feat(provenance): govern source snapshots and data inputs

scripts/export_operator_yolo_dataset.py

@@ -9,6 +9,7 @@ new provider data.
 from __future__ import annotations
 
 import argparse
+import hashlib
 import json
 import os
 import shutil
@@ -16,6 +17,19 @@ import sys
 from pathlib import Path
 from typing import Any, Iterable
 
+SCRIPT_DIR = Path(__file__).resolve().parent
+if str(SCRIPT_DIR) not in sys.path:
+    sys.path.insert(0, str(SCRIPT_DIR))
+
+from training_dataset_eligibility import (  # noqa: E402
+    TrainingEligibilityError,
+    assert_frozen_manifest_training_eligible,
+)
+from training_release_manifest import (  # noqa: E402
+    TrainingReleaseError,
+    create_training_release_manifest,
+)
+
 
 DEFAULT_MANIFEST_PATH = Path("/app/storage/operator-data/operator_samples_manifest.json")
 DEFAULT_OUTPUT_DIR = Path("/app/storage/operator-data/yolo-building-dataset")
@@ -24,6 +38,14 @@ Transformer: Any = None
 Image: Any = None
 
 
+def file_sha256(path: Path) -> str:
+    digest = hashlib.sha256()
+    with path.open("rb") as stream:
+        for chunk in iter(lambda: stream.read(1024 * 1024), b""):
+            digest.update(chunk)
+    return digest.hexdigest()
+
+
 def parse_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser(
         description="Export operator real-data samples to a YOLO detection dataset.",
@@ -45,6 +67,12 @@ def parse_args() -> argparse.Namespace:
         default=os.environ.get("OPERATOR_YOLO_VAL_SAMPLES", "turnhout"),
         help="Comma/space separated sample slugs assigned to validation. Defaults to turnhout.",
     )
+    parser.add_argument(
+        "--review-audit",
+        type=Path,
+        required=True,
+        help="Passed corpus audit with accepted human-review evidence for this frozen source manifest.",
+    )
     parser.add_argument(
         "--force",
         action="store_true",
@@ -218,12 +246,16 @@ def ensure_yolo_directories(output_dir: Path) -> None:
 
 def main() -> int:
     args = parse_args()
+    manifest = json.loads(args.manifest_path.read_text(encoding="utf-8-sig"))
+    try:
+        assert_frozen_manifest_training_eligible(args.manifest_path, verify_live=True)
+    except TrainingEligibilityError as exc:
+        raise SystemExit(str(exc)) from exc
     ensure_dependencies()
     if args.force and args.output_dir.exists():
         shutil.rmtree(args.output_dir)
     args.output_dir.mkdir(parents=True, exist_ok=True)
     ensure_yolo_directories(args.output_dir)
-    manifest = json.loads(args.manifest_path.read_text(encoding="utf-8-sig"))
     samples = manifest.get("samples") or []
     if not samples:
         raise SystemExit("Operator sample manifest contains no samples")
@@ -234,12 +266,26 @@ def main() -> int:
     if not any(item["split"] == "val" for item in exported):
         raise SystemExit("YOLO dataset export produced no validation samples")
     dataset_yaml = write_dataset_yaml(args.output_dir)
+    try:
+        release_paths = create_training_release_manifest(
+            train_yaml=dataset_yaml,
+            corpus_manifest=args.manifest_path,
+            review_audit_path=args.review_audit,
+        )
+    except TrainingReleaseError as exc:
+        raise SystemExit(str(exc)) from exc
     summary = {
         "status": "ok",
         "dataset_yaml": str(dataset_yaml),
+        "training_release_manifest": str(release_paths["release_manifest"]),
+        "training_release_manifest_sha256": file_sha256(release_paths["release_manifest"]),
+        "training_release_freeze": str(release_paths["release_freeze"]),
+        "training_asset_manifest": str(release_paths["asset_manifest"]),
         "output_dir": str(args.output_dir),
         "class_names": ["building"],
         "sample_count": len(exported),
+        "source_manifest": str(args.manifest_path),
+        "source_manifest_sha256": file_sha256(args.manifest_path),
         "train_sample_count": sum(1 for item in exported if item["split"] == "train"),
         "val_sample_count": sum(1 for item in exported if item["split"] == "val"),
         "label_count": sum(item["label_count"] for item in exported),