Initial SQL-only 1C adapter baseline

scripts/parse_1c_extension_zipped_info.py

@@ -0,0 +1,84 @@
+#!/usr/bin/env python3
+"""Parse _ExtensionsInfo._ExtensionZippedInfo blobs.
+
+Observed structure starts with a 4-byte marker, followed by a 20-byte SHA1 key
+that points to a ConfigCAS file. The rest contains small extension metadata,
+including a UTF-16LE serialized text fragment.
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+from pathlib import Path
+from typing import Any
+
+from inspect_1c_sql_files import GUID_RE
+
+
+def extract_utf16le_fragment(data: bytes) -> str:
+    starts = [pos for pos in (data.find(b"{\x00"), data.find(b'"\x00#\x00"\x00')) if pos >= 0]
+    if not starts:
+        return ""
+    start = min(starts)
+    fragment = data[start:]
+    if len(fragment) % 2:
+        fragment = fragment[:-1]
+    try:
+        return fragment.decode("utf-16-le", errors="ignore").strip("\x00")
+    except UnicodeError:
+        return ""
+
+
+def parse_blob(path: Path, cas_dir: Path | None) -> dict[str, Any]:
+    data = path.read_bytes()
+    root_key = data[4:24].hex() if len(data) >= 24 else ""
+    text = extract_utf16le_fragment(data)
+    cas_path = cas_dir / root_key if cas_dir and root_key else None
+    return {
+        "file_name": path.name,
+        "path": str(path),
+        "bytes": len(data),
+        "marker_hex": data[:4].hex(),
+        "root_cas_key": root_key,
+        "root_cas_exists": bool(cas_path and cas_path.is_file()),
+        "root_cas_path": str(cas_path) if cas_path and cas_path.is_file() else None,
+        "text_fragment": text,
+        "guids": sorted(set(match.lower() for match in GUID_RE.findall(text))),
+    }
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser(description="Parse exported _ExtensionZippedInfo blobs.")
+    parser.add_argument("input", type=Path, help="A blob file or directory with *_ExtensionZippedInfo.bin files.")
+    parser.add_argument("--cas-dir", type=Path)
+    parser.add_argument("--output", type=Path, required=True)
+    args = parser.parse_args()
+
+    files = [args.input] if args.input.is_file() else sorted(args.input.glob("*ExtensionZippedInfo.bin"))
+    items = [parse_blob(path, args.cas_dir) for path in files]
+    report = {
+        "schema": "onec_extension_zipped_info.v1",
+        "input": str(args.input),
+        "cas_dir": str(args.cas_dir) if args.cas_dir else None,
+        "file_count": len(items),
+        "root_cas_found_count": sum(1 for item in items if item["root_cas_exists"]),
+        "items": items,
+    }
+    args.output.parent.mkdir(parents=True, exist_ok=True)
+    args.output.write_text(json.dumps(report, ensure_ascii=False, indent=2) + "\n", encoding="utf-8")
+    print(
+        json.dumps(
+            {
+                "output": str(args.output),
+                "files": len(items),
+                "root_cas_found": report["root_cas_found_count"],
+            },
+            ensure_ascii=False,
+        )
+    )
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())