Tailfindr (#1904)

* adding tailfindr module

* replacing the container

* removing todo

* removing todo and fixing meta.yml

* Update modules/tailfindr/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update modules/tailfindr/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update modules/tailfindr/meta.yml

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update modules/tailfindr/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update tests/modules/tailfindr/test.yml

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update modules/tailfindr/meta.yml

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update modules/tailfindr/meta.yml

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update modules/tailfindr/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update modules/tailfindr/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update main.nf

* Update meta.yml

* Update modules/tailfindr/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update meta.yml

* Update main.nf

* adding pytest

* Update main.nf

* Update modules/tailfindr/main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

* Update test.yml

* Update main.nf

* Update main.nf

Co-authored-by: Jose Espinosa-Carrasco <kadomu@gmail.com>

modules/tailfindr/main.nf

@@ -0,0 +1,38 @@
+process TAILFINDR {
+    tag "$meta.id"
+    label 'process_high'
+
+    conda (params.enable_conda ? "bioconda::ont-fast5-api=0.4.1 bioconda::r-tailfindr=1.3" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/mulled-v2-f24f1399a77784f913670cbb36a0f17b78e0631b:80e40d512cd5a71665e3e00e8d0ad1462fc58f76-0':
+        'quay.io/biocontainers/mulled-v2-f24f1399a77784f913670cbb36a0f17b78e0631b:80e40d512cd5a71665e3e00e8d0ad1462fc58f76-0' }"
+
+    input:
+    tuple val(meta), path(fast5)
+
+    output:
+    tuple val(meta), path("*.csv.gz"), emit: csv_gz
+    path("versions.yml"), emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ? ", ${task.ext.args}": ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    """
+    R --vanilla --slave -e "library(tailfindr);
+    find_tails(fast5_dir = './' ,
+    save_dir = './' ${args},
+    csv_filename = \'${meta.id}.csv\',
+    num_cores = ${task.cpus})";
+    gzip ${meta.id}.csv
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        tailfindr: \$(Rscript -e "cat(paste(packageVersion('tailfindr'), collapse='.'))")
+        ont-fast5-api: \$(pip show ont-fast5-api | grep Version | awk '{print \$2}')
+    END_VERSIONS
+    """
+}

modules/tailfindr/meta.yml

@@ -0,0 +1,42 @@
+name: "tailfindr"
+description: Estimating poly(A)-tail lengths from basecalled fast5 files produced by Nanopore sequencing of RNA and DNA
+keywords:
+  - polyA
+  - fast5
+tools:
+  - "tailfindr":
+      description: "An R package for estimating poly(A)-tail lengths in Oxford Nanopore RNA and DNA reads."
+      homepage: "https://github.com/adnaniazi/tailfindr"
+      documentation: "https://github.com/adnaniazi/tailfindr/blob/master/README.md"
+      tool_dev_url: "https://github.com/adnaniazi/tailfindr"
+      doi: "10.1261/rna.071332.119"
+      licence: "['AGPL v3']"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test' ]
+  - fast5:
+      type: file
+      description: fast5 file
+      pattern: "*.fast5"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test' ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - csv_gz:
+      type: file
+      description: Compressed csv file
+      pattern: "*.csv.gz"
+
+authors:
+  - "@lucacozzuto"

tests/config/pytest_modules.yml

@@ -2147,6 +2147,10 @@ tabix/tabix:
   - modules/tabix/tabix/**
   - tests/modules/tabix/tabix/**
 
+tailfindr:
+  - modules/tailfindr/**
+  - tests/modules/tailfindr/**
+
 tbprofiler/profile:
   - modules/tbprofiler/profile/**
   - tests/modules/tbprofiler/profile/**

tests/modules/tailfindr/main.nf

@@ -0,0 +1,15 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { TAILFINDR } from '../../../modules/tailfindr/main.nf'
+
+workflow test_tailfindr {
+    
+    input = [
+        [ id: 'test' ], // meta map
+        file('https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/tailfindr/test.fast5', checkIfExists: true)
+    ]
+
+    TAILFINDR ( input )
+}

tests/modules/tailfindr/nextflow.config

@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+}

tests/modules/tailfindr/test.yml

@@ -0,0 +1,9 @@
+- name: "tailfindr"
+  command: nextflow run ./tests/modules/tailfindr -entry test_tailfindr -c ./tests/config/nextflow.config -c ./tests/modules/tailfindr/nextflow.config
+  tags:
+    - "tailfindr"
+  files:
+    - path: "output/tailfindr/test.csv.gz"
+      md5sum: 329e856d529dfd1ab31b0dedffc71f3c
+    - path: output/tailfindr/versions.yml
+      md5sum: 3a4e99bd95d9489da261b12a0ad740bb