fix: harden benchmark scripts against filesystem and type errors

Wrap reduce_baselines I/O in BenchmarkDataError; guard machine_info type;
move float(mean) into try block in load_results; assert argparse exit code 2.

Author: clean6378-max-it

scripts/check_benchmark_regression.py

@@ -33,12 +33,12 @@ def load_results(results_path: str | Path) -> dict[str, float]:
             raise BenchmarkDataError(f"{path} benchmarks[{index}] must be an object")
         try:
             name = entry["name"]
-            mean = entry["stats"]["mean"]
-        except (KeyError, TypeError) as exc:
+            mean = float(entry["stats"]["mean"])
+        except (KeyError, TypeError, ValueError) as exc:
             raise BenchmarkDataError(
                 f"{path} benchmarks[{index}] missing 'name' or 'stats.mean'"
             ) from exc
-        results[str(name)] = float(mean)
+        results[str(name)] = mean
     return results
 
 

scripts/reduce_baselines.py

@@ -31,6 +31,8 @@ def reduce_baselines(
         raw = json.loads(path.read_text(encoding="utf-8"))
     except json.JSONDecodeError as exc:
         raise BenchmarkDataError(f"invalid JSON in {path}: {exc}") from exc
+    except OSError as exc:
+        raise BenchmarkDataError(f"cannot read {path}: {exc}") from exc
 
     try:
         entries = raw["benchmarks"]
@@ -55,15 +57,19 @@ def reduce_baselines(
             continue
         groups[group][str(name)] = mean * slack
 
-    machine_info = raw.get("machine_info", {})
+    machine_info = raw.get("machine_info")
+    machine = machine_info.get("system") if isinstance(machine_info, dict) else None
     output: dict[str, object] = {
         "_note": "CI gates the ubuntu benchmarks job when mean exceeds baseline by >20%.",
         "updated": datetime.now(UTC).strftime("%Y-%m-%dT%H:%M:%SZ"),
-        "machine": machine_info.get("system"),
+        "machine": machine,
         "groups": groups,
     }
     out = Path(out_path)
-    out.write_text(json.dumps(output, indent=2) + "\n", encoding="utf-8")
+    try:
+        out.write_text(json.dumps(output, indent=2) + "\n", encoding="utf-8")
+    except OSError as exc:
+        raise BenchmarkDataError(f"cannot write {out}: {exc}") from exc
     return output
 
 

tests/test_reduce_baselines.py

@@ -74,5 +74,6 @@ def test_reduce_baselines_cli_rejects_non_positive_slack(tmp_path) -> None:
         [{"group": "parse", "name": "test_parse_session_small", "stats": {"mean": 0.0001}}],
     )
 
-    with pytest.raises(SystemExit):
+    with pytest.raises(SystemExit) as exc_info:
         main([str(raw), str(tmp_path / "out.json"), "--slack", "0"])
+    assert exc_info.value.code == 2