Fix traffic archive extraction workflow

2026-06-19 20:48:08 -05:00
parent 20daac34e5
commit 355b135d69
3 changed files with 126 additions and 28 deletions
@@ -21,13 +21,24 @@ pack_root="${work_dir}/${mod_name}"
 packed_dir="${work_dir}/05_packed"
 resource_dir=$(dirname "${resource_path}")
 resource_file=$(basename "${resource_path}")
+resource_hash=3419764573789342681

-mkdir -p "${raw_dir}" "${json_dir}" "${patched_json_dir}" "${patched_raw_dir}" "${pack_root}/${resource_dir}" "${packed_dir}"
+mkdir -p "${raw_dir}/${resource_dir}" "${json_dir}" "${patched_json_dir}" "${patched_raw_dir}" "${pack_root}/${resource_dir}" "${packed_dir}"

-tools/cp77_toolbox.sh extract "${archive_path}" \
-  --gamepath "${game_dir}" \
-  --outpath "${raw_dir}" \
-  --regex "${resource_regex}"
+if [[ -f "${archive_path}" ]]; then
+  kark_path="${raw_dir}/${resource_file}.kark"
+  decompressed_base="${raw_dir}/${resource_file}.raw"
+  python3 tools/extract_archive_segment.py "${archive_path}" "${resource_hash}" "${kark_path}"
+  tools/cp77_toolbox.sh oodle decompress "${kark_path}" "${decompressed_base}" || true
+  if [[ ! -f "${decompressed_base}.bin" ]]; then
+    echo "failed to decompress ${kark_path}" >&2
+    exit 1
+  fi
+  mv "${decompressed_base}.bin" "${raw_dir}/${resource_path}"
+else
+  echo "raw segment extraction needs a single archive file, got: ${archive_path}" >&2
+  exit 2
+fi

 tools/cp77_toolbox.sh convert serialize "${raw_dir}" --outpath "${json_dir}"

@@ -0,0 +1,77 @@
+#!/usr/bin/env python3
+"""Extract a single compressed RED4 archive segment by resource hash."""
+
+from __future__ import annotations
+
+import argparse
+import struct
+from pathlib import Path
+
+
+FILE_ENTRY_SIZE = 56
+FILE_SEGMENT_SIZE = 16
+
+
+def parse_hash(value: str) -> int:
+    return int(value, 0)
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("archive", type=Path)
+    parser.add_argument("resource_hash", type=parse_hash)
+    parser.add_argument("outpath", type=Path)
+    args = parser.parse_args()
+
+    with args.archive.open("rb") as archive:
+        header = archive.read(40)
+        if len(header) != 40:
+            raise SystemExit("archive header is too short")
+        magic, _version, index_pos, _index_size, *_rest = struct.unpack("<IIQIQIQ", header)
+        if magic != 1380009042:
+            raise SystemExit(f"unexpected archive magic: {magic}")
+
+        archive.seek(index_pos + 16)
+        file_count, segment_count, _dependency_count = struct.unpack("<III", archive.read(12))
+
+        found: tuple[int, int] | None = None
+        for _ in range(file_count):
+            entry = archive.read(FILE_ENTRY_SIZE)
+            (
+                key,
+                _timestamp,
+                _flags,
+                segments_start,
+                segments_end,
+                _dependencies_start,
+                _dependencies_end,
+                _sha1,
+            ) = struct.unpack("<QqIIIII20s", entry)
+            if key == args.resource_hash:
+                if segments_end - segments_start != 1:
+                    raise SystemExit(
+                        f"resource has {segments_end - segments_start} segments; expected 1"
+                    )
+                found = (segments_start, segments_end)
+
+        if found is None:
+            raise SystemExit(f"resource hash {args.resource_hash} not found")
+
+        segments = []
+        for _ in range(segment_count):
+            segments.append(struct.unpack("<QII", archive.read(FILE_SEGMENT_SIZE)))
+
+        offset, z_size, size = segments[found[0]]
+        archive.seek(offset)
+        data = archive.read(z_size)
+        if len(data) != z_size:
+            raise SystemExit("archive ended before segment data was fully read")
+
+    args.outpath.parent.mkdir(parents=True, exist_ok=True)
+    args.outpath.write_bytes(data)
+    print(f"wrote {args.outpath} z_size={z_size} size={size}")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())