Add HLA-HD v1.7.1 module and BAM-input subworkflow

.github/actions/nf-test-action/action.yml

@@ -13,6 +13,14 @@ inputs:
   paths:
     description: "Test paths"
     required: true
+  jfrog_username:
+    description: "JFrog registry username"
+    required: false
+    default: ""
+  jfrog_password:
+    description: "JFrog registry password or token"
+    required: false
+    default: ""
 
 runs:
   using: "composite"
@@ -32,6 +40,14 @@ runs:
       with:
         python-version: "3.11"
 
+    - name: Install cross-repo component dependencies
+      shell: bash
+      run: |
+        mapfile -t metas < <(find subworkflows -name meta.yml)
+        if (( ${#metas[@]} > 0 )); then
+          bash .github/scripts/install-components.sh "${metas[@]}"
+        fi
+
     - name: Install nf-test
       uses: nf-core/setup-nf-test@v1
       with:
@@ -72,6 +88,34 @@ runs:
         nextflow secrets set ONCOKB_TOKEN $ONCOKB_TOKEN
 
 
+    - name: Login to JFrog Container Registry (docker)
+      if: ${{ inputs.profile == 'docker' && inputs.jfrog_username != '' }}
+      uses: docker/login-action@9780b0c442fbb1117ed29e0efdff1e18412f7567 # v3
+      with:
+        registry: mskcc.jfrog.io
+        username: ${{ inputs.jfrog_username }}
+        password: ${{ inputs.jfrog_password }}
+
+    - name: Configure JFrog auth for Apptainer/Singularity
+      if: ${{ contains(inputs.profile, 'singularity') && inputs.jfrog_username != '' }}
+      shell: bash
+      env:
+        JFROG_USERNAME: ${{ inputs.jfrog_username }}
+        JFROG_PASSWORD: ${{ inputs.jfrog_password }}
+      run: |
+        # Scope JFrog creds to mskcc.jfrog.io only via a Docker-format auth
+        # file. Setting APPTAINER_DOCKER_USERNAME/PASSWORD globally would send
+        # JFrog basic-auth to every docker:// pull (incl. ghcr.io, quay.io),
+        # which the other registries reject with 403.
+        AUTH=$(printf '%s:%s' "$JFROG_USERNAME" "$JFROG_PASSWORD" | base64 -w0)
+        for d in "$HOME/.apptainer" "$HOME/.singularity"; do
+          mkdir -p "$d"
+          cat > "$d/docker-config.json" <<EOF
+        {"auths":{"mskcc.jfrog.io":{"auth":"$AUTH"}}}
+        EOF
+          chmod 600 "$d/docker-config.json"
+        done
+
       # TODO Skip failing conda tests and document their failures
       # https://github.com/nf-core/modules/issues/7017
     - name: Run nf-test

