add tool identifier from bio.tools and intput and output ontologies

nf_core/components/components_utils.py

@@ -4,6 +4,7 @@
 from typing import List, Optional, Tuple
 
 import questionary
+import requests
 import rich.prompt
 
 import nf_core.utils
@@ -152,3 +153,29 @@ def get_components_to_install(subworkflow_dir: str) -> Tuple[List[str], List[str
                 elif link.startswith("../"):
                     subworkflows.append(name.lower())
     return modules, subworkflows
+
+
+def get_biotools_id(tool_name) -> str:
+    """
+    Try to find a bio.tools ID for 'tool'
+    """
+    url = f"https://bio.tools/api/t/?q={tool_name}&format=json"
+    try:
+        # Send a GET request to the API
+        response = requests.get(url)
+        response.raise_for_status()  # Raise an error for bad status codes
+        # Parse the JSON response
+        data = response.json()
+
+        # Iterate through the tools in the response to find the tool name
+        for tool in data["list"]:
+            if tool["name"].lower() == tool_name:
+                return tool["biotoolsCURIE"]
+
+        # If the tool name was not found in the response
+        log.warning(f"Could not find a bio.tools ID for '{tool_name}'")
+        return ""
+
+    except requests.exceptions.RequestException as e:
+        log.warning(f"Could not find a bio.tools ID for '{tool_name}': {e}")
+        return ""

nf_core/components/create.py

@@ -20,6 +20,7 @@
 import nf_core
 import nf_core.utils
 from nf_core.components.components_command import ComponentCommand
+from nf_core.components.components_utils import get_biotools_id
 from nf_core.lint_utils import run_prettier_on_file
 
 log = logging.getLogger(__name__)
@@ -60,6 +61,7 @@ def __init__(
         self.file_paths: Dict[str, Path] = {}
         self.not_empty_template = not empty_template
         self.migrate_pytest = migrate_pytest
+        self.tool_identifier = ""
 
     def create(self):
         """
@@ -150,6 +152,8 @@ def create(self):
             if self.component_type == "modules":
                 # Try to find a bioconda package for 'component'
                 self._get_bioconda_tool()
+                # Try to find a biotools entry for 'component'
+                self.tool_identifier = get_biotools_id(self.component)
 
             # Prompt for GitHub username
             self._get_username()

nf_core/module-template/meta.yml

@@ -20,6 +20,7 @@ tools:
       tool_dev_url: "{{ tool_dev_url }}"
       doi: ""
       licence: {{ tool_licence }}
+      identifier: {{ tool_identifier }}
 
 {% if not_empty_template -%}
 ## TODO nf-core: Add a description of all of the variables used as input
@@ -39,6 +40,14 @@ input:
         type: file
         description: {{ 'Sorted BAM/CRAM/SAM file' if not_empty_template else "" }}
         pattern: {{ '"*.{bam,cram,sam}"' if not_empty_template else "" }}
+        ontologies:
+          {% if not_empty_template -%}
+          - edam: "http://edamontology.org/format_25722"
+          - edam: "http://edamontology.org/format_2573"
+          - edam: "http://edamontology.org/format_3462"
+          {% else %}
+          - edam: ""
+          {%- endif %}
 
 {% if not_empty_template -%}
 ## TODO nf-core: Add a description of all of the variables used as output
@@ -56,14 +65,22 @@ output:
         description: |
           Groovy Map containing sample information
           e.g. `[ id:'sample1', single_end:false ]`
-  {% endif %}
+  {%- endif %}
   {% if not_empty_template -%}
     ## TODO nf-core: Delete / customise this example output
   {%- endif %}
     - {{ '"*.bam":' if not_empty_template else '"*":' }}
         type: file
         description: {{ 'Sorted BAM/CRAM/SAM file' if not_empty_template else "" }}
         pattern: {{ '"*.{bam,cram,sam}"' if not_empty_template else "" }}
+        ontologies:
+          {% if not_empty_template -%}
+          - edam: "http://edamontology.org/format_25722"
+          - edam: "http://edamontology.org/format_2573"
+          - edam: "http://edamontology.org/format_3462"
+          {% else -%}
+          - edam: ""
+          {%- endif %}
 
 authors:
   - "{{ author }}"

nf_core/modules/lint/__init__.py

@@ -11,10 +11,11 @@
 
 import questionary
 import rich
-import yaml
+import ruamel.yaml
 
 import nf_core.modules.modules_utils
 import nf_core.utils
+from nf_core.components.components_utils import get_biotools_id
 from nf_core.components.lint import ComponentLint, LintExceptionError, LintResult
 from nf_core.lint_utils import console
 
@@ -253,6 +254,8 @@ def update_meta_yml_file(self, mod):
         """
         meta_yml = self.read_meta_yml(mod)
         corrected_meta_yml = meta_yml.copy()
+        yaml = ruamel.yaml.YAML()
+        yaml.preserve_quotes = True
 
         # Obtain inputs and outputs from main.nf and meta.yml
         # Used to compare only the structure of channels and elements
@@ -332,6 +335,14 @@ def update_meta_yml_file(self, mod):
                                                 )
                             break
 
+        # Add bio.tools identifier
+        for i, tool in enumerate(corrected_meta_yml["tools"]):
+            tool_name = list(tool.keys())[0]
+            if "identifier" not in tool[tool_name]:
+                corrected_meta_yml["tools"][i][tool_name]["identifier"] = get_biotools_id(
+                    mod.component_name if "/" not in mod.component_name else mod.component_name.split("/")[0]
+                )
+
         with open(mod.meta_yml, "w") as fh:
             log.info(f"Updating {mod.meta_yml}")
-            yaml.dump(corrected_meta_yml, fh, sort_keys=False, Dumper=nf_core.utils.custom_yaml_dumper())
+            yaml.dump(corrected_meta_yml, fh)

nf_core/modules/lint/meta_yml.py

@@ -3,7 +3,7 @@
 from pathlib import Path
 from typing import Union
 
-import yaml
+import ruamel.yaml
 from jsonschema import exceptions, validators
 
 from nf_core.components.lint import ComponentLint
@@ -173,6 +173,8 @@ def read_meta_yml(module_lint_object: ComponentLint, module: NFCoreComponent) ->
         dict: The `meta.yml` file as a dictionary
     """
     meta_yaml = None
+    yaml = ruamel.yaml.YAML()
+    yaml.preserve_quotes = True
     # Check if we have a patch file, get original file in that case
     if module.is_patched:
         lines = ModulesDiffer.try_apply_patch(
@@ -183,11 +185,11 @@ def read_meta_yml(module_lint_object: ComponentLint, module: NFCoreComponent) ->
             reverse=True,
         ).get("meta.yml")
         if lines is not None:
-            meta_yaml = yaml.safe_load("".join(lines))
+            meta_yaml = yaml.load("".join(lines))
     if meta_yaml is None:
         try:
             with open(module.meta_yml) as fh:
-                meta_yaml = yaml.safe_load(fh)
+                meta_yaml = yaml.load(fh)
         except FileNotFoundError:
             return None
     return meta_yaml