From e18a2de9c090b4e0f01b6cea1c1432070cd18253 Mon Sep 17 00:00:00 2001
From: Zehra Hazal Sezer <65447556+zehrahazalsezer@users.noreply.github.com>
Date: Wed, 18 Oct 2023 17:04:42 +0200
Subject: [PATCH] htseq/count module is implemented (#4085)

* htseq/count module is implemented

* Update meta.yml

bam.bai changed to .bai

* Added meta2

* Implemented changes from review

* update pytest runner

* Update modules/nf-core/htseq/count/main.nf

Co-authored-by: Nicolas Vannieuwkerke <101190534+nvnieuwk@users.noreply.github.com>

---------

Co-authored-by: Maxime U Garcia <max.u.garcia@gmail.com>
Co-authored-by: Nicolas Vannieuwkerke <101190534+nvnieuwk@users.noreply.github.com>
Co-authored-by: mashehu <mashehu3@gmail.com>
---
 modules/nf-core/htseq/count/main.nf           | 49 +++++++++++++++++
 modules/nf-core/htseq/count/meta.yml          | 53 +++++++++++++++++++
 tests/config/pytest_modules.yml               |  4 ++
 tests/modules/nf-core/htseq/count/main.nf     | 40 ++++++++++++++
 .../nf-core/htseq/count/nextflow.config       |  6 +++
 tests/modules/nf-core/htseq/count/test.yml    | 19 +++++++
 6 files changed, 171 insertions(+)
 create mode 100644 modules/nf-core/htseq/count/main.nf
 create mode 100644 modules/nf-core/htseq/count/meta.yml
 create mode 100644 tests/modules/nf-core/htseq/count/main.nf
 create mode 100644 tests/modules/nf-core/htseq/count/nextflow.config
 create mode 100644 tests/modules/nf-core/htseq/count/test.yml

diff --git a/modules/nf-core/htseq/count/main.nf b/modules/nf-core/htseq/count/main.nf
new file mode 100644
index 00000000000..1fc6510b2a7
--- /dev/null
+++ b/modules/nf-core/htseq/count/main.nf
@@ -0,0 +1,49 @@
+process HTSEQ_COUNT {
+    tag "$meta.id"
+    label 'process_medium'
+
+    conda "bioconda::htseq=2.0.2"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/htseq:2.0.2--py310ha14a713_0':
+        'biocontainers/htseq:2.0.2--py310ha14a713_0' }"
+
+    input:
+    tuple val(meta), path(input), path(index)
+    tuple val(meta2), path(gtf)
+
+    output:
+    tuple val(meta), path("*.txt"), emit: txt
+    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}"
+
+    """
+    htseq-count \\
+        ${input} \\
+        ${gtf} \\
+        ${args} \\
+        > ${prefix}.txt
+
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        htseq: \$(echo \$(htseq-count --version ) | sed 's/^.*htseq-count //; s/Using.*\$//' ))
+    END_VERSIONS
+    """
+
+    stub:
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}.txt
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        : \$(echo \$(htseq-count --version ) | sed 's/^.*htseq-count //; s/Using.*\$//' ))
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/htseq/count/meta.yml b/modules/nf-core/htseq/count/meta.yml
new file mode 100644
index 00000000000..87b88d73fd1
--- /dev/null
+++ b/modules/nf-core/htseq/count/meta.yml
@@ -0,0 +1,53 @@
+---
+name: "htseq_count"
+description: count how many reads map to each feature
+keywords:
+  - htseq
+  - count
+  - gtf
+  - annotation
+tools:
+  - "htseq/count":
+      description: "HTSeq is a Python library to facilitate processing and analysis of data from high-throughput sequencing (HTS) experiments."
+      homepage: "https://htseq.readthedocs.io/en/latest/"
+      documentation: "https://htseq.readthedocs.io/en/latest/index.html"
+      doi: "10.1093/bioinformatics/btu638"
+      licence: "['GPL v3']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - meta2:
+      type: map
+      description: |
+        .gtf file information
+        e.g. `[ id:'test' ]`
+  - input:
+      type: file
+      description: Sorted BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+  - index:
+      type: file
+      description: Contains indexed bam file
+      pattern: "*.bai"
+  - gtf:
+      type: file
+      description: Contains the features in the GTF format
+      pattern: "*.gtf"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. `[ id:'test', single_end:false ]`
+  - txt:
+      type: file
+      description: File containing feature counts output
+      pattern: ".txt"
+
+authors:
+  - "@zehrahazalsezer"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 99384ebe424..5a1f7454cb6 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -1916,6 +1916,10 @@ hpsuissero:
   - modules/nf-core/hpsuissero/**
   - tests/modules/nf-core/hpsuissero/**
 
+htseq/count:
+  - modules/nf-core/htseq/count/**
+  - tests/modules/nf-core/htseq/count/**
+
 hypo:
   - modules/nf-core/hypo/**
   - tests/modules/nf-core/hypo/**
diff --git a/tests/modules/nf-core/htseq/count/main.nf b/tests/modules/nf-core/htseq/count/main.nf
new file mode 100644
index 00000000000..9b653cbce29
--- /dev/null
+++ b/tests/modules/nf-core/htseq/count/main.nf
@@ -0,0 +1,40 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { HTSEQ_COUNT } from '../../../../../modules/nf-core/htseq/count/main.nf'
+
+workflow test_htseq_count_bam {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_bam_bai'], checkIfExists: true)
+
+    ]
+    gtf = [
+        [ id:'test2'], // meta map
+        file(params.test_data['homo_sapiens']['genome']['genome_gtf'], checkIfExists: true)
+    ]
+
+
+    HTSEQ_COUNT (input,gtf)
+}
+
+workflow test_htseq_count_cram {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true)
+
+    ]
+    gtf = [
+        [ id:'test2'], // meta map
+        file(params.test_data['homo_sapiens']['genome']['genome_gtf'], checkIfExists: true)
+    ]
+
+
+    HTSEQ_COUNT (input,gtf)
+}
+
diff --git a/tests/modules/nf-core/htseq/count/nextflow.config b/tests/modules/nf-core/htseq/count/nextflow.config
new file mode 100644
index 00000000000..153c686dc53
--- /dev/null
+++ b/tests/modules/nf-core/htseq/count/nextflow.config
@@ -0,0 +1,6 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    docker.registry = 'quay.io'
+
+}
diff --git a/tests/modules/nf-core/htseq/count/test.yml b/tests/modules/nf-core/htseq/count/test.yml
new file mode 100644
index 00000000000..9d45ed672fb
--- /dev/null
+++ b/tests/modules/nf-core/htseq/count/test.yml
@@ -0,0 +1,19 @@
+- name: htseq count test_htseq_count_bam
+  command: nextflow run ./tests/modules/nf-core/htseq/count -entry test_htseq_count_bam -c ./tests/config/nextflow.config
+  tags:
+    - htseq
+    - htseq/count
+  files:
+    - path: output/htseq/test.txt
+      md5sum: 5e37b2292f1d21945acb532afa5eb615
+    - path: output/htseq/versions.yml
+
+- name: htseq count test_htseq_count_cram
+  command: nextflow run ./tests/modules/nf-core/htseq/count -entry test_htseq_count_cram -c ./tests/config/nextflow.config
+  tags:
+    - htseq
+    - htseq/count
+  files:
+    - path: output/htseq/test.txt
+      md5sum: 53f9b007bd72115e8b72a40f859bc896
+    - path: output/htseq/versions.yml