up

bench/reachability-benchmark/baselines/codeql/normalize.py

@@ -0,0 +1,74 @@
+#!/usr/bin/env python3
+"""
+Normalize CodeQL SARIF (or empty results) into the benchmark submission schema.
+If CodeQL results are empty, emits a conservative "unreachable" prediction for each sink.
+"""
+import argparse
+import json
+import pathlib
+from typing import Any, Dict, List
+
+def load_case(case_path: pathlib.Path) -> Dict[str, Any]:
+    import yaml
+    return yaml.safe_load(case_path.read_text())
+
+def load_codeql_results(path: pathlib.Path) -> Dict[str, Any]:
+    if not path.exists():
+        return {"results": []}
+    try:
+        return json.loads(path.read_text())
+    except json.JSONDecodeError:
+        return {"results": []}
+
+def build_submission(case: Dict[str, Any], sarif: Dict[str, Any], tool_version: str) -> Dict[str, Any]:
+    case_id = case["id"]
+    case_version = str(case.get("version", "1.0.0"))
+    sinks = case.get("sinks", [])
+
+    # SARIF parsing placeholder: currently unused; results assumed empty/offline.
+    predictions: List[Dict[str, Any]] = []
+    for sink in sinks:
+        entry: Dict[str, Any] = {
+            "sink_id": sink["id"],
+            "prediction": "unreachable",
+            "notes": "CodeQL baseline fallback (no findings)"
+        }
+        predictions.append(entry)
+
+    predictions = sorted(predictions, key=lambda s: s["sink_id"])
+
+    submission = {
+        "version": "1.0.0",
+        "tool": {"name": "codeql", "version": tool_version},
+        "run": {"platform": "codeql-baseline-offline"},
+        "cases": [
+            {
+                "case_id": case_id,
+                "case_version": case_version,
+                "sinks": predictions
+            }
+        ]
+    }
+    return submission
+
+def main() -> None:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--case", required=True, help="Path to case.yaml")
+    parser.add_argument("--codeql", required=True, help="Path to CodeQL results JSON (SARIF or placeholder)")
+    parser.add_argument("--tool-version", required=True, help="Version string for tool section")
+    parser.add_argument("--output", required=True, help="Destination submission.json")
+    args = parser.parse_args()
+
+    case_path = pathlib.Path(args.case).resolve()
+    codeql_path = pathlib.Path(args.codeql).resolve()
+    out_path = pathlib.Path(args.output).resolve()
+    out_path.parent.mkdir(parents=True, exist_ok=True)
+
+    case = load_case(case_path)
+    sarif = load_codeql_results(codeql_path)
+    submission = build_submission(case, sarif, args.tool_version)
+
+    out_path.write_text(json.dumps(submission, indent=2, sort_keys=True))
+
+if __name__ == "__main__":
+    main()