feat: give platform pages structured data

scripts/generate_site.py

@@ -1342,6 +1342,47 @@ def system_structured_data(slug: str, consoles: dict[str, list]) -> dict:
     return {"hasPart": parts} if parts else {}
 
 
+def platform_structured_data(name: str, coverage: dict) -> dict:
+    """A platform page's declarations for a machine reader.
+
+    The page printed thousands of hashes while its JSON-LD named only the
+    page: a crawler could not learn which md5 Batocera requires for a file.
+    The software and its version are the subject; each declared file, with
+    the hashes the platform checks, is a part. Past SYSTEM_PARTS_IN_FULL the
+    parts give way to the platform-files table, as system pages do.
+    """
+    config = coverage.get("config") or {}
+    about: dict = {"@type": "SoftwareApplication", "name": str(coverage.get("platform", name))}
+    if config.get("version"):
+        about["softwareVersion"] = str(config["version"])
+    if config.get("homepage"):
+        about["url"] = str(config["homepage"])
+    parts = []
+    for system_id, system in sorted((config.get("systems") or {}).items()):
+        for entry in system.get("files") or []:
+            if not isinstance(entry, dict) or not entry.get("name"):
+                continue
+            part: dict = {
+                "@type": "CreativeWork",
+                "name": str(entry.get("destination") or entry["name"]),
+                "url": urllib.parse.urljoin(SITE_URL, f"platforms/{name}/#{system_id}"),
+            }
+            identifiers = [
+                f"{field}:{value}"
+                for field in ("sha1", "md5", "crc32")
+                for value in _hash_values(entry.get(field))
+            ]
+            if identifiers:
+                part["identifier"] = identifiers
+            parts.append(part)
+    data: dict = {"about": about}
+    if len(parts) > SYSTEM_PARTS_IN_FULL:
+        data["isBasedOn"] = urllib.parse.urljoin(SITE_URL, "downloads/platform-files.csv")
+    elif parts:
+        data["hasPart"] = parts
+    return data
+
+
 def dataset_structured_data(exports: list[dict], db: dict) -> dict:
     """The data page as a schema.org Dataset a harvester can index.
 
@@ -4329,6 +4370,7 @@ def _generate(args: argparse.Namespace) -> None:
             str(docs / "platforms" / f"{name}.md"),
             generate_platform_page(name, cov, registry, emulator_files),
         )
+        PAGE_DATA[f"platforms/{name}.md"] = platform_structured_data(name, cov)
 
     # Generate system pages
     print("Generating system pages...")

tests/test_site_exports.py

@@ -477,5 +477,22 @@ class ArchiveMembersAreAcceptedHashes(unittest.TestCase):
         )
 
 
+
+class PlatformPagesCarryTheirDeclarations(unittest.TestCase):
+    """Platform pages were the only content pages without structured data:
+    5,640 md5 and 2,047 sha1 the platforms declare lived in the prose only."""
+
+    def test_parts_and_version(self):
+        from generate_site import platform_structured_data
+
+        data = platform_structured_data("batocera", {"platform": "Batocera", "config": {
+            "version": "43.1", "systems": {"sony-playstation": {"files": [
+                {"name": "scph5501.bin", "destination": "scph5501.bin", "md5": "490f666e"}]}}}})
+        self.assertEqual(data["about"]["softwareVersion"], "43.1")
+        part = data["hasPart"][0]
+        self.assertEqual(part["identifier"], ["md5:490f666e"])
+        self.assertTrue(part["url"].endswith("platforms/batocera/#sony-playstation"))
+
+
 if __name__ == "__main__":
     unittest.main()