Add STIMULUS/CHECKTORCHMODEL

modules/nf-core/stimulus/checktorchmodel/environment.yml

@@ -0,0 +1,9 @@
+channels:
+  - conda-forge
+  - bioconda
+
+dependencies:
+  - python=3.12
+  - pip
+  - pip:
+    - stimulus-py==0.2.4

modules/nf-core/stimulus/checktorchmodel/main.nf

@@ -0,0 +1,61 @@
+process STIMULUS_CHECKTORCHMODEL {
+    tag "$data_config - $data"
+    label 'process_medium'
+
+    container "docker.io/mathysgrapotte/stimulus-py:0.2.4.dev"
+
+    input:
+    path(data)
+    path(data_config)
+    path(model)
+    path(model_config)
+    path(initial_weights)
+
+    output:
+    path "*_modelcheck.log", emit: log
+    path "versions.yml"    , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args        = task.ext.args ?: ''
+    prefix          = task.ext.prefix ?: model.baseName.replaceFirst(/\.py/, "")
+    def weights_arg = initial_weights ? "--initial_weights ${initial_weights}" : ""
+    """
+    # initialize Ray
+    ray start --head --port=6379 --temp-dir /tmp/ray
+
+    # wait or it to start
+    sleep 10
+
+    # run the model check
+    stimulus-check-model \
+        -d ${data} \
+        -m ${model} \
+        -e ${data_config} \
+        -c ${model_config} \
+        --ray_results_dirpath ${workDir} \
+        $args > ${prefix}_modelcheck.log
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        Python: \$(python --version | cut -d ' ' -f 2)
+        Stimulus-py: \$(pip show stimulus-py | grep Version | cut -d ' ' -f 2)
+    END_VERSIONS
+    """
+
+    stub:
+    def args         = task.ext.args ?: ''
+    prefix           = task.ext.prefix ?: model.baseName.replaceFirst(/\.py/, "")
+    def STIMULUS_VER = '0.2.2' // container not used in stub, change manually
+    """
+    touch ${prefix}_modelcheck.log
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        Python: \$(python --version | cut -d ' ' -f 2)
+        Stimulus-py: ${STIMULUS_VER}
+    END_VERSIONS
+    """
+}

modules/nf-core/stimulus/checktorchmodel/meta.yml

@@ -0,0 +1,64 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/meta-schema.json
+name: "stimulus_checktorchmodel"
+description: Sanity check a Pytorch model using raytune
+keywords:
+  - machine learning
+  - neural network
+  - unit test
+tools:
+  - "stimulus":
+      description: "Stochastic Testing and Input Manipulation for Unbiased Learning Systems"
+      homepage: "https://github.com/mathysgrapotte/stimulus-py"
+      documentation: "https://github.com/mathysgrapotte/stimulus-py"
+      tool_dev_url: "https://github.com/mathysgrapotte/stimulus-py"
+      licence: ["MIT"]
+
+input:
+  - - data:
+        type: file
+        description: A CSV file with untransformed data
+        pattern: "*.csv"
+        ontologies:
+          - edam: "http://edamontology.org/format_3752"
+  - - data_config:
+        type: file
+        description: A YAML file describing data transformations
+        pattern: "*.{yml,yaml}"
+        ontologies:
+          - edam: "http://edamontology.org/format_3464"
+  - - model:
+        type: file
+        description: Python source file containing the model
+        pattern: "*.py"
+        ontologies:
+          - edam: "http://edamontology.org/format_3996"
+  - - model_config:
+        type: file
+        description: YAML file containing RayTune tuning parameters
+        pattern: "*.{yml,yaml}"
+        ontologies:
+          - edam: "http://edamontology.org/format_3750"
+  - - initial_weights:
+        type: file
+        description: An optional SafeTensors file containing initial weights
+        pattern: "*.safetensors"
+
+output:
+  - log:
+      - "*_modelcheck.log":
+          type: file
+          description: Model check output
+          pattern: "*.log"
+  - versions:
+      - "versions.yml":
+          type: file
+          description: File containing software versions
+          pattern: "versions.yml"
+
+authors:
+  - "@mathysgrapotte"
+  - "@alessiovignoli"
+  - "@itrujnara"
+maintainers:
+  - "@itrujnara"

modules/nf-core/stimulus/checktorchmodel/tests/main.nf.test

@@ -0,0 +1,63 @@
+// nf-core modules test stimulus/checktorchmodel
+nextflow_process {
+
+    name "Test Process STIMULUS_CHECKTORCHMODEL"
+    script "../main.nf"
+    process "STIMULUS_CHECKTORCHMODEL"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "stimulus"
+    tag "stimulus/checktorchmodel"
+
+    test("basic - csv - py") {
+
+        when {
+            process {
+                """
+                input[0] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_stimulus_split.csv', checkIfExists: true)
+                input[1] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_sub_config_0.yaml', checkIfExists: true)
+                input[2] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_model.py', checkIfExists: true)
+                input[3] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_model_cpu.yaml', checkIfExists: true)
+                input[4] = []
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert file(process.out.log[0]).readLines().get(1).contains("Configuration for experiment") },
+                { assert snapshot(process.out.versions).match() }
+            )
+        }
+
+    }
+
+    // TODO nf-core: Change the test name preferably indicating the test-data and file-format used but keep the " - stub" suffix.
+    test("basic - csv - stub") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_stimulus_split.csv', checkIfExists: true)
+                input[1] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_sub_config_0.yaml', checkIfExists: true)
+                input[2] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_model.py', checkIfExists: true)
+                input[3] = file(params.modules_testdata_base_path + '../../deepmodeloptim/testdata/titanic/titanic_model_cpu.yaml', checkIfExists: true)
+                input[4] = []
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+
+    }
+
+}

modules/nf-core/stimulus/checktorchmodel/tests/main.nf.test.snap

@@ -0,0 +1,37 @@
+{
+    "basic - csv - py": {
+        "content": [
+            [
+                "versions.yml:md5,84d45b8dcf715e8f42c82339832c0f29"
+            ]
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.3"
+        },
+        "timestamp": "2025-02-11T16:24:18.958811"
+    },
+    "basic - csv - stub": {
+        "content": [
+            {
+                "0": [
+                    "titanic_model_modelcheck.log:md5,d41d8cd98f00b204e9800998ecf8427e"
+                ],
+                "1": [
+                    "versions.yml:md5,70fcb1abbe6e6a2f1dad708b98db1843"
+                ],
+                "log": [
+                    "titanic_model_modelcheck.log:md5,d41d8cd98f00b204e9800998ecf8427e"
+                ],
+                "versions": [
+                    "versions.yml:md5,70fcb1abbe6e6a2f1dad708b98db1843"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.3"
+        },
+        "timestamp": "2025-02-11T16:24:24.093869"
+    }
+}