fix: exclude generated artifacts from wiki sync

tests/test_gitea_sync_wiki.py

@@ -0,0 +1,49 @@
+from __future__ import annotations
+
+import importlib.util
+import pathlib
+import sys
+import tempfile
+import unittest
+
+
+ROOT = pathlib.Path(__file__).resolve().parents[1]
+SCRIPT = ROOT / "tools" / "gitea" / "gitea-sync-wiki.py"
+TOOLS = SCRIPT.parent
+if str(TOOLS) not in sys.path:
+    sys.path.insert(0, str(TOOLS))
+SPEC = importlib.util.spec_from_file_location("gitea_sync_wiki", SCRIPT)
+assert SPEC and SPEC.loader
+wiki_sync = importlib.util.module_from_spec(SPEC)
+sys.modules[SPEC.name] = wiki_sync
+SPEC.loader.exec_module(wiki_sync)
+
+
+class WikiSourceDiscoveryTests(unittest.TestCase):
+    def test_generated_and_incidental_govoplan_files_are_not_docs(self) -> None:
+        with tempfile.TemporaryDirectory() as temporary:
+            root = pathlib.Path(temporary)
+            paths = {
+                "readme": root / "README.md",
+                "architecture": root / "docs" / "DATASOURCE_ARCHITECTURE.md",
+                "plan": root / "workflow-plan.md",
+                "audit": root / "audit-reports" / "full" / "manifest.json",
+                "runtime": root / "runtime" / "release" / "manifest.json",
+                "incidental": root / "tools" / "semgrep" / "govoplan.yml",
+                "manifest": root / "manifest.json",
+            }
+            for path in paths.values():
+                path.parent.mkdir(parents=True, exist_ok=True)
+                path.write_text("content\n", encoding="utf-8")
+
+            self.assertTrue(wiki_sync.is_repo_doc(root, paths["readme"]))
+            self.assertTrue(wiki_sync.is_repo_doc(root, paths["architecture"]))
+            self.assertTrue(wiki_sync.is_repo_doc(root, paths["plan"]))
+            self.assertFalse(wiki_sync.is_repo_doc(root, paths["audit"]))
+            self.assertFalse(wiki_sync.is_repo_doc(root, paths["runtime"]))
+            self.assertFalse(wiki_sync.is_repo_doc(root, paths["incidental"]))
+            self.assertFalse(wiki_sync.is_repo_doc(root, paths["manifest"]))
+
+
+if __name__ == "__main__":
+    unittest.main()

tools/gitea/gitea-sync-wiki.py

@@ -27,12 +27,17 @@ EXCLUDED_PARTS = {
     ".git",
     ".gitea",
     ".venv",
+    ".mypy_cache",
+    ".pytest_cache",
+    ".ruff_cache",
     "node_modules",
     "__pycache__",
+    "audit-reports",
     "dist",
     "build",
     ".cache",
     ".module-test-build",
+    "runtime",
 }
 EXCLUDED_NAMES = {
     "package-lock.json",
@@ -40,6 +45,14 @@ EXCLUDED_NAMES = {
     "yarn.lock",
     "uv.lock",
 }
+REPO_DOC_NAME_TOKENS = {
+    "architecture",
+    "backlog",
+    "plan",
+    "roadmap",
+    "todo",
+    "workflow",
+}
 
 
 @dataclasses.dataclass(frozen=True)
@@ -216,7 +229,8 @@ def is_repo_doc(repo_root_path: pathlib.Path, path: pathlib.Path) -> bool:
         return False
     if rel.parts[0] in {"docs", "doc", "codex"} and path.suffix.lower() in {".md", ".txt", ".csv"}:
         return True
-    if re.search(r"(backlog|todo|roadmap|plan|architecture|workflow|manifest)", path.name, re.IGNORECASE):
+    name_tokens = set(re.split(r"[^a-z0-9]+", path.stem.lower()))
+    if name_tokens & REPO_DOC_NAME_TOKENS:
         return True
     return False