Add origin extrinsic metadata endpoint

swh/web/api/tests/views/test_origin.py

@@ -930,6 +930,16 @@ def test_api_origin_intrinsic_metadata(api_client, origin):
     assert rv.data[ORIGIN_METADATA_KEY] == ORIGIN_METADATA_VALUE
 
 
+def test_api_origin_extrinsic_metadata(api_client, origin):
+    url = reverse(
+        "api-origin-extrinsic-metadata", query_params={"origin_url": origin["url"]}
+    )
+    rv = check_api_get_responses(api_client, url, status_code=200)
+
+    assert ORIGIN_METADATA_KEY in rv.data
+    assert rv.data[ORIGIN_METADATA_KEY] == ORIGIN_METADATA_VALUE
+
+
 def test_api_origin_metadata_search_invalid(api_client, mocker):
     mock_idx_storage = mocker.patch("swh.web.utils.archive.idx_storage")
     url = reverse("api-1-origin-metadata-search")

swh/web/api/views/origin.py

@@ -536,3 +536,40 @@ def api_origin_intrinsic_metadata(request: Request):
         enrich_fn=enrich_origin,
         request=request,
     )
+
+
+@api_route(r"/extrinsic-metadata/origin/", "api-origin-extrinsic-metadata")
+@api_doc("/extrinsic-metadata/origin/", category="Metadata")
+@format_docstring()
+def api_origin_extrinsic_metadata(request: Request):
+    """
+    .. http:get:: /api/1/origin/(origin_url)/extrinsic-metadata
+
+        Get extrinsic metadata of a software origin (as a JSON-LD/CodeMeta dictionary).
+
+        :query str origin_url: parameter for origin url
+
+        :>json string ???: extrinsic metadata field of the origin
+
+        {common_headers}
+
+        :statuscode 200: no error
+        :statuscode 404: requested origin cannot be found in the archive
+
+        **Example:**
+
+        .. parsed-literal::
+
+            :swh_web_api:`extrinsic-metadata/origin/origin_url=https://github.com/node-red/node-red-nodegen`
+    """
+    origin_url = request.GET.get("origin_url")
+    if origin_url is None:
+        raise BadInputExc("An origin URL must be provided as query parameter.")
+    return api_lookup(
+        archive.lookup_origin_extrinsic_metadata,
+        origin_url,
+        lookup_similar_urls=False,
+        notfound_msg=f"Origin with url {origin_url} not found",
+        enrich_fn=enrich_origin,
+        request=request,
+    )

swh/web/tests/data.py

@@ -16,7 +16,10 @@ from swh.counters import get_counters
 from swh.indexer.fossology_license import FossologyLicenseIndexer
 from swh.indexer.mimetype import MimetypeIndexer
 from swh.indexer.storage import get_indexer_storage
-from swh.indexer.storage.model import OriginIntrinsicMetadataRow
+from swh.indexer.storage.model import (
+    OriginExtrinsicMetadataRow,
+    OriginIntrinsicMetadataRow,
+)
 from swh.loader.git.from_disk import GitLoaderFromArchive
 from swh.model.hashutil import DEFAULT_ALGORITHMS, hash_to_bytes, hash_to_hex
 from swh.model.model import (
@@ -314,18 +317,30 @@ def _init_tests_data():
                     )
                 )
                 if b"master" in branch_name:
-                    # Add some origin intrinsic metadata for tests
+                    # Add some origin intrinsic and extrinsic metadata for tests
                     metadata = common_metadata
                     metadata.update(origin.get("metadata", {}))
                     revision = storage.revision_get([branch_data.target])[0]
-                    origin_metadata = OriginIntrinsicMetadataRow(
+                    origin_intrinsic_metadata = OriginIntrinsicMetadataRow(
                         id=origin["url"],
                         from_directory=revision.directory,
                         indexer_configuration_id=idx_tool["id"],
                         metadata=metadata,
                         mappings=[],
                     )
-                    idx_storage.origin_intrinsic_metadata_add([origin_metadata])
+                    idx_storage.origin_intrinsic_metadata_add(
+                        [origin_intrinsic_metadata]
+                    )
+                    origin_extrinsic_metadata = OriginExtrinsicMetadataRow(
+                        id=origin["url"],
+                        from_remd_id=revision.directory,
+                        indexer_configuration_id=idx_tool["id"],
+                        metadata=metadata,
+                        mappings=[],
+                    )
+                    idx_storage.origin_extrinsic_metadata_add(
+                        [origin_extrinsic_metadata]
+                    )
                     search.origin_update([{"url": origin["url"], "jsonld": metadata}])
 
                     ORIGIN_MASTER_REVISION[origin["url"]] = hash_to_hex(

swh/web/utils/archive.py

@@ -447,6 +447,38 @@ def lookup_origin_intrinsic_metadata(
     return result
 
 
+def lookup_origin_extrinsic_metadata(
+    origin_url: str, lookup_similar_urls: bool = True
+) -> Dict[str, Any]:
+    """Return extrinsic metadata for origin whose origin matches given
+    origin.
+
+    Args:
+        origin_url: origin url
+        lookup_similar_urls: if :const:`True`, lookup origin with and
+            without trailing slash in its URL
+
+    Raises:
+        NotFoundExc when the origin is not found
+
+    Returns:
+        origin metadata.
+
+    """
+    origins = [
+        lookup_origin(origin_url, lookup_similar_urls=lookup_similar_urls)["url"]
+    ]
+    origin_info = storage.origin_get(origins)[0]
+    if not origin_info:
+        raise NotFoundExc(f"Origin with url {origin_url} not found!")
+
+    match = _first_element(idx_storage.origin_extrinsic_metadata_get(origins))
+    result = {}
+    if match:
+        result = match.metadata
+    return result
+
+
 def _to_sha1_bin(sha1_hex):
     _, sha1_git_bin = query.parse_hash_with_algorithms_or_throws(
         sha1_hex, ["sha1"], "Only sha1_git is supported."  # HACK: sha1_git really