Metaspace submit

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

@@ -71,6 +71,7 @@ runs:
       shell: bash
       env:
         SENTIEON_LICSRVR_IP: ${{ env.SENTIEON_LICSRVR_IP }}
+        METASPACE_API_KEY: ${{ env.METASPACE_API_KEY }}
         SENTIEON_AUTH_MECH: "GitHub Actions - token"
       run: |
         NFT_WORKDIR=~ \

.github/workflows/nf-test.yml

@@ -166,6 +166,7 @@ jobs:
         if: ${{steps.filter.outputs.filtered_paths != '[]'}}
         uses: ./.github/actions/nf-test-action
         env:
+          METASPACE_API_KEY: ${{ secrets.METASPACE_API_KEY }}
           SENTIEON_ENCRYPTION_KEY: ${{ secrets.SENTIEON_ENCRYPTION_KEY }}
           SENTIEON_LICENSE_MESSAGE: ${{ secrets.SENTIEON_LICENSE_MESSAGE }}
           SENTIEON_LICSRVR_IP: ${{ secrets.SENTIEON_LICSRVR_IP }}

modules/nf-core/metaspace/submit/environment.yml

@@ -0,0 +1,8 @@
+channels:
+- conda-forge
+- bioconda
+dependencies:
+- bioconda::metaspace2020=2.0.9
+- conda-forge::python=3.11.11
+- conda-forge::pip=25.0.1
+- conda-forge::pyyaml=6.0.2

modules/nf-core/metaspace/submit/main.nf

@@ -0,0 +1,43 @@
+process METASPACE_SUBMIT {
+    label 'process_low'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://community-cr-prod.seqera.io/docker/registry/v2/blobs/sha256/c3/c317f9380b8b631acacad83ab362b2badb42e8782f6bfa03e5befe59f2382283/data':
+        'community.wave.seqera.io/library/python_pip_metaspace-converter:958b8906de66e072' }"
+
+    input:
+    path imzml
+    path ibd
+    path config
+
+    output:
+    path("ds_id.txt")   , emit: ds_id
+    path("versions.yml"), emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def metaspaceApi = secrets.METASPACE_API_KEY ?
+        "export API_KEY=\$(mktemp);echo -n \"${secrets.METASPACE_API_KEY}\" > \$API_KEY; " :
+        ""
+
+    """
+    $metaspaceApi
+    """
+    template 'submit.py'
+
+    stub:
+
+    """
+    touch ds_id.txt
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        python: \$(python --version | cut -d ' ' -f 2)
+        metaspace: \$(python3 -c 'import metaspace; print(metaspace.__version__)')
+        yaml: \$(python3 -c 'import yaml; print(yaml.__version__)')
+    END_VERSIONS
+    """
+}

modules/nf-core/metaspace/submit/meta.yml

@@ -0,0 +1,44 @@
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/meta-schema.json
+name: "metaspace_submit"
+description: Submit imaging mass spectrometry data to METASPACE for metabolite annotation
+keywords:
+- imaging
+- mass_spectrometry
+- metabolomics
+- annotation
+tools:
+- "metaspace2020":
+    description: "Python package providing programmatic access to the METASPACE platform"
+    homepage: "https://metaspace2020.readthedocs.io"
+    documentation: "https://metaspace2020.readthedocs.io"
+    tool_dev_url: "https://github.com/metaspace2020/metaspace/tree/master/metaspace/python-client"
+    licence: ["Apache-2.0 license"]
+    identifier: biotools:metaspace
+
+input:
+- - imzml:
+      type: file
+      description: Path to .imzML file
+- - ibd:
+      type: file
+      description: Path to .ibd file
+- - config:
+      type: file
+      description: Path to .yml file containing dataset configuration and metadata
+output:
+- ds_id:
+  - "ds_id.txt":
+      type: file
+      description: File containing METASPACE dataset ID
+      pattern: "ds_id.txt"
+
+- versions:
+  - "versions.yml":
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+
+authors:
+- "@bisho2122"
+maintainers:
+- "@bisho2122"

