docs(pbf): archive Full production provenance and engineering history

navsea_reencode_unified_semantic_pbf.py

@@ -0,0 +1,92 @@
+from __future__ import annotations
+
+import json
+import os
+import shutil
+import subprocess
+from concurrent.futures import ProcessPoolExecutor
+from pathlib import Path
+
+import mapbox_vector_tile
+
+from navsea_tile_reencode_guard import decode_tile_with_extents, encode_layers_preserving_extents
+
+
+BASE_ROOT = Path("/home/wwwroot/pbf-delivery-full-20260418-rebuild")
+OUTPUT_ROOT = Path("/home/wwwroot/pbf-delivery-full-semantic-unified-20260808")
+MAP_JSON = Path(__file__).resolve().parent / "src/pbf/semantic_sprite_key_map_2026-08-08.json"
+FIELDS = ("chart_icon_image", "chart_fill_pattern", "chart_line_pattern")
+
+
+def hardlink_or_copy(src: Path, dst: Path) -> None:
+    dst.parent.mkdir(parents=True, exist_ok=True)
+    if dst.exists():
+        dst.unlink()
+    try:
+        os.link(src, dst)
+    except OSError:
+        shutil.copy2(src, dst)
+
+
+def rewrite_one(args):
+    src_str, dst_str, value_map = args
+    src = Path(src_str)
+    dst = Path(dst_str)
+    raw = src.read_bytes()
+    decoded, extents = decode_tile_with_extents(raw)
+    changed = 0
+    layers = []
+    for layer_name, layer in decoded.items():
+        features = []
+        for feature in layer.get("features", []):
+            props = dict(feature.get("properties") or {})
+            feature_changed = False
+            for field in FIELDS:
+                value = props.get(field)
+                if isinstance(value, str) and value in value_map:
+                    props[field] = value_map[value]
+                    feature_changed = True
+            if feature_changed:
+                changed += 1
+            out = {"geometry": feature["geometry"], "properties": props}
+            if feature.get("id") is not None:
+                out["id"] = feature["id"]
+            features.append(out)
+        layers.append({"name": layer_name, "features": features})
+    if not changed:
+        hardlink_or_copy(src, dst)
+        return 0, 0
+    dst.parent.mkdir(parents=True, exist_ok=True)
+    # The output tree is initially made with hard links to BASE_ROOT.  Break
+    # that link before writing so the legacy baseline can never be modified.
+    if dst.exists():
+        dst.unlink()
+    dst.write_bytes(encode_layers_preserving_extents(layers, extents))
+    return 1, changed
+
+
+def main() -> None:
+    mapping = json.loads(MAP_JSON.read_text(encoding="utf-8"))
+    mapping["fill-daytime-405"] = mapping["fill-daytime-401"]
+    if OUTPUT_ROOT.exists():
+        shutil.rmtree(OUTPUT_ROOT)
+    OUTPUT_ROOT.mkdir(parents=True, exist_ok=True)
+    subprocess.run(["cp", "-al", f"{BASE_ROOT}/.", str(OUTPUT_ROOT)], check=True)
+
+    command = ["rg", "-a", "-l", "--no-messages", "-F"]
+    for key in mapping:
+        command.extend(["-e", key])
+    command.append(str(BASE_ROOT))
+    result = subprocess.run(command, check=True, capture_output=True, text=True)
+    paths = [Path(line) for line in result.stdout.splitlines() if line.strip()]
+    jobs = [(str(src), str(OUTPUT_ROOT / src.relative_to(BASE_ROOT)), mapping) for src in paths]
+    changed_tiles = changed_features = 0
+    with ProcessPoolExecutor(max_workers=12) as executor:
+        for tile_changed, feature_count in executor.map(rewrite_one, jobs, chunksize=64):
+            changed_tiles += tile_changed
+            changed_features += feature_count
+    print(json.dumps({"base_tiles": len(list(BASE_ROOT.rglob("*.pbf"))), "scanned_tiles": len(paths), "changed_tiles": changed_tiles, "changed_features": changed_features, "output": str(OUTPUT_ROOT)}, ensure_ascii=False, indent=2))
+
+
+if __name__ == "__main__":
+    main()