up

scripts/bench/README.md

@@ -1,10 +1,12 @@
 # Bench scripts
 
-- `determinism-run.sh`: runs BENCH-DETERMINISM-401-057 harness (`src/Bench/StellaOps.Bench/Determinism`), writes artifacts to `out/bench-determinism`, and enforces threshold via `BENCH_DETERMINISM_THRESHOLD` (default 0.95). Defaults to 10 runs per scanner/SBOM pair. Pass `DET_EXTRA_INPUTS` (space-separated globs) to include frozen feeds in `inputs.sha256`; `DET_RUN_EXTRA_ARGS` to forward extra args to the harness.
+- `determinism-run.sh`: runs BENCH-DETERMINISM-401-057 harness (`src/Bench/StellaOps.Bench/Determinism`), writes artifacts to `out/bench-determinism`, and enforces threshold via `BENCH_DETERMINISM_THRESHOLD` (default 0.95). Defaults to 10 runs per scanner/SBOM pair. Pass `DET_EXTRA_INPUTS` (space-separated globs) to include frozen feeds in `inputs.sha256`; `DET_RUN_EXTRA_ARGS` to forward extra args to the harness; `DET_REACH_GRAPHS`/`DET_REACH_RUNTIME` to hash reachability datasets and emit `dataset.sha256` + `results-reach.*`.
 
 Usage:
 ```sh
 BENCH_DETERMINISM_THRESHOLD=0.97 \
 DET_EXTRA_INPUTS="offline/feeds/*.tar.gz" \
+DET_REACH_GRAPHS="offline/reachability/graphs/*.json" \
+DET_REACH_RUNTIME="offline/reachability/runtime/*.ndjson" \
 scripts/bench/determinism-run.sh
 ```

scripts/bench/determinism-run.sh

@@ -28,5 +28,28 @@ printf "timestamp=%s\n" "$(date -u +"%Y-%m-%dT%H:%M:%SZ")" >> "$OUT/summary.txt"
 
 awk -v rate="$det_rate" -v th="$THRESHOLD" 'BEGIN {if (rate+0 < th+0) {printf("determinism_rate %s is below threshold %s\n", rate, th); exit 1}}'
 
+if [ -n "${DET_REACH_GRAPHS:-}" ]; then
+  echo "[bench-determinism] running reachability dataset hash"
+  reach_graphs=${DET_REACH_GRAPHS}
+  reach_runtime=${DET_REACH_RUNTIME:-}
+  # prefix relative globs with repo root for consistency
+  case "$reach_graphs" in
+    /*) ;; 
+    *) reach_graphs="${ROOT}/${reach_graphs}" ;;
+  esac
+  case "$reach_runtime" in
+    /*|"") ;; 
+    *) reach_runtime="${ROOT}/${reach_runtime}" ;;
+  esac
+  python run_reachability.py \
+    --graphs ${reach_graphs} \
+    --runtime ${reach_runtime} \
+    --output results
+  # copy reachability outputs
+  cp results/results-reach.csv "$OUT"/ || true
+  cp results/results-reach.json "$OUT"/ || true
+  cp results/dataset.sha256 "$OUT"/ || true
+fi
+
 tar -C "$OUT" -czf "$OUT/bench-determinism-artifacts.tgz" .
 echo "[bench-determinism] artifacts at $OUT"