.github/scripts/install-components.sh

@@ -0,0 +1,73 @@
+#!/usr/bin/env bash
+# Install cross-repo nf-core/MSK component dependencies declared in a
+# subworkflow's meta.yml.
+#
+# Usage: install-components.sh META_YML [META_YML …]
+#
+# Each META_YML may declare a `components:` list. Bare-string entries
+# (e.g. `- hlahd`) refer to components in this repo and are skipped.
+# Dict entries (e.g. `- {name: samtools/view, git_remote: …, org_path: …}`)
+# are sparse-checked-out into `modules/<org_path>/<name>/`.
+#
+# Notes:
+# - Only leaf components are installed; transitive `components:` in the
+#   fetched modules are not recursively resolved.
+# - An existing `modules/<org>/<name>/` directory is treated as a cache hit
+#   and not refreshed even if the declared ref has changed. CI always
+#   starts from a fresh checkout, so this is fine in practice; for local
+#   re-runs, delete the directory to force a refetch.
+# - Default ref is `master`. Set `git_sha:` or `branch:` per component to
+#   override.
+
+set -euo pipefail
+
+DEFAULT_REMOTE="https://github.com/nf-core/modules.git"
+DEFAULT_REF="master"
+
+# Allowlist for path components built from meta.yml input.
+NAME_RE='^[a-z0-9_]+(/[a-z0-9_]+)?$'
+ORG_RE='^[a-z0-9_-]+$'
+
+TMPROOT=$(mktemp -d)
+trap 'rm -rf "$TMPROOT"' EXIT
+
+declare -A seen
+
+for meta in "$@"; do
+  [[ -f "$meta" ]] || { echo "skip: $meta not found"; continue; }
+  n=$(yq '.components | length // 0' "$meta")
+  for i in $(seq 0 $((n - 1))); do
+    kind=$(yq ".components[$i] | tag" "$meta")
+    [[ "$kind" == "!!str" ]] && continue  # bare = local, skip
+
+    name=$(yq -r   ".components[$i].name"                            "$meta")
+    org=$(yq -r    ".components[$i].org_path         // \"nf-core\"" "$meta")
+    remote=$(yq -r ".components[$i].git_remote       // \"$DEFAULT_REMOTE\"" "$meta")
+    ref=$(yq -r    ".components[$i].git_sha // .components[$i].branch // \"$DEFAULT_REF\"" "$meta")
+
+    [[ "$name" =~ $NAME_RE ]] || { echo "ERROR: invalid component name: $name (in $meta)" >&2; exit 1; }
+    [[ "$org"  =~ $ORG_RE  ]] || { echo "ERROR: invalid org_path: $org (in $meta)" >&2; exit 1; }
+
+    key="$remote|$org|$name|$ref"
+    [[ -n "${seen[$key]:-}" ]] && continue
+    seen[$key]=1
+
+    dest="modules/$org/$name"
+    if [[ -d "$dest" ]]; then
+      echo "✓ $dest already present"
+      continue
+    fi
+
+    echo "→ fetching $org/$name from $remote@$ref"
+    tmp="$TMPROOT/$org-${name//\//_}-$ref"
+    mkdir -p "$tmp"
+    git -C "$tmp" init -q
+    git -C "$tmp" remote add origin "$remote"
+    git -C "$tmp" config core.sparseCheckout true
+    echo "modules/$org/$name/" > "$tmp/.git/info/sparse-checkout"
+    git -C "$tmp" fetch --depth 1 origin "$ref" -q
+    git -C "$tmp" checkout -q FETCH_HEAD
+    mkdir -p "$(dirname "$dest")"
+    mv "$tmp/modules/$org/$name" "$dest"
+  done
+done

.github/skip_nf_test.json

@@ -38,7 +38,9 @@
         "subworkflows/msk/phylowgs",
         "subworkflows/msk/generate_mutated_peptides",
         "subworkflows/msk/neoantigen_editing",
-        "subworkflows/msk/traceback"
+        "subworkflows/msk/traceback",
+        "modules/msk/hlahd",
+        "subworkflows/msk/hlahd_from_bam"
     ],
     "docker": [],
     "singularity": []

.github/workflows/nf-test.yml

@@ -135,6 +135,8 @@ jobs:
           shard: ${{ matrix.shard }}
           total_shards: ${{ env.TOTAL_SHARDS }}
           paths: "${{ join(fromJson(steps.filter.outputs.filtered_paths), ' ') }}"
+          jfrog_username: ${{ secrets.JFROG_USERNAME }}
+          jfrog_password: ${{ secrets.JFROG_PASSWORD }}
 
   confirm-pass:
     runs-on: ubuntu-latest

modules/msk/hlahd/environment.yml

@@ -0,0 +1,7 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/environment-schema.json
+channels:
+  - conda-forge
+  - bioconda
+dependencies:
+  - "YOUR-TOOL=HERE"

modules/msk/hlahd/main.nf

@@ -0,0 +1,64 @@
+process HLAHD {
+    tag "$meta.id"
+    label 'process_high'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'docker://mskcc.jfrog.io/omicswf-docker-prod-local/mskcc-omics-workflows/hlahd:1.7.1':
+        'mskcc.jfrog.io/omicswf-docker-prod-local/mskcc-omics-workflows/hlahd:1.7.1' }"
+
+    input:
+    tuple val(meta), path(fastq_1), path(fastq_2)
+
+    output:
+    tuple val(meta), path("${prefix}/result/${prefix}_final.result.txt"), emit: result
+    tuple val(meta), path("${prefix}/result/${prefix}_*.est.txt"),       emit: result_per_locus
+    path "versions.yml",                                           emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args        = task.ext.args  ?: ''
+    def min_read    = task.ext.args2 ?: '100'
+    prefix          = task.ext.prefix ?: "${meta.id}"
+    def install_dir = '/opt/hlahd/current'
+    """
+    if [[ \$(ulimit -n) -lt 1024 ]]; then ulimit -n 1024; fi
+
+    ln -sf /usr/bin/python3 ./python
+    export PATH=\$PWD:${install_dir}/bin:\$PATH
+
+    mkdir -p ${prefix}
+
+    bash ${install_dir}/bin/hlahd.sh \\
+        -t ${task.cpus} \\
+        -m ${min_read} \\
+        -f ${install_dir}/freq_data \\
+        ${args} \\
+        ${fastq_1} \\
+        ${fastq_2} \\
+        ${install_dir}/HLA_gene.split.3.50.0.txt \\
+        ${install_dir}/dictionary \\
+        ${prefix} \\
+        .
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        hlahd: \$(bash ${install_dir}/bin/hlahd.sh 2>&1 | grep -oP 'HLA-HD version \\K[0-9.]+' | head -1)
+    END_VERSIONS
+    """
+
+    stub:
+    prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    mkdir -p ${prefix}/result
+    touch ${prefix}/result/${prefix}_final.result.txt
+    touch ${prefix}/result/${prefix}_A.est.txt
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        hlahd: \$(bash /opt/hlahd/current/bin/hlahd.sh 2>&1 | grep -oP 'HLA-HD version \\K[0-9.]+' | head -1)
+    END_VERSIONS
+    """
+}

