fix: structured data names releases, not pins

The home page is the site: named RetroBIOS, part of nothing. A source repository states the core_version of its build when it names a release, never a commit id or a pin, and no placeholder description.

scripts/generate_site.py

@@ -2834,6 +2834,36 @@ def emulator_file_aliases(files: list[dict], anchors: list[str]) -> list[str]:
     return aliases
 
 
+# A commit id: hexadecimal with a letter and a digit, 7 to 40 long. A version
+# that holds one (0.8.139-2-g25e1ee2fd5) names no release.
+_COMMIT_ID = re.compile(
+    r"(?<![0-9a-f])(?=[0-9a-f]*[a-f])(?=[0-9a-f]*[0-9])[0-9a-f]{7,40}(?![0-9a-f])", re.I)
+
+
+def release_versions(profile: dict) -> dict[str, str]:
+    """The `core_version` of a profile by build mode ("" when one covers
+    every build), kept only where it names a release: a number and no
+    commit id. A commit id of digits only reads as a number, so a run that
+    starts one of the profile's own pins is one too (supermodel3's
+    "0.3a-1481545-mm")."""
+    pins = [str(value).casefold()
+            for field in ("source_commit", "upstream_commit")
+            for value in (profile[field].values() if isinstance(profile.get(field), dict)
+                          else [profile.get(field)])
+            if value]
+
+    def names_a_commit(value: str) -> bool:
+        return bool(_COMMIT_ID.search(value)) or any(
+            pin.startswith(run) for run in re.findall(r"[0-9a-f]{7,40}", value.casefold())
+            for pin in pins)
+
+    stated = profile.get("core_version")
+    pairs = stated.items() if isinstance(stated, dict) else [("", stated)]
+    versions = {str(mode): str(value or "").strip() for mode, value in pairs}
+    return {mode: value for mode, value in versions.items()
+            if re.search(r"[0-9]", value) and not names_a_commit(value)}
+
+
 def emulator_structured_data(profile: dict) -> dict:
     """What an emulator page states, for a reader of its JSON-LD.
 
@@ -2841,27 +2871,26 @@ def emulator_structured_data(profile: dict) -> dict:
     the core loads, their hashes, the source read or when.
     """
     data: dict = {}
-    # Every repository the profile reads, each with its own pin: xemu's
-    # libretro build lives in another repository at another commit, and a
-    # port's upstream_commit was never published.
+    # Every repository the profile reads. The source repository carries the
+    # release the profile describes, never a commit id: a pin is no version.
+    # The upstream of a port is another code base and states none.
+    versions = release_versions(profile)
     about = []
     seen: set[str] = set()
     for field in ("source", "upstream"):
         urls = profile.get(field)
-        pins = profile.get(f"{field}_commit")
         for mode, url in (urls.items() if isinstance(urls, dict) else [("", urls)]):
             if not url or not str(url).startswith("http") or str(url) in seen:
                 continue
             seen.add(str(url))
-            pin = pins.get(mode) if isinstance(pins, dict) else pins
             entry = {
                 "@type": "SoftwareSourceCode",
                 "name": str(profile.get("emulator", "")),
                 "codeRepository": str(url),
-                "description": " ".join(filter(None, [mode, field])),
             }
-            if pin:
-                entry["version"] = str(pin)
+            version = (versions.get(mode) or versions.get("")) if field == "source" else None
+            if version:
+                entry["version"] = version
             about.append(entry)
     if about:
         data["about"] = about[0] if len(about) == 1 else about
@@ -3083,6 +3112,10 @@ def decorate_markdown_pages(docs: Path) -> dict[str, str]:
             },
             **PAGE_DATA.get(relative.as_posix(), {}),
         }
+        if schema_type == "WebSite":
+            # The home page is the site: named as the site, part of nothing.
+            structured["name"] = SITE_NAME
+            del structured["isPartOf"]
         target, text = sitestarlight.finalize_page(
             relative.as_posix(), markdown, level=level,
             weight=search_weight(relative, level),

tests/test_audit_regressions.py

@@ -4360,31 +4360,44 @@ class EmulatorPagesStateTheirDataForMachines(unittest.TestCase):
 
         data = emulator_structured_data({
             "emulator": "DuckStation", "source": "https://github.com/stenzek/duckstation",
-            "source_commit": "4122fed9", "profiled_date": "2026-03-23",
+            "source_commit": "4122fed9", "core_version": "v0.1-12070",
+            "profiled_date": "2026-03-23",
             "files": [{"name": "scph5501.bin", "md5": "490F666E", "sha1": ["a" * 40]},
                       {"name": "free.bin"}],
         })
         self.assertEqual(data["about"]["codeRepository"], "https://github.com/stenzek/duckstation")
-        self.assertEqual(data["about"]["version"], "4122fed9")
+        # The release the profile describes, never the pin (round 5, 3.5).
+        self.assertEqual(data["about"]["version"], "v0.1-12070")
+        self.assertNotIn("description", data["about"])
         self.assertEqual(data["dateModified"], "2026-03-23")
         self.assertEqual(data["hasPart"][0]["identifier"], ["sha1:" + "a" * 40, "md5:490f666e"])
         self.assertNotIn("identifier", data["hasPart"][1])
 
-    def test_every_repository_with_its_own_pin(self):
-        """xemu's libretro build lives in another repository at another
-        commit, and a port's upstream_commit was never published."""
+    def test_every_repository_with_its_own_release(self):
+        """xemu's libretro build lives in another repository. Each source
+        repository states the release of its build, a version holding a
+        commit id states none, and a port's upstream is another code base."""
         from scripts.generate_site import emulator_structured_data
 
         data = emulator_structured_data({
-            "emulator": "xemu",
-            "source": {"standalone": "https://github.com/xemu-project/xemu",
-                       "libretro": "https://github.com/LibretroAdmin/xemu-libretro"},
+            "emulator": "MesenCE",
+            "source": {"standalone": "https://github.com/nesdev-org/MesenCE",
+                       "libretro": "https://github.com/libretro/MesenCE"},
             "source_commit": {"standalone": "f9b14039", "libretro": "d9804db7"},
-            "upstream": "https://github.com/mborgerson/xemu", "upstream_commit": "aa11",
+            "core_version": {"standalone": "2.2.1", "libretro": "2.0.0"},
+            "upstream": "https://github.com/SourMesen/Mesen2", "upstream_commit": "aa11",
         })
         self.assertEqual(
             [(a["codeRepository"].rsplit("/", 1)[1], a.get("version")) for a in data["about"]],
-            [("xemu", "f9b14039"), ("xemu-libretro", "d9804db7"), ("xemu", "aa11")])
+            [("MesenCE", "2.2.1"), ("MesenCE", "2.0.0"), ("Mesen2", None)])
+        xemu = emulator_structured_data({
+            "emulator": "xemu",
+            "source": {"standalone": "https://github.com/xemu-project/xemu",
+                       "libretro": "https://github.com/LibretroAdmin/xemu-libretro"},
+            "core_version": {"standalone": "0.8.139-2-g25e1ee2fd5",
+                             "libretro": "libretro-experimental"},
+        })
+        self.assertEqual([a.get("version") for a in xemu["about"]], [None, None])