feat(site-publish): add split-surface publishing

Authored-By: OpenAI (GPT-5) <noreply@openai.com>

scripts/deploy.py

@@ -78,27 +78,51 @@ def _immutable_head(endpoint, bucket, key, aws_env):
     raise RuntimeError(f"head-object failed for s3://{bucket}/{key}: {error.strip()}")
 
 
-def _existing_immutable_keys(endpoint, bucket, aws_env):
+def _list_object_keys(endpoint, bucket, aws_env):
-    """Enumerate publisher-owned immutable keys so route moves cannot delete them."""
+    """List a bucket through the AWS CLI paginator in one client process."""
     args = ["aws", "--endpoint-url", endpoint, "s3api", "list-objects-v2",
             "--bucket", bucket, "--query", "Contents[].Key", "--output", "json"]
     result = _aws_capture(args, aws_env)
     if result.returncode != 0:
         error = f"{result.stdout}\n{result.stderr}".strip()
         raise RuntimeError(f"list-objects-v2 failed for s3://{bucket}: {error}")
-    keys = json.loads(result.stdout) or []
+    return sorted(json.loads(result.stdout) or [])
-    immutable = []
+
-    for key in keys:
+
-        info = _immutable_head(endpoint, bucket, key, aws_env)
+def _has_publication_address(key):
-        if info is None:
+    return bool(re.search(r"(?<![0-9a-f])[0-9a-f]{64}(?![0-9a-f])", key.lower()))
+
+
+def delete_stale_mutable(endpoint, bucket, object_prefix, html_dir, artifact, aws_env):
+    """Delete stale keys only from the currently served mutable partition."""
+    prefix = f"{object_prefix}/" if object_prefix else ""
+    expected = {
+        f"{prefix}{path.relative_to(html_dir).as_posix()}"
+        for path in html_dir.rglob("*") if path.is_file()
+    }
+    immutable_paths = [rule["path"] for rule in artifact["cache_rules"] if _is_immutable(rule)]
+    stale = []
+    for key in _list_object_keys(endpoint, bucket, aws_env):
+        if prefix and not key.startswith(prefix):
             continue
-        directives = {
+        relative = key[len(prefix):]
-            part.strip().lower().split("=", 1)[0]
+        if key in expected or _has_publication_address(key):
-            for part in (info.get("CacheControl") or "").split(",")
+            continue
-        }
+        if any(relative == path or relative.startswith(f"{path.rstrip('/')}/")
-        if "immutable" in directives:
+               for path in immutable_paths):
-            immutable.append(key)
+            continue
-    return sorted(immutable)
+        if any(fnmatch.fnmatch(relative, pattern) for pattern in artifact["excludes"]):
+            continue
+        stale.append(key)
+    for offset in range(0, len(stale), 1000):
+        batch = stale[offset:offset + 1000]
+        with tempfile.TemporaryDirectory() as tmp:
+            request = Path(tmp) / "delete.json"
+            request.write_text(json.dumps({
+                "Objects": [{"Key": key} for key in batch], "Quiet": True,
+            }))
+            run(["aws", "--endpoint-url", endpoint, "s3api", "delete-objects",
+                 "--bucket", bucket, "--delete", f"file://{request}"], env=aws_env)
 
 
 def _immutable_digests(source, cache_control, content_type):
@@ -191,11 +215,9 @@ def s3_sync(artifact, route, site_dir, credential_env_names=None):
     })
     bucket = artifact["bucket"]
     object_prefix = route["path"].strip("/")
-    bucket_destination = f"s3://{bucket}/"
+    destination = f"s3://{bucket}/{object_prefix + '/' if object_prefix else ''}"
-    destination = f"{bucket_destination}{object_prefix + '/' if object_prefix else ''}"
     default_cache = next(rule["cache_control"] for rule in artifact["cache_rules"] if not rule["path"])
     immutable_paths = [rule["path"] for rule in artifact["cache_rules"] if _is_immutable(rule)]
-    existing_immutable = _existing_immutable_keys(endpoint, bucket, aws_env)
     # `excludes` are patterns (site.yaml `excludes:` list) that should never
     # be uploaded *and* should never be deleted from the bucket — escape hatch
     # for assets managed out-of-band (e.g. large PDFs uploaded via aws-cli).
@@ -203,39 +225,23 @@ def s3_sync(artifact, route, site_dir, credential_env_names=None):
     if artifact["excludes"]:
         print(f"Excluding patterns: {artifact['excludes']}")
     print(f"Syncing artifact {artifact['name']} → {destination} via {endpoint}")
-    # `sync --delete` handles new/changed/orphaned files. Partitioned
+    # Sync handles new and changed mutable files. Stale mutable keys are then
+    # deleted explicitly only inside the current route prefix; old route
+    # prefixes become unreachable with their old Ingress and no immutable key
+    # can match the deletion set. Partitioned
     # `cp --recursive` calls then re-upload each file once to refresh metadata
     # (cache-control, content-type) on objects sync skipped as unchanged.
     # A no-op deploy therefore transfers the artifact bytes once.
     # AWS CLI guesses Content-Type from file extension on local→S3 uploads,
     # so a fresh upload always carries the right MIME type.
-    stage = None
-    sync_source = html_dir
     sync_excludes = [*exclude_args,
-                     *(arg for key in existing_immutable for arg in ("--exclude", key)),
                      *(arg for path in immutable_paths for arg in ("--exclude", f"{path}/*"))]
-    if object_prefix:
+    run(["aws", "--endpoint-url", endpoint, "s3", "sync", f"{html_dir}/", destination,
-        stage = tempfile.TemporaryDirectory()
+         "--only-show-errors", "--cache-control", default_cache,
-        sync_source = Path(stage.name)
+         *sync_excludes], env=aws_env)
-        staged_artifact = sync_source / object_prefix
+    delete_stale_mutable(
-        staged_artifact.parent.mkdir(parents=True, exist_ok=True)
+        endpoint, bucket, object_prefix, html_dir, artifact, aws_env,
-        staged_artifact.symlink_to(html_dir.resolve(), target_is_directory=True)
+    )
-        sync_excludes = [
-            *(arg for pattern in artifact["excludes"]
-              for arg in ("--exclude", f"{object_prefix}/{pattern}")),
-            *(arg for key in existing_immutable for arg in ("--exclude", key)),
-            *(arg for path in immutable_paths
-              for arg in ("--exclude", f"{object_prefix}/{path}/*")),
-        ]
-    try:
-        # Sync the complete bucket authority so moving a route prefix also
-        # deletes objects under its old prefix instead of leaving them public.
-        run(["aws", "--endpoint-url", endpoint, "s3", "sync", f"{sync_source}/", bucket_destination,
-             "--delete", "--only-show-errors", "--cache-control", default_cache,
-             *sync_excludes], env=aws_env)
-    finally:
-        if stage:
-            stage.cleanup()
     print("Re-stamping metadata on all objects...")
     specific_paths = [rule["path"] for rule in artifact["cache_rules"] if rule["path"]]
     default_filters = [arg for path in specific_paths for arg in ("--exclude", f"{path}/*")]

