Use compact URIs for ForgeFed and ActivityStreams

It makes resulting documents (usually) shorter, and tests more readable.

swh/indexer/codemeta.py

@@ -9,6 +9,7 @@ import itertools
 import json
 import os.path
 import re
+from typing import Any, List
 
 from pyld import jsonld
 
@@ -120,11 +121,18 @@ def _document_loader(url, options=None):
         raise Exception(url)
 
 
-def compact(doc):
-    """Same as `pyld.jsonld.compact`, but in the context of CodeMeta."""
-    return jsonld.compact(
-        doc, CODEMETA_CONTEXT_URL, options={"documentLoader": _document_loader}
-    )
+def compact(doc, forgefed: bool):
+    """Same as `pyld.jsonld.compact`, but in the context of CodeMeta.
+
+    Args:
+        forgefed: Whether to add ForgeFed and ActivityStreams as compact URIs.
+          This is typically used for extrinsic metadata documents, which frequently
+          use properties from these namespaces.
+    """
+    contexts: List[Any] = [CODEMETA_CONTEXT_URL]
+    if forgefed:
+        contexts.append({"as": ACTIVITYSTREAMS_URI, "forge": FORGEFED_URI})
+    return jsonld.compact(doc, contexts, options={"documentLoader": _document_loader})
 
 
 def expand(doc):
@@ -202,4 +210,7 @@ def merge_documents(documents):
                     elif value not in merged_document[key]:
                         merged_document[key].append(value)
 
-    return compact(merged_document)
+    # XXX: we should set forgefed=True when merging extrinsic-metadata documents.
+    # however, this function is only used to merge multiple files of the same
+    # directory (which is only for intrinsic-metadata), so it is not an issue for now
+    return compact(merged_document, forgefed=False)

swh/indexer/metadata_dictionary/base.py

@@ -62,7 +62,7 @@ class BaseMapping:
         raise NotImplementedError(f"{self.__class__.__name__}.translate")
 
     def normalize_translation(self, metadata: Dict[str, Any]) -> Dict[str, Any]:
-        return compact(metadata)
+        raise NotImplementedError(f"{self.__class__.__name__}.normalize_translation")
 
 
 class BaseExtrinsicMapping(BaseMapping):
@@ -82,6 +82,9 @@ class BaseExtrinsicMapping(BaseMapping):
         """
         raise NotImplementedError(f"{cls.__name__}.extrinsic_metadata_formats")
 
+    def normalize_translation(self, metadata: Dict[str, Any]) -> Dict[str, Any]:
+        return compact(metadata, forgefed=True)
+
 
 class BaseIntrinsicMapping(BaseMapping):
     """Base class for intrinsic-metadata mappings to inherit from
@@ -99,6 +102,9 @@ class BaseIntrinsicMapping(BaseMapping):
         """
         raise NotImplementedError(f"{cls.__name__}.detect_metadata_files")
 
+    def normalize_translation(self, metadata: Dict[str, Any]) -> Dict[str, Any]:
+        return compact(metadata, forgefed=False)
+
 
 class SingleFileIntrinsicMapping(BaseIntrinsicMapping):
     """Base class for all intrinsic metadata mappings that use a single file as input."""

swh/indexer/tests/metadata_dictionary/test_github.py

@@ -5,6 +5,14 @@
 
 from swh.indexer.metadata_dictionary import MAPPINGS
 
+CONTEXT = [
+    "https://doi.org/10.5063/schema/codemeta-2.0",
+    {
+        "as": "https://www.w3.org/ns/activitystreams#",
+        "forge": "https://forgefed.org/ns#",
+    },
+]
+
 
 def test_compute_metadata_none():
     """
@@ -111,11 +119,11 @@ def test_compute_metadata_github():
     """
     result = MAPPINGS["GitHubMapping"]().translate(content)
     assert result == {
-        "@context": "https://doi.org/10.5063/schema/codemeta-2.0",
+        "@context": CONTEXT,
         "type": "https://forgefed.org/ns#Repository",
-        "https://forgefed.org/ns#forks": {
-            "https://www.w3.org/ns/activitystreams#totalItems": 1,
-            "type": "https://www.w3.org/ns/activitystreams#OrderedCollection",
+        "forge:forks": {
+            "as:totalItems": 1,
+            "type": "as:OrderedCollection",
         },
         "license": "https://spdx.org/licenses/GPL-3.0",
         "name": "SoftwareHeritage/swh-indexer",