Notebook modules

modules/jupyternotebook/functions.nf

@@ -0,0 +1,68 @@
+//
+//  Utility functions used in nf-core DSL2 module files
+//
+
+//
+// Extract name of software tool from process name using $task.process
+//
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

modules/jupyternotebook/main.nf

@@ -0,0 +1,87 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+include { dump_params_yml } from "./parametrize"
+
+params.options = [:]
+options        = initOptions(params.options)
+params.parametrize = true
+params.implicit_params = true
+params.meta_params = true
+
+process JUPYTERNOTEBOOK {
+    tag "$meta.id"
+    label 'process_low'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
+
+    //NB: You likely want to override this with a container containing all required
+    //dependencies for you analysis. The container at least needs to contain the
+    //yaml and rmarkdown R packages.
+    //TODO: what container to use as default image?
+    conda (params.enable_conda ? "ipykernel jupytext nbconvert papermill" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/YOUR-TOOL-HERE"
+    } else {
+        container "quay.io/biocontainers/YOUR-TOOL-HERE"
+    }
+
+    input:
+    tuple val(meta), path(notebook)
+    val(parameters)
+    path(input_files)
+
+    output:
+    tuple val(meta), path("*.html"), emit: report
+    path("artifacts/*"), emit: artifacts
+    path "*.version.txt"          , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+
+    // Dump parameters to yaml file.
+    // Using a yaml file over using the CLI params because
+    //  * no issue with escaping
+    //  * allows to pass nested maps instead of just single values
+    def params_cmd = ""
+    def render_cmd = ""
+    if (params.parametrize) {
+        nb_params = [:]
+        if (params.implicit_params) {
+            nb_params["cpus"] = task.cpus
+            nb_params["artifact_dir"] = "artifacts"
+            nb_params["input_dir"] = "./"
+        }
+        if (params.meta_params) {
+            nb_params["meta"] = meta
+        }
+        nb_params += parameters
+        params_cmd = dump_params_yml(nb_params)
+        render_cmd = "papermill -f .params.yml"
+    } else {
+        render_cmd = "papermill"
+    }
+
+    """
+    # Create output directory
+    mkdir artifacts
+
+    # Set parallelism for BLAS/MKL etc. to avoid over-booking of resources
+    export MKL_NUM_THREADS="${task.cpus}"
+    export OPENBLAS_NUM_THREADS="${task.cpus}"
+    export OMP_NUM_THREADS="${task.cpus}"
+    export NUMBA_NUM_THREADS="${task.cpus}"
+
+    # dump parameters to yaml
+    ${params_cmd}
+
+    # Convert notebook to ipynb using jupytext, execute using papermill, convert using nbconvert
+    jupytext --to notebook --output - --set-kernel - ${notebook}  \\
+        | ${render_cmd} \\
+        | jupyter nbconvert --stdin --to html --output ${notebook.baseName}.html
+
+    # TODO how to output versions of multiple tools?
+    echo \$(jupytext --version) > ${software}.version.txt
+    """
+}

modules/jupyternotebook/meta.yml

@@ -0,0 +1,47 @@
+name: jupyternotebook
+## TODO nf-core: Add a description of the module and list keywords
+description: write your description here
+keywords:
+  - sort
+tools:
+  - jupyternotebook:
+      ## TODO nf-core: Add a description and other details for the software below
+      description: 
+      homepage: 
+      documentation: 
+      tool_dev_url: 
+      doi: ""
+      licence: 
+
+## TODO nf-core: Add a description of all of the variables used as input
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  ## TODO nf-core: Delete / customise this example input
+  - bam:
+      type: file
+      description: BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+
+## TODO nf-core: Add a description of all of the variables used as output
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  ## TODO nf-core: Delete / customise this example output
+  - bam:
+      type: file
+      description: Sorted BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+
+authors:
+  - "@grst"

modules/jupyternotebook/parametrize.nf

@@ -0,0 +1 @@
+../rmarkdown/parametrize.nf

modules/rmarkdown/functions.nf

@@ -0,0 +1,68 @@
+//
+//  Utility functions used in nf-core DSL2 module files
+//
+
+//
+// Extract name of software tool from process name using $task.process
+//
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

modules/rmarkdown/main.nf

@@ -0,0 +1,91 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+include { dump_params_yml } from "./parametrize"
+
+params.options = [:]
+options        = initOptions(params.options)
+params.parametrize = true
+params.implicit_params = true
+params.meta_params = true
+
+process RMARKDOWN {
+    tag "$meta.id"
+    label 'process_low'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
+
+    //NB: You likely want to override this with a container containing all required
+    //dependencies for you analysis. The container at least needs to contain the
+    //yaml and rmarkdown R packages.
+    conda (params.enable_conda ? "r-tidyverse=1.2.1" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/r-tidyverse:1.2.1"
+    } else {
+        container "quay.io/biocontainers/r-tidyverse:1.2.1"
+    }
+
+    input:
+    tuple val(meta), path(notebook)
+    val(parameters)
+    path(input_files)
+
+    output:
+    tuple val(meta), path("*.html"), emit: report
+    path("artifacts/*"), emit: artifacts, optional: true
+    path "*.version.txt", emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+
+    // Dump parameters to yaml file.
+    // Using a yaml file over using the CLI params because
+    //  * no issue with escaping
+    //  * allows to pass nested maps instead of just single values
+    def params_cmd = ""
+    def render_cmd = ""
+    if (params.parametrize) {
+        nb_params = [:]
+        if (params.implicit_params) {
+            nb_params["cpus"] = task.cpus
+            nb_params["artifact_dir"] = "artifacts"
+            nb_params["input_dir"] = "./"
+        }
+        if (params.meta_params) {
+            nb_params["meta"] = meta
+        }
+        nb_params += parameters
+        params_cmd = dump_params_yml(nb_params)
+        render_cmd = (
+            "params = yaml::read_yaml('.params.yml')\n" +
+            "rmarkdown::render('${notebook}', params=params, envir=new.env())"
+        )
+    } else {
+        render_cmd = "rmarkdown::render('${notebook}')"
+    }
+
+    """
+    # Create output directory
+    mkdir artifacts
+
+    # Set parallelism for BLAS/MKL etc. to avoid over-booking of resources
+    export MKL_NUM_THREADS="${task.cpus}"
+    export OPENBLAS_NUM_THREADS="${task.cpus}"
+    export OMP_NUM_THREADS="${task.cpus}"
+
+    # dump parameters to yaml
+    ${params_cmd}
+
+    # work around  https://github.com/rstudio/rmarkdown/issues/1508
+    mv "${notebook}" "${notebook}.orig"
+    cp -L "${notebook}.orig" "${notebook}"
+
+    # Render notebook
+    Rscript - <<EOF
+    ${render_cmd}
+    \nEOF
+
+    echo \$(Rscript -e "cat(paste(packageVersion('rmarkdown'), collapse='.'))") > ${software}.version.txt
+    """
+}