Merge branch 'no_libraries' into feature/SWS1-14

.github/workflows/unit-tests-srv.yml

@@ -51,11 +51,16 @@ jobs:
           env_vars: OS,PYTHON
           name: codecov-umbrella
           fail_ci_if_error: true
+      - name: Set working directory fir SonarCloud
+        run: mkdir -p $GITHUB_WORKSPACE/.scannerwork && chmod -R 777 $GITHUB_WORKSPACE/.scannerwork && pwd && ls -a
+
       - name: SonarCloud Scan
-        uses: SonarSource/sonarcloud-github-action@master
+        uses: SonarSource/sonarqube-scan-action@v4.1.0
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}  # Needed to get PR information, if any
           SONAR_TOKEN: ${{ secrets.SONAR_TOKEN }}
+        with:
+          args: -Dsonar.working.directory=$GITHUB_WORKSPACE/.scannerwork
 #      - name: Clean Mongo DB
 #        run: make clean-mongo-db
 #      - name: Upload coverage to Codecov

ted_sws/notice_packager/resources/templates/mets_xml_dmd_rdf.jinja2

@@ -12,6 +12,7 @@
     <cdm:work rdf:about="&resource;ted/{{ work.identifier }}">
         {# <rdf:type rdf:resource="http://publications.europa.eu/ontology/cdm#work"/> #}
         <rdf:type rdf:resource="http://publications.europa.eu/ontology/cdm#procurement_public"/>
+        <cdm:work_id rdf:datatype="http://www.w3.org/2001/XMLSchema#string">{{ work.uri }}</cdm:work_id>
         <cdm:work_id_document rdf:datatype="http://www.w3.org/2001/XMLSchema#string">ted:{{ work.identifier }}</cdm:work_id_document>
         <cdm:work_id_document rdf:datatype="http://www.w3.org/2001/XMLSchema#string">oj:{{ work.oj_identifier }}</cdm:work_id_document>
         <cdm:work_has_resource-type rdf:resource="http://publications.europa.eu/resource/authority/resource-type/PROCUREMENT_NOTICE"/>

tests/test_data/notice_packager/templates/2021_S_004_003545_0.mets.xml.dmd.rdf

@@ -12,6 +12,7 @@
     <cdm:work rdf:about="&resource;ted/2021_S_004_003545">
 
         <rdf:type rdf:resource="http://publications.europa.eu/ontology/cdm#procurement_public"/>
+        <cdm:work_id rdf:datatype="http://www.w3.org/2001/XMLSchema#string">http://data.europa.eu/a4g/resource/2021/003545_2021</cdm:work_id>
         <cdm:work_id_document rdf:datatype="http://www.w3.org/2001/XMLSchema#string">ted:2021_S_004_003545</cdm:work_id_document>
         <cdm:work_id_document rdf:datatype="http://www.w3.org/2001/XMLSchema#string">oj:JOS_2021_004_R_003545</cdm:work_id_document>
         <cdm:work_has_resource-type rdf:resource="http://publications.europa.eu/resource/authority/resource-type/PROCUREMENT_NOTICE"/>

tests/unit/dags/__init__.py

@@ -40,3 +40,4 @@
 
 FULL_BRANCH_TASK_IDS = [START_PROCESSING_NOTICE_TASK_ID, INDEX_NOTICE_XML_CONTENT_TASK_ID,
                         NORMALISE_NOTICE_METADATA_TASK_ID] + TRANSFORM_BRANCH_TASK_IDS + PACKAGE_BRANCH_TASK_IDS + PUBLISH_BRANCH_TASK_IDS
+

tests/unit/notice_packager/conftest.py

@@ -98,3 +98,9 @@ def invalid_rdf_files_path():
 @pytest.fixture
 def notice_id():
     return "196390_2018"
+
+
+@pytest.fixture
+def work_id_predicate():
+    """Returns the URI predicate for the CDM work identifier."""
+    return "http://publications.europa.eu/ontology/cdm#work_id"

tests/unit/notice_packager/test_template_generator.py

@@ -9,8 +9,10 @@
 import re
 
 import pytest
+from rdflib import Graph, Literal, XSD
 
 from ted_sws.notice_packager.adapters.template_generator import TemplateGenerator
+from ted_sws.notice_packager.model.metadata import PackagerMetadata
 from tests import TEST_DATA_PATH
 
 
@@ -57,3 +59,43 @@ def test_mets2action_mets_xml_generator_with_wrong_action(template_sample_metada
     template_sample_metadata.mets.type = "wrong_action"
     with pytest.raises(ValueError):
         TemplateGenerator.mets2action_mets_xml_generator(template_sample_metadata)
+
+
+def test_mets_dmd_rdf_has_work_id_after_generation(template_sample_metadata: PackagerMetadata,
+                                                   work_id_predicate: str):
+    """Test that generated METS DMD RDF contains a work_id predicate."""
+    mets_dmd_rdf: str = TemplateGenerator.mets_xml_dmd_rdf_generator(template_sample_metadata)
+    mets_graph: Graph = Graph().parse(data=mets_dmd_rdf, format="xml")
+
+    work_id_predicate_exists: bool = mets_graph.query(
+        f""" ASK WHERE {{ ?subject <{work_id_predicate}> ?object . }} """).askAnswer
+
+    assert work_id_predicate_exists
+
+
+def test_mets_dmd_rdf_has_work_id_as_string_after_generation(template_sample_metadata: PackagerMetadata,
+                                                             work_id_predicate: str):
+    """Test that work_id in METS DMD RDF is of type xsd:string."""
+    mets_dmd_rdf: str = TemplateGenerator.mets_xml_dmd_rdf_generator(template_sample_metadata)
+    mets_graph: Graph = Graph().parse(data=mets_dmd_rdf, format="xml")
+    string_datatype = XSD.string
+
+    work_id_predicate_exists: bool = mets_graph.query(
+        f""" ASK WHERE {{ ?subject <{work_id_predicate}> ?object . FILTER(datatype(?object) = <{string_datatype}>) }} """).askAnswer
+
+    assert work_id_predicate_exists
+
+
+def test_mets_dmd_rdf_has_correct_work_id_value_after_generation(template_sample_metadata: PackagerMetadata,
+                                                                 work_id_predicate: str):
+    """Test that work_id value in METS DMD RDF matches the metadata work URI."""
+    mets_dmd_rdf: str = TemplateGenerator.mets_xml_dmd_rdf_generator(template_sample_metadata)
+    mets_graph: Graph = Graph().parse(data=mets_dmd_rdf, format="xml")
+
+    assert template_sample_metadata.work.uri
+    work_id_value_literal = Literal(template_sample_metadata.work.uri, datatype=XSD.string)
+
+    work_id_is_same: bool = mets_graph.query(
+        f""" ASK WHERE {{ ?subject <{work_id_predicate}> {work_id_value_literal.n3()} . }} """).askAnswer
+
+    assert work_id_is_same