modules/msk/hlahd/meta.yml

@@ -0,0 +1,67 @@
+name: "hlahd"
+description: HLA typing from paired-end FASTQ reads using HLA-HD
+keywords:
+  - HLA
+  - immunology
+  - typing
+  - genomics
+tools:
+  - "hlahd":
+      description:
+        "HLA-HD (HLA typing from High-quality Dictionary) performs HLA typing
+        from paired-end FASTQ reads using bowtie2 alignment against HLA allele dictionaries."
+      homepage: "https://w3.genome.med.kyoto-u.ac.jp/HLA-HD/"
+      documentation: "https://w3.genome.med.kyoto-u.ac.jp/HLA-HD/"
+      licence:
+        - "ACADEMIC SOFTWARE LICENSE"
+      identifier: ""
+input:
+  - - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. `[ id:'sample1', single_end:false ]`
+    - fastq_1:
+        type: file
+        description: First read of paired-end FASTQ input
+        pattern: "*.{fastq,fastq.gz,fq,fq.gz}"
+        ontologies:
+          - edam: http://edamontology.org/format_1930
+    - fastq_2:
+        type: file
+        description: Second read of paired-end FASTQ input
+        pattern: "*.{fastq,fastq.gz,fq,fq.gz}"
+        ontologies:
+          - edam: http://edamontology.org/format_1930
+output:
+  result:
+    - - meta:
+          type: map
+          description: |
+            Groovy Map containing sample information
+            e.g. `[ id:'sample1', single_end:false ]`
+      - ${prefix}/result/${prefix}_final.result.txt:
+          type: file
+          description: Final HLA typing result file containing best-call alleles for all loci
+          pattern: "**/result/*_final.result.txt"
+  result_per_locus:
+    - - meta:
+          type: map
+          description: |
+            Groovy Map containing sample information
+            e.g. `[ id:'sample1', single_end:false ]`
+      - ${prefix}/result/${prefix}_*.est.txt:
+          type: file
+          description: Per-locus HLA estimation files (one file per HLA gene)
+          pattern: "**/result/*_*.est.txt"
+  versions:
+    - versions.yml:
+        type: file
+        description: File containing software versions
+        pattern: "versions.yml"
+        ontologies:
+          - edam: http://edamontology.org/format_3750
+authors:
+  - "@johnoooh"
+maintainers:
+  - "@johnoooh"

modules/msk/hlahd/tests/main.nf.test

@@ -0,0 +1,66 @@
+nextflow_process {
+
+    name "Test Process HLAHD"
+    script "../main.nf"
+    process "HLAHD"
+
+    tag "modules"
+    tag "modules_msk"
+    tag "hlahd"
+
+    test("hlahd - fastq pair - result txt") {
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test_sample', single_end:false ], // meta map
+                    file(params.test_data_mskcc['hlahd']['fastq_1'], checkIfExists: true),
+                    file(params.test_data_mskcc['hlahd']['fastq_2'], checkIfExists: true)
+                    ]
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(
+                    process.out.result.collect { meta, f ->
+                        [meta, path(f).readLines().findAll { it.startsWith("A\t") }]
+                    },
+                    process.out.versions
+                    ).match()
+                }
+            )
+        }
+
+    }
+
+    test("hlahd - fastq pair - stub") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id:'test_sample', single_end:false ], // meta map
+                    file(params.test_data_mskcc['hlahd']['fastq_1'], checkIfExists: true),
+                    file(params.test_data_mskcc['hlahd']['fastq_2'], checkIfExists: true)
+                    ]
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert path(process.out.result.get(0).get(1)).exists() },
+                { assert snapshot(process.out.versions).match() }
+            )
+        }
+
+    }
+
+}