modules/nf-core/metaspace/submit/templates/submit.py

@@ -0,0 +1,91 @@
+#!/usr/bin/env python
+
+import yaml
+import metaspace
+import platform
+import os
+
+with open("${config}") as f:
+    config = yaml.safe_load(f)
+
+api_key_path = os.getenv('API_KEY')
+
+if api_key_path and os.path.isfile(api_key_path):
+    with open(api_key_path, "r") as f:
+        API_key = f.read()
+else:
+    API_key = api_key_path
+
+sm = metaspace.SMInstance(api_key=API_key)
+
+ds_name = config['Dataset_name']
+
+Annot_settings = config['Annotation_settings']
+
+metadata = {k: v for k, v in config.items() if k not in ['User_info','Dataset_name','Annotation_settings']}
+metadata = {'Data_Type': 'Imaging MS', **metadata}
+
+dbs = [tuple(item) for item in Annot_settings["Metabolite_database"]]
+
+if len(Annot_settings['Adducts']) == 0:
+    adduct_list = None
+else:
+    adduct_list = Annot_settings['Adducts']
+
+if len(Annot_settings['Chemical_modifications']) == 0:
+    chem_modif = None
+else:
+    chem_modif = Annot_settings['Chemical_modifications']
+
+scoring_model_id_map = {
+    'MSM': 2,
+    'METASPACE_ML_Animal': 3,
+    'METASPACE_ML_Plant': 4
+}
+
+scoring_model_id = scoring_model_id_map[Annot_settings['Analysis_version']]
+
+# Submit dataset
+
+ds_id = sm.submit_dataset(
+    imzml_fn = "${imzml}",
+    ibd_fn = "${ibd}",
+    name = ds_name,
+    metadata = metadata,
+    is_public = False,
+    databases = dbs,
+    adducts = adduct_list,
+    chem_mods = chem_modif,
+    scoring_model = scoring_model_id,
+    ppm = float(Annot_settings['ppm'])
+)
+
+with open("ds_id.txt", 'w') as f:
+    f.write(ds_id)
+
+# Versions
+versions = {"${task.process}": {"python": platform.python_version(),
+                                "metaspace": metaspace.__version__,
+                                "yaml": yaml.__version__}}
+
+def format_yaml_like(data: dict, indent: int = 0) -> str:
+    """Formats a dictionary to a YAML-like string.
+
+    Args:
+        data (dict): The dictionary to format.
+        indent (int): The current indentation level.
+
+    Returns:
+        str: A string formatted as YAML.
+    """
+    yaml_str = ""
+    for key, value in data.items():
+        spaces = "    " * indent
+        if isinstance(value, dict):
+            yaml_str += f"{spaces}{key}:\\n{format_yaml_like(value, indent + 1)}"
+        else:
+            yaml_str += f"{spaces}{key}: {value}\\n"
+    return yaml_str
+
+with open("versions.yml", "w") as f:
+    f.write(format_yaml_like(versions))

modules/nf-core/metaspace/submit/tests/main.nf.test

@@ -0,0 +1,52 @@
+nextflow_process {
+
+    name "Test Process METASPACE_SUBMIT"
+    script "../main.nf"
+    process "METASPACE_SUBMIT"
+    config "./nextflow.config"
+
+    tag "modules"
+    tag "modules_nfcore"
+    tag "metaspace"
+    tag "metaspace/submit"
+
+    test("metaspace_submit_test") {
+
+        when {
+            process {
+                """
+                input[0] = file(params.test_data_base_path + 'test_metaspace_submit/test.imzML', checkIfExists: true)
+                input[1] = file(params.test_data_base_path + 'test_metaspace_submit/test.ibd', checkIfExists: true)
+                input[2] = file(params.test_data_base_path + 'test_metaspace_submit/ds_config.yml')
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success }
+            )
+        }
+    }
+    test("metaspace_submit - stub") {
+        tag "versions_test"
+
+        options '-stub'
+        when {
+            process {
+                """
+                input[0] = file(params.test_data_base_path + 'test_metaspace_submit/test.imzML', checkIfExists: true)
+                input[1] = file(params.test_data_base_path + 'test_metaspace_submit/test.ibd', checkIfExists: true)
+                input[2] = file(params.test_data_base_path + 'test_metaspace_submit/ds_config.yml')
+                """
+            }
+        }
+        then {
+            assertAll(
+                { assert process.success },
+                { assert snapshot(process.out).match() }
+            )
+        }
+    }
+
+}

modules/nf-core/metaspace/submit/tests/main.nf.test.snap

@@ -0,0 +1,25 @@
+{
+    "metaspace_submit - stub": {
+        "content": [
+            {
+                "0": [
+                    "ds_id.txt:md5,d41d8cd98f00b204e9800998ecf8427e"
+                ],
+                "1": [
+                    "versions.yml:md5,db6244919f766cddc2cd20502db9403d"
+                ],
+                "ds_id": [
+                    "ds_id.txt:md5,d41d8cd98f00b204e9800998ecf8427e"
+                ],
+                "versions": [
+                    "versions.yml:md5,db6244919f766cddc2cd20502db9403d"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.9.2",
+            "nextflow": "24.10.5"
+        },
+        "timestamp": "2025-03-31T21:55:03.550137605"
+    }
+}

modules/nf-core/metaspace/submit/tests/nextflow.config

@@ -0,0 +1,11 @@
+env {
+    // NOTE This is how nf-core/metaspace/submit users will use METASPACE submission in real world use
+    API_KEY = "$METASPACE_API_KEY"
+
+    // NOTE This is how nf-core/metaspace/submit users will test out METASPACE submission with a user API key
+    // nextflow secrets set METASPACE_API_KEY "USER_API_KEY"
+}
+
+params {
+    test_data_base_path   = "https://raw.githubusercontent.com/nf-core/test-datasets/spatialmetabo/"
+}