Sync from simready-oem-library-pm@f65d2a11

tools/hf_space/runner.py

@@ -403,6 +403,7 @@ def _validate_zip_streaming(*, api: HfApi, dataset: str, token: str | None,
                              flat_target: Path | None = None,
                              prefetched_zip_entries: list | None = None,
                              prefetched_dataset_head: str | None = None,
+                             continue_on_preliminary: bool = False,
                              ) -> dict | None:
     """Validate a zip-bundled dataset by streaming one archive at a time.
 
@@ -545,6 +546,8 @@ def _validate_zip_streaming(*, api: HfApi, dataset: str, token: str | None,
         "--use-plugin", "--no-use-kit", "--workers", daemon_workers,
         "--profile", profile, "--version", version,
     ]
+    if continue_on_preliminary:
+        daemon_cmd.append("--continue-on-preliminary")
     out(f"  spawning {n_daemons} validator daemon(s) (spec load happens once each)…")
     for di in range(n_daemons):
         try:
@@ -1428,6 +1431,7 @@ def run(
                 flat_target=None,
                 prefetched_zip_entries=probe_zip_entries,
                 prefetched_dataset_head=dataset_head,
+                continue_on_preliminary=True,
             )
             out_dir = work / "out"
             out_dir.mkdir(parents=True, exist_ok=True)
@@ -1566,6 +1570,7 @@ def run(
             flat_target=flat_target,
             prefetched_zip_entries=probe_zip_entries,
             prefetched_dataset_head=dataset_head,
+            continue_on_preliminary=preliminary,
         )
 
         out_dir = work / "out"

tools/validation/plugins/simready-report/skills/simready-report/validate.py

@@ -1594,6 +1594,14 @@ def main() -> int:
                          "real-time bar AND populate per-asset pass/fail in "
                          "the Files expander as assets complete. Writes are "
                          "atomic (tmp + rename).")
+    ap.add_argument("--continue-on-preliminary", action="store_true",
+                    help="Don't short-circuit USD validation when "
+                         "preliminary structure checks (PKG.01/06, "
+                         "AA.002, etc.) find issues. Both phases run and "
+                         "their findings appear in results.json. Used by "
+                         "the dashboard's Preliminary scan tab — the "
+                         "sample asset gets a full ruleset sweep even "
+                         "when its packaging structure is non-conformant.")
     ap.add_argument("--daemon", action="store_true",
                     help="Long-running mode: load foundation specs ONCE at "
                          "startup, then read JSON request lines from stdin "
@@ -1774,7 +1782,7 @@ def _run_validation_body(args, asset_profile_map: dict[str, str]) -> int:
     # the foundation source-of-truth is caught by the out-of-band
     # tools/spec_sync/ job.
     preliminary_issues = run_preliminary_checks(target)
-    if preliminary_issues:
+    if preliminary_issues and not args.continue_on_preliminary:
         out_dir.mkdir(parents=True, exist_ok=True)
         print(f"PRELIMINARY CHECK: {len(preliminary_issues)} issue(s) — "
               f"skipping USD validation until these are addressed", flush=True)
@@ -1809,6 +1817,10 @@ def _run_validation_body(args, asset_profile_map: dict[str, str]) -> int:
             json.dumps(results_json, indent=2), encoding="utf-8"
         )
         return 0
+    if preliminary_issues and args.continue_on_preliminary:
+        print(f"PRELIMINARY CHECK: {len(preliminary_issues)} issue(s) — "
+              f"continuing with full USD validation per --continue-on-preliminary",
+              flush=True)
 
     assets, layout_findings = discover_assets(target, exclude=out_dir)
     print(f"Discovered {len(assets)} USD asset(s) under {target}", flush=True)
@@ -1988,6 +2000,35 @@ def _run_validation_body(args, asset_profile_map: dict[str, str]) -> int:
     if coverage["declared"] is not None and coverage["missing"]:
         print(f"  profile coverage: {coverage['loaded']}/{coverage['declared']} features loaded "
               f"({len(coverage['missing'])} silently dropped — see PROBLEMS.md P1)", flush=True)
+    # If preliminary checks fired alongside USD validation (the
+    # --continue-on-preliminary path), graft their findings into the
+    # final report so the dashboard's per-rule aggregator sees both
+    # phases' issues. Each preliminary finding is attached to its
+    # path's row (a synthetic row is created if no USD asset row
+    # already exists for that path).
+    if preliminary_issues:
+        by_path: dict[str, list[dict]] = {}
+        for f in preliminary_issues:
+            by_path.setdefault(f["path"], []).append(f)
+        row_by_path = {r.get("rel_path"): r for r in results if r.get("rel_path")}
+        for rel, extra in by_path.items():
+            row = row_by_path.get(rel)
+            if row is None:
+                row = {
+                    "asset_path": str(target / rel) if rel != "." else str(target),
+                    "rel_path": rel,
+                    "validation_status": "fail",
+                    "profile": args.profile,
+                    "profile_version": args.version,
+                    "issues": [],
+                    "passed": False,
+                }
+                results.append(row)
+                row_by_path[rel] = row
+            row.setdefault("issues", []).extend(extra)
+            row["passed"] = False
+            row["validation_status"] = "fail"
+
     (out_dir / "results.json").write_text(json.dumps({
         "generated": datetime.now(timezone.utc).isoformat(),
         "target": str(target),
@@ -1998,6 +2039,7 @@ def _run_validation_body(args, asset_profile_map: dict[str, str]) -> int:
         "profile_coverage": coverage,
         "thumbnail_provenance": thumbnail_provenance,
         "layout_findings": layout_findings,
+        "preliminary_findings": preliminary_issues,
         "results": results,
     }, indent=2), encoding="utf-8")