tests/test_contract.py

@@ -389,9 +389,7 @@ class PublishingTests(unittest.TestCase):
             with patch.dict(os.environ, {
                 "DIST_S3_ACCESS_KEY": "dist-key", "DIST_S3_SECRET_KEY": secret
             }, clear=False), patch.object(deploy, "run", side_effect=capture), \
-                    patch.object(deploy, "_existing_immutable_keys", return_value=[
+                    patch.object(deploy, "_list_object_keys", return_value=[]), \
-                        "old-dist/releases/preserved.js"
-                    ]), \
                     patch.object(deploy, "publish_immutable_rule") as immutable_publish, \
                     redirect_stdout(output):
                 deploy.s3_sync(artifact, route, root)
@@ -401,9 +399,9 @@ class PublishingTests(unittest.TestCase):
             self.assertTrue(all(call_env["AWS_ACCESS_KEY_ID"] == "dist-key" for _, call_env in commands))
             self.assertTrue(all("DIST_S3_SECRET_KEY" not in call_env for _, call_env in commands))
             rendered = [" ".join(command) for command, _ in commands]
-            self.assertIn("s3://baseline-dist/", rendered[0])
+            self.assertIn("s3://baseline-dist/dist/", rendered[0])
-            self.assertIn("dist/releases/*", rendered[0])
+            self.assertIn("releases/*", rendered[0])
-            self.assertIn("old-dist/releases/preserved.js", rendered[0])
+            self.assertNotIn("--delete", rendered[0])
             self.assertTrue(all("s3://baseline-dist/dist/" in command for command in rendered[1:]))
             self.assertTrue(any("channels/" in command and
                                 "public, max-age=0, must-revalidate" in command
@@ -414,6 +412,29 @@ class PublishingTests(unittest.TestCase):
                 immutable_publish.call_args.args[2]["cache_control"],
             )
 
+    def test_stale_cleanup_cannot_target_old_prefix_or_content_address(self):
+        cfg = normalize_site_config(fixture("split-site.yaml"), "baseline.fritzlab.net")
+        artifact = next(item for item in cfg["artifacts"] if item["name"] == "distributions")
+        digest_key = "dist/old-" + ("a" * 64) + ".js"
+        keys = ["old-dist/channels/stable.json", digest_key,
+                "dist/releases/legacy.js", "dist/channels/stale.json"]
+        with tempfile.TemporaryDirectory() as tmp:
+            html = Path(tmp)
+            (html / "channels").mkdir()
+            (html / "channels" / "current.json").write_text("current")
+            commands, requests = [], []
+            def capture(command, **_kwargs):
+                commands.append(command)
+                requests.append(json.loads(Path(command[-1].removeprefix("file://")).read_text()))
+            with patch.object(deploy, "_list_object_keys", return_value=keys), \
+                    patch.object(deploy, "run", side_effect=capture):
+                deploy.delete_stale_mutable(
+                    "http://garage-s3.storage.svc:3900", "baseline-dist", "dist",
+                    html, artifact, {},
+                )
+            self.assertEqual(1, len(commands))
+            self.assertEqual([{"Key": "dist/channels/stale.json"}], requests[0]["Objects"])
+
     def test_absent_artifact_is_detected_before_publish(self):
         cfg = normalize_site_config(fixture("split-site.yaml"), "baseline.fritzlab.net")
         with tempfile.TemporaryDirectory() as tmp, redirect_stderr(io.StringIO()), \