Add publiccode.yml package handler\n\nImplements a new DatafileHandler to parse publiccode.yml files.\npubliccode.yml is a metadata standard for public sector open source\nsoftware. See https://github.com/publiccodeyml/publiccode.yml\n\nResolves #2851

Author: kumarasantosh

src/packagedcode/__init__.py

@@ -33,6 +33,7 @@
 from packagedcode import opam
 from packagedcode import phpcomposer
 from packagedcode import pubspec
+from packagedcode import publiccode
 from packagedcode import pypi
 from packagedcode import readme
 from packagedcode import rpm
@@ -77,6 +78,8 @@
     conda.CondaMetaYamlHandler,
     conda.CondaYamlHandler,
 
+    publiccode.PubliccodeYmlHandler,
+
     conan.ConanFileHandler,
     conan.ConanDataHandler,
 

src/packagedcode/publiccode.py

@@ -0,0 +1,133 @@
+#
+# Copyright (c) nexB Inc. and others. All rights reserved.
+# ScanCode is a trademark of nexB Inc.
+# SPDX-License-Identifier: Apache-2.0
+# See http://www.apache.org/licenses/LICENSE-2.0 for the license text.
+# See https://github.com/nexB/scancode-toolkit for support or download.
+# See https://aboutcode.org for more information about nexB OSS projects.
+#
+
+import logging
+import os
+
+import saneyaml
+
+from packagedcode import models
+
+"""
+Handle publiccode.yml metadata files.
+publiccode.yml is a metadata standard for public sector open source software.
+See https://github.com/publiccodeyml/publiccode.yml
+"""
+
+TRACE = os.environ.get('SCANCODE_DEBUG_PACKAGE', False)
+
+logger = logging.getLogger(__name__)
+
+
+class PubliccodeYmlHandler(models.DatafileHandler):
+    datasource_id = 'publiccode_yml'
+    path_patterns = ('*/publiccode.yml', '*/publiccode.yaml')
+    default_package_type = 'publiccode'
+    default_primary_language = None
+    description = 'publiccode.yml metadata file'
+    documentation_url = 'https://github.com/publiccodeyml/publiccode.yml'
+
+    @classmethod
+    def parse(cls, location, package_only=False):
+        with open(location, 'rb') as f:
+            data = saneyaml.load(f.read())
+
+        if not data or not isinstance(data, dict):
+            return
+
+        # Validate: a publiccode.yml must have 'publiccodeYmlVersion'
+        if 'publiccodeYmlVersion' not in data:
+            return
+
+        name = data.get('name')
+        version = data.get('softwareVersion')
+        vcs_url = data.get('url')
+        homepage_url = data.get('landingURL') or vcs_url
+
+        # License is under legal.license (SPDX expression)
+        legal = data.get('legal') or {}
+        declared_license = legal.get('license')
+        copyright_statement = legal.get('mainCopyrightOwner') or legal.get('repoOwner')
+
+        # Description: prefer English, fall back to first available language
+        description = _get_description(data)
+
+        # Keywords from categories
+        categories = data.get('categories') or []
+        keywords = ', '.join(categories) if categories else None
+
+        # Parties from maintenance.contacts
+        parties = []
+        maintenance = data.get('maintenance') or {}
+        for contact in maintenance.get('contacts') or []:
+            contact_name = contact.get('name')
+            contact_email = contact.get('email')
+            if contact_name or contact_email:
+                parties.append(
+                    models.Party(
+                        type=models.party_person,
+                        name=contact_name,
+                        email=contact_email,
+                        role='maintainer',
+                    )
+                )
+
+        # Extra data
+        extra_data = {}
+        schema_version = data.get('publiccodeYmlVersion')
+        if schema_version:
+            extra_data['publiccodeYmlVersion'] = schema_version
+        platforms = data.get('platforms')
+        if platforms:
+            extra_data['platforms'] = platforms
+        development_status = data.get('developmentStatus')
+        if development_status:
+            extra_data['developmentStatus'] = development_status
+        software_type = data.get('softwareType')
+        if software_type:
+            extra_data['softwareType'] = software_type
+
+        yield models.PackageData(
+            datasource_id=cls.datasource_id,
+            type=cls.default_package_type,
+            name=name,
+            version=version,
+            vcs_url=vcs_url,
+            homepage_url=homepage_url,
+            description=description,
+            declared_license_expression=declared_license,
+            copyright=copyright_statement,
+            keywords=keywords,
+            parties=parties,
+            extra_data=extra_data or None,
+        )
+
+
+def _get_description(data):
+    """
+    Extract the best available description from publiccode.yml's
+    multilingual 'description' block. Prefer English, fall back to
+    any available language. Returns longDescription, else shortDescription.
+    """
+    description_block = data.get('description') or {}
+    if not description_block:
+        return
+
+    lang_data = (
+        description_block.get('en')
+        or description_block.get('eng')
+        or next(iter(description_block.values()), None)
+    )
+    if not lang_data:
+        return
+
+    long_desc = lang_data.get('longDescription', '').strip()
+    short_desc = lang_data.get('shortDescription', '').strip()
+
+    return long_desc or short_desc or None

