Fix NuGet packages.lock.json parsing for Project and CentralTransitive types

AUTHORS.rst

@@ -49,6 +49,7 @@ The following organizations or individuals have contributed to ScanCode:
 - Li Ha @linexb
 - Mankaran Singh @MankaranSingh
 - Marc-Etienne Vargenau @vargenau
+- Marco Berger @marcoberger
 - Martin Petkov @MartinPetkov
 - Maximilian Huber @maxhbr
 - Michael Herzog @mjherzog

CHANGELOG.rst

@@ -16,6 +16,13 @@ with the licensedcode-data and licensedcode-index being published
 in two seperate wheels. Also adds linux/macos ARM support in
 release archives and pypi wheels.
 
+- Fix NuGet ``packages.lock.json`` parsing to support ``Project`` and
+  ``CentralTransitive`` dependency types. ``Project`` entries are skipped
+  because they are project references, while ``CentralTransitive`` entries
+  are treated as transitive package dependencies. This prevents parsing from
+  aborting for lockfiles generated by projects using project references and
+  Central Package Management.
+
 - Remove the licensedcode data and built license indexes from the
   main scancode-toolkit built wheel, and release them as
   seperate wheels which scancode-toolkit depends on.

src/packagedcode/nuget.py

@@ -218,28 +218,36 @@ def parse(cls, location, package_only=False):
             extra_data = dict(
                 target_framework=target_framework,
             )
+
             for package_name, package_info in packages.items():
-                dependencies = cls.get_dependencies(package_info=package_info, scope=target_framework)
-                resolved_package_mapping = dict(
-                datasource_id=cls.datasource_id,
-                type=cls.default_package_type,
-                primary_language=cls.default_primary_language,
-                name=package_name,
-                dependencies=[
-                    dep.to_dict() for dep in dependencies
-                ],
-                is_virtual=True,
-                version=package_info.get('resolved'),
-                )
-                resolved_package = models.PackageData.from_data(resolved_package_mapping)
                 package_type = package_info.get('type')
+
+                if package_type == "Project":
+                    continue
+
                 if package_type == "Direct":
                     is_direct = True
-                elif package_type == "Transitive":
+                elif package_type in {"Transitive", "CentralTransitive"}:
                     is_direct = False
                 else:
-                    raise Exception(f"Unknown package type: {package_type} for package {package_name} in {location}")
-
+                    raise Exception(
+                        f"Unknown package type: {package_type} "
+                        f"for package {package_name} in {location}"
+                    )
+
+                dependencies = cls.get_dependencies(package_info=package_info, scope=target_framework)
+                resolved_package_mapping = dict(
+                    datasource_id=cls.datasource_id,
+                    type=cls.default_package_type,
+                    primary_language=cls.default_primary_language,
+                    name=package_name,
+                    dependencies=[
+                        dep.to_dict() for dep in dependencies
+                    ],
+                    is_virtual=True,
+                    version=package_info.get('resolved'),
+                )
+                resolved_package = models.PackageData.from_data(resolved_package_mapping)
 
                 version = package_info.get('resolved')
                 requested = package_info.get('requested')
@@ -256,12 +264,12 @@ def parse(cls, location, package_only=False):
                     is_direct=is_direct,
                 )
                 top_dependencies.append(dependency.to_dict())
-        package_data = dict(
-            datasource_id=cls.datasource_id,
-            type=cls.default_package_type,
-            primary_language=cls.default_primary_language,
-            extra_data=extra_data,
-            dependencies=top_dependencies,
-        )
-        yield models.PackageData.from_data(package_data, package_only)
 
+            package_data = dict(
+                datasource_id=cls.datasource_id,
+                type=cls.default_package_type,
+                primary_language=cls.default_primary_language,
+                extra_data=extra_data,
+                dependencies=top_dependencies,
+            )
+            yield models.PackageData.from_data(package_data, package_only)

tests/packagedcode/data/nuget/packages-with-project-and-central-transitive.lock.json

@@ -0,0 +1,33 @@
+{
+  "version": 2,
+  "dependencies": {
+    "net8.0": {
+      "Direct.Package": {
+        "type": "Direct",
+        "requested": "[1.0.0, )",
+        "resolved": "1.0.0",
+        "contentHash": "direct-package-content-hash",
+        "dependencies": {
+          "Transitive.Package": "2.0.0"
+        }
+      },
+      "Transitive.Package": {
+        "type": "Transitive",
+        "resolved": "2.0.0",
+        "contentHash": "transitive-package-content-hash"
+      },
+      "CentralTransitive.Package": {
+        "type": "CentralTransitive",
+        "requested": "[3.0.0, )",
+        "resolved": "3.0.0",
+        "contentHash": "central-transitive-package-content-hash"
+      },
+      "Local.Project": {
+        "type": "Project",
+        "dependencies": {
+          "CentralTransitive.Package": "[3.0.0, )"
+        }
+      }
+    }
+  }
+}

tests/packagedcode/test_nuget.py

@@ -62,13 +62,45 @@ def test_parse_as_package_only(self):
         package = nuget.NugetNuspecHandler.parse(location=test_file, package_only=True)
         expected_loc = self.get_test_loc('nuget/Castle.Core.nuspec-package-only.json.expected')
         self.check_packages_data(package, expected_loc, regen=REGEN_TEST_FIXTURES, package_only=True)
-    
+
     def test_parse_nuget_package_lock_json(self):
         test_file = self.get_test_loc('nuget/packages.lock.json')
         package = nuget.NugetPackagesLockHandler.parse(location=test_file)
         expected_loc = self.get_test_loc('nuget/packages.lock.json.expected')
         self.check_packages_data(package, expected_loc, regen=REGEN_TEST_FIXTURES, package_only=True)
-
+
+    def test_parse_nuget_package_lock_json_with_project_and_central_transitive_types(self):
+        test_file = self.get_test_loc(
+            'nuget/packages-with-project-and-central-transitive.lock.json'
+        )
+
+        packages = list(
+            nuget.NugetPackagesLockHandler.parse(
+                location=test_file,
+                package_only=True,
+            )
+        )
+
+        assert len(packages) == 1
+
+        package = packages[0].to_dict()
+        dependencies = package['dependencies']
+        dependencies_by_purl = {
+            dependency['purl']: dependency
+            for dependency in dependencies
+        }
+
+        assert 'pkg:nuget/Local.Project@1.0.0' not in dependencies_by_purl
+
+        assert dependencies_by_purl['pkg:nuget/Direct.Package@1.0.0']['is_direct'] is True
+        assert dependencies_by_purl['pkg:nuget/Direct.Package@1.0.0']['extracted_requirement'] == '[1.0.0, )'
+
+        assert dependencies_by_purl['pkg:nuget/Transitive.Package@2.0.0']['is_direct'] is False
+        assert dependencies_by_purl['pkg:nuget/Transitive.Package@2.0.0']['extracted_requirement'] == '2.0.0'
+
+        assert dependencies_by_purl['pkg:nuget/CentralTransitive.Package@3.0.0']['is_direct'] is False
+        assert dependencies_by_purl['pkg:nuget/CentralTransitive.Package@3.0.0']['extracted_requirement'] == '[3.0.0, )'
+
     def test_package_lock_json_is_package_data_file(self):
         test_file = self.get_test_loc('nuget/packages.lock.json')
         assert nuget.NugetPackagesLockHandler.is_datafile(test_file)