUGE_BUILD"
+    script "../main.nf"
+    process "CENTRIFUGE_BUILD"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "centrifuge"
+    tag "centrifuge/build"
+
+    test("sarscov2 - fasta") {
+
+        when {
+            process {
+                """
+                input[0] = [ [id: 'test'], file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true) ]
+                input[1] = file(params.test_data['sarscov2']['metagenome']['seqid2taxid_map'], checkIfExists: true)
+                input[2] = file(params.test_data['sarscov2']['metagenome']['nodes_dmp'], checkIfExists: true)
+                input[3] = file(params.test_data['sarscov2']['metagenome']['names_dmp'], checkIfExists: true)
+                input[4] = []
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+    }
+
+    test("sarscov2 - fasta - stub") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = [ [id: 'test'], file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true) ]
+                input[1] = file(params.test_data['sarscov2']['metagenome']['seqid2taxid_map'], checkIfExists: true)
+                input[2] = file(params.test_data['sarscov2']['metagenome']['nodes_dmp'], checkIfExists: true)
+                input[3] = file(params.test_data['sarscov2']['metagenome']['names_dmp'], checkIfExists: true)
+                input[4] = []
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+
+    }
+
+}
diff --git a/modules/nf-core/centrifuge/build/tests/main.nf.test.snap b/modules/nf-core/centrifuge/build/tests/main.nf.test.snap
diff --git a/modules/nf-core/centrifuge/build/tests/tags.yml b/modules/nf-core/centrifuge/build/tests/tags.yml
@@ -0,0 +1,2 @@
+centrifuge/build:
+  - "modules/nf-core/centrifuge/build/**"