tests/packagedcode/data/publiccode/publiccode.yml

@@ -0,0 +1,47 @@
+publiccodeYmlVersion: "0.4"
+
+name: Medusa
+url: "https://example.com/italia/medusa.git"
+landingURL: "https://example.com/medusa"
+softwareVersion: "1.0.3"
+
+platforms:
+  - web
+  - linux
+
+categories:
+  - financial-reporting
+  - accounting
+
+developmentStatus: stable
+softwareType: "standalone/desktop"
+
+description:
+  en:
+    shortDescription: >
+      A short description of this software.
+    longDescription: >
+      A very long description of this software. It explains what it does,
+      who it is for, and why you might want to use it in a public
+      administration context.
+    features:
+      - Feature one
+      - Feature two
+
+legal:
+  license: AGPL-3.0-or-later
+  mainCopyrightOwner: City of Example
+  repoOwner: City of Example
+
+maintenance:
+  type: "contract"
+  contacts:
+    - name: Francesco Rossi
+      email: f.rossi@example.com
+      affiliation: City of Example
+
+localisation:
+  localisationReady: true
+  availableLanguages:
+    - en
+    - it

tests/packagedcode/test_publiccode.py

@@ -0,0 +1,46 @@
+#
+# Copyright (c) nexB Inc. and others. All rights reserved.
+# ScanCode is a trademark of nexB Inc.
+# SPDX-License-Identifier: Apache-2.0
+#
+
+import os
+import pytest
+
+from packagedcode.publiccode import PubliccodeYmlHandler
+
+TESTDATA_DIR = os.path.join(os.path.dirname(__file__), 'data', 'publiccode')
+
+
+def test_publiccode_yml_basic():
+    location = os.path.join(TESTDATA_DIR, 'publiccode.yml')
+    packages = list(PubliccodeYmlHandler.parse(location))
+    assert len(packages) == 1
+    pkg = packages[0]
+
+    assert pkg.name == 'Medusa'
+    assert pkg.version == '1.0.3'
+    assert pkg.vcs_url == 'https://example.com/italia/medusa.git'
+    assert pkg.homepage_url == 'https://example.com/medusa'
+    assert pkg.declared_license_expression == 'AGPL-3.0-or-later'
+    assert pkg.copyright == 'City of Example'
+    assert 'financial-reporting' in pkg.keywords
+    assert len(pkg.parties) == 1
+    assert pkg.parties[0].name == 'Francesco Rossi'
+    assert pkg.parties[0].email == 'f.rossi@example.com'
+    assert pkg.parties[0].role == 'maintainer'
+
+
+def test_publiccode_yml_no_version_key_returns_nothing(tmp_path):
+    """A YAML file without publiccodeYmlVersion should yield nothing."""
+    f = tmp_path / 'publiccode.yml'
+    f.write_text('name: something\nversion: 1.0\n')
+    packages = list(PubliccodeYmlHandler.parse(str(f)))
+    assert packages == []
+
+
+def test_publiccode_yml_path_patterns():
+    assert PubliccodeYmlHandler.path_patterns == (
+        '*/publiccode.yml',
+        '*/publiccode.yaml',
+    )