Merge pull request #2190 from mashehu/add-yaml-schema

add schema based validation for meta.ymls to linting

Author: mashehu

nf_core/module-template/modules/meta.yml

@@ -1,10 +1,14 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/yaml-schema.json
 name: "{{ component_name_underscore }}"
 {% if not_empty_template -%}
 ## TODO nf-core: Add a description of the module and list keywords
 {% endif -%}
 description: write your description here
 keywords:
   - sort
+  - example
+  - genomics
 tools:
   - "{{ component }}":
       {% if not_empty_template -%}
@@ -32,9 +36,9 @@ input:
   ## TODO nf-core: Delete / customise this example input
   {%- endif %}
   - {{ 'bam:' if not_empty_template else "input:" }}
-    type: file
-    description: {{ 'Sorted BAM/CRAM/SAM file' if not_empty_template else "" }}
-    pattern: {{ '"*.{bam,cram,sam}"' 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 "" }}
 
 {% if not_empty_template -%}
 ## TODO nf-core: Add a description of all of the variables used as output
@@ -55,9 +59,9 @@ output:
   ## TODO nf-core: Delete / customise this example output
   {%- endif %}
   - {{ 'bam:' if not_empty_template else "output:" }}
-    type: file
-    description: {{ 'Sorted BAM/CRAM/SAM file' if not_empty_template else "" }}
-    pattern: {{ '"*.{bam,cram,sam}"' 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 "" }}
 
 authors:
   - "{{ author }}"

nf_core/modules/lint/meta_yml.py

@@ -1,5 +1,7 @@
+import json
 from pathlib import Path
 
+import jsonschema.validators
 import yaml
 
 from nf_core.modules.modules_differ import ModulesDiffer
@@ -10,17 +12,15 @@ def meta_yml(module_lint_object, module):
     Lint a ``meta.yml`` file
 
     The lint test checks that the module has
-    a ``meta.yml`` file and that it contains
-    the required keys: ``name``, input`` and
-    ``output``.
+    a ``meta.yml`` file and that it follows the
+    JSON schema defined in the ``modules/yaml-schema.json``
+    file in the nf-core/modules repository.
 
     In addition it checks that the module name
     and module input is consistent between the
     ``meta.yml`` and the ``main.nf``.
 
     """
-    required_keys = ["name", "output"]
-    required_keys_lists = ["input", "output"]
     # Check if we have a patch file, get original file in that case
     meta_yaml = None
     if module.is_patched:
@@ -42,21 +42,31 @@ def meta_yml(module_lint_object, module):
             module.failed.append(("meta_yml_exists", "Module `meta.yml` does not exist", module.meta_yml))
             return
 
-    # Confirm that all required keys are given
-    contains_required_keys = True
-    all_list_children = True
-    for rk in required_keys:
-        if rk not in meta_yaml.keys():
-            module.failed.append(("meta_required_keys", f"`{rk}` not specified in YAML", module.meta_yml))
-            contains_required_keys = False
-        elif rk in meta_yaml.keys() and not isinstance(meta_yaml[rk], list) and rk in required_keys_lists:
-            module.failed.append(("meta_required_keys", f"`{rk}` is not a list", module.meta_yml))
-            all_list_children = False
-    if contains_required_keys:
-        module.passed.append(("meta_required_keys", "`meta.yml` contains all required keys", module.meta_yml))
+    # Confirm that the meta.yml file is valid according to the JSON schema
+    valid_meta_yml = True
+    try:
+        with open(Path(module_lint_object.modules_repo.local_repo_dir, "modules/yaml-schema.json"), "r") as fh:
+            schema = json.load(fh)
+        jsonschema.validators.validate(instance=meta_yaml, schema=schema)
+        module.passed.append(("meta_yml_valid", "Module `meta.yml` is valid", module.meta_yml))
+    except jsonschema.exceptions.ValidationError as e:
+        valid_meta_yml = False
+        hint = ""
+        if len(e.path) > 0:
+            hint = f"\nCheck the entry for `{e.path[0]}`."
+        if e.message.startswith("None is not of type 'object'") and len(e.path) > 2:
+            hint = f"\nCheck that the child entries of {e.path[0]+'.'+e.path[2]} are indented correctly."
+        module.failed.append(
+            (
+                "meta_yml_valid",
+                f"The `meta.yml` of the module {module.module_name} is not valid: {e.message}.{hint}",
+                module.meta_yml,
+            )
+        )
+        return
 
     # Confirm that all input and output channels are specified
-    if contains_required_keys and all_list_children:
+    if valid_meta_yml:
         if "input" in meta_yaml:
             meta_input = [list(x.keys())[0] for x in meta_yaml["input"]]
             for input in module.inputs:

nf_core/pipeline-template/modules/nf-core/custom/dumpsoftwareversions/meta.yml

@@ -1,3 +1,4 @@
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/subworkflows/yaml-schema.json
 name: "{{ subworkflow_name }}"
 ## TODO nf-core: Add a description of the subworkflow and list keywords
 description: Sort SAM/BAM/CRAM file

nf_core/subworkflow-template/subworkflows/meta.yml

@@ -43,7 +43,7 @@ def test_modules_lint_empty(self):
 
 
 def test_modules_lint_new_modules(self):
-    """lint all modules in nf-core/modules repo clone"""
+    """lint a new module"""
     module_lint = nf_core.modules.ModuleLint(dir=self.nfcore_modules)
     module_lint.lint(print_results=True, all_modules=True)
     assert len(module_lint.failed) == 0, f"Linting failed with {[x.__dict__ for x in module_lint.failed]}"

tests/modules/lint.py

@@ -18,7 +18,7 @@
 """
 
 ORG_SHA = "002623ccc88a3b0cb302c7d8f13792a95354d9f2"
-CORRECT_SHA = "63fd3cdb1be733041db74c15542a7b5b8f4095ed"
+CORRECT_SHA = "0245a9277d51a47c8aa68d264d294cf45312fab8"
 SUCCEED_SHA = "ba15c20c032c549d77c5773659f19c2927daf48e"
 FAIL_SHA = "67b642d4471c4005220a342cad3818d5ba2b5a73"
 BISMARK_ALIGN = "bismark/align"

tests/modules/patch.py

@@ -45,6 +45,17 @@ def create_modules_repo_dummy(tmp_dir):
         with requests_cache.disabled():
             module_create.create()
 
+    # Remove doi from meta.yml which makes lint fail
+    meta_yml = os.path.join(root_dir, "modules", "nf-core", "bpipe", "test", "meta.yml")
+    with open(meta_yml, "r") as fh:
+        lines = fh.readlines()
+    for line_index in range(len(lines)):
+        if "doi" in lines[line_index]:
+            to_pop = line_index
+    lines.pop(to_pop)
+    with open(meta_yml, "w") as fh:
+        fh.writelines(lines)
+
     return root_dir
 
 

tests/test_modules.py

@@ -80,8 +80,8 @@ def mock_anaconda_api_calls(rsps: responses.RequestsMock, module, version):
     anaconda_mock = {
         "latest_version": version.split("--")[0],
         "summary": "",
-        "doc_url": "",
-        "dev_url": "",
+        "doc_url": "http://test",
+        "dev_url": "http://test",
         "files": [{"version": version.split("--")[0]}],
         "license": "",
     }