feat(clew): dump Ren'Py compiled script objects

src/clew/clew.py

@@ -4,6 +4,7 @@ import argparse
 import logging
 import os
 import shutil
+import subprocess
 import sys
 from collections import deque
 from pathlib import Path
@@ -19,7 +20,6 @@ def parse_args(argv: list[str]) -> argparse.Namespace:
     )
     parser.add_argument(
         "--game-root",
-        required=True,
         type=Path,
         help="Game installation root to scan.",
     )
@@ -34,6 +34,11 @@ def parse_args(argv: list[str]) -> argparse.Namespace:
         action="store_true",
         help="Replace Clew-managed output directories.",
     )
+    parser.add_argument(
+        "--dump-renpy-objects",
+        action="store_true",
+        help="Dump compiled Ren'Py scripts from output/extracted/ to output/ast-raw/.",
+    )
     return parser.parse_args(argv)
 
 
@@ -77,16 +82,96 @@ def extract_game_files(game_root: Path, output_path: Path) -> None:
             shutil.copy2(item, destination)
 
 
-def run(args: argparse.Namespace) -> None:
-    game_root = args.game_root.resolve()
-    output_root = resolve_output_root(args.output_root)
+def is_game_root(path: Path) -> bool:
+    return (
+        (path / "game").is_dir()
+        and (path / "renpy").is_dir()
+        and (path / "lib" / "linux-x86_64" / "python").is_file()
+    )
 
-    if not game_root.is_dir():
-        LOG.critical(f'game root is not a directory: {game_root}')
+
+def find_game_roots(extracted_root: Path) -> list[Path]:
+    game_roots = []
+    for root, directories, _files in os.walk(extracted_root):
+        path = Path(root)
+        if is_game_root(path):
+            game_roots.append(path)
+            directories[:] = []
+    return sorted(game_roots)
+
+
+def find_compiled_scripts(game_root: Path) -> list[Path]:
+    scripts = []
+    game_directory = game_root / "game"
+    for suffix in ("