added more tests

Author: notactuallyfinn

src/hermes/commands/harvest/codemeta.py

@@ -8,15 +8,16 @@
 import glob
 import json
 import pathlib
-import typing as t
+from typing import Union
 
 from hermes.commands.harvest.base import HermesHarvestCommand, HermesHarvestPlugin
 from hermes.commands.harvest.util.validate_codemeta import validate_codemeta
-from hermes.model.errors import HermesValidationError
+from hermes.model.error import HermesValidationError
+from hermes.model import SoftwareMetadata
 
 
 class CodeMetaHarvestPlugin(HermesHarvestPlugin):
-    def __call__(self, command: HermesHarvestCommand) -> t.Tuple[t.Dict, t.Dict]:
+    def __call__(self, command: HermesHarvestCommand) -> tuple[SoftwareMetadata, dict]:
         """
         Implementation of a harvester that provides data from a codemeta.json file format.
 
@@ -39,7 +40,7 @@ def __call__(self, command: HermesHarvestCommand) -> t.Tuple[t.Dict, t.Dict]:
             raise HermesValidationError(codemeta_file)
 
         codemeta = json.loads(codemeta_str)
-        return codemeta, {'local_path': str(codemeta_file)}
+        return SoftwareMetadata(codemeta), {'local_path': str(codemeta_file)}
 
     def _validate(self, codemeta_file: pathlib.Path) -> bool:
         with open(codemeta_file, "r") as fi:
@@ -55,7 +56,7 @@ def _validate(self, codemeta_file: pathlib.Path) -> bool:
 
         return True
 
-    def _get_single_codemeta(self, path: pathlib.Path) -> t.Optional[pathlib.Path]:
+    def _get_single_codemeta(self, path: pathlib.Path) -> Union[pathlib.Path, None]:
         # Find CodeMeta files in directories and subdirectories
         # TODO: Do we really want to search recursive? Maybe add another option to enable pointing to a single file?
         #       (So this stays "convention over configuration")

test/hermes_test/model/test_api_e2e.py

@@ -1,5 +1,12 @@
+# SPDX-FileCopyrightText: 2026 German Aerospace Center (DLR)
+#
+# SPDX-License-Identifier: Apache-2.0
+
+# SPDX-FileContributor: Michael Fritzsche
+
 import pytest
 from hermes.commands.harvest.cff import CffHarvestPlugin, CffHarvestSettings
+from hermes.commands.harvest.codemeta import CodeMetaHarvestPlugin
 from hermes.model import SoftwareMetadata
 
 
@@ -18,15 +25,88 @@
     email: max@muster.mann""",
             SoftwareMetadata({
                 "@type": "SoftwareSourceCode",
-                "schema:author": {
+                "author": {
                     "@list": [{
                         "@type": "Person",
                         "email": ["max@muster.mann"],
                         "familyName": ["Mustermann"],
                         "givenName": ["Max"]
                     }]
                 },
-                "schema:name": ["Temp"]
+                "name": ["Temp"]
+            })
+        ),
+        (
+            """# SPDX-FileCopyrightText: 2022 German Aerospace Center (DLR), Helmholtz-Zentrum Dresden-Rossendorf
+#
+# SPDX-License-Identifier: CC0-1.0
+
+# SPDX-FileContributor: Michael Meinel
+
+cff-version: 1.2.0
+title: hermes
+message: >-
+  If you use this software, please cite it using the
+  metadata from this file.
+version: 0.9.0
+license: "Apache-2.0"
+abstract: "Tool to automate software publication. Not stable yet."
+type: software
+authors:
+  - given-names: Michael
+    family-names: Meinel
+    email: michael.meinel@dlr.de
+    affiliation: German Aerospace Center (DLR)
+    orcid: "https://orcid.org/0000-0001-6372-3853"
+  - given-names: Stephan
+    family-names: Druskat
+    email: stephan.druskat@dlr.de
+    affiliation: German Aerospace Center (DLR)
+    orcid: "https://orcid.org/0000-0003-4925-7248"
+identifiers:
+  - type: doi
+    value: 10.5281/zenodo.13221384
+    description: Version 0.8.1b1
+""",
+            SoftwareMetadata({
+                "@type": ["http://schema.org/SoftwareSourceCode"],
+                "http://schema.org/author": [
+                    {
+                        "@list": [
+                            {
+                                "@id": "https://orcid.org/0000-0001-6372-3853",
+                                "@type": ["http://schema.org/Person"],
+                                "http://schema.org/affiliation": [
+                                    {
+                                        "@type": ["http://schema.org/Organization"],
+                                        "http://schema.org/name": [{"@value": "German Aerospace Center (DLR)"}]
+                                    }
+                                ],
+                                "http://schema.org/email": [{"@value": "michael.meinel@dlr.de"}],
+                                "http://schema.org/familyName": [{"@value": "Meinel"}],
+                                "http://schema.org/givenName": [{"@value": "Michael"}]
+                            },
+                            {
+                                "@id": "https://orcid.org/0000-0003-4925-7248",
+                                "@type": ["http://schema.org/Person"],
+                                "http://schema.org/affiliation": [
+                                    {
+                                        "@type": ["http://schema.org/Organization"],
+                                        "http://schema.org/name": [{"@value": "German Aerospace Center (DLR)"}]
+                                    }
+                                ],
+                                "http://schema.org/email": [{"@value": "stephan.druskat@dlr.de"}],
+                                "http://schema.org/familyName": [{"@value": "Druskat"}],
+                                "http://schema.org/givenName": [{"@value": "Stephan"}]
+                            }
+                        ]
+                    }
+                ],
+                "http://schema.org/description": [{"@value": "Tool to automate software publication. Not stable yet."}],
+                "http://schema.org/identifier": [{"@id": "https://doi.org/10.5281/zenodo.13221384"}],
+                "http://schema.org/license": [{"@id": "https://spdx.org/licenses/Apache-2.0"}],
+                "http://schema.org/name": [{"@value": "hermes"}],
+                "http://schema.org/version": [{"@value": "0.9.0"}]
             })
         )
     ]
@@ -51,6 +131,42 @@ def __init__(self, args, settings):
     cff_file.write_text(cff)
 
     result = CffHarvestPlugin().__call__(command)
-    # FIXME: update to compare the SoftwareMetadata objects instead of the data_dicts
+    # FIXME: update to compare the SoftwareMetadata objects instead of the data_dicts (in multiple places)
     # after merge with refactor/data-model and/or refactor/423-implement-public-api
     assert result[0].data_dict == res.data_dict
+
+
+@pytest.mark.parametrize(
+    "codemeta, res",
+    [
+        (
+            """{
+    "@context": "https://doi.org/10.5063/schema/codemeta-2.0",
+    "type": "SoftwareSourceCode",
+    "description": "for testing",
+    "name": "Test"
+}""",
+            SoftwareMetadata({
+                "@type": ["http://schema.org/SoftwareSourceCode"],
+                "http://schema.org/description": [{"@value": "for testing"}],
+                "http://schema.org/name": [{"@value": "Test"}]
+            })
+        )
+    ]
+)
+def test_codemeta_harvest(tmp_path, codemeta, res):
+    class Args:
+        def __init__(self, path):
+            self.path = path
+
+    class Command:
+        def __init__(self, args):
+            self.args = args
+
+    command = Command(Args(tmp_path))
+
+    codemeta_file = tmp_path / "codemeta.json"
+    codemeta_file.write_text(codemeta)
+
+    result = CodeMetaHarvestPlugin().__call__(command)
+    assert result[0].data_dict == res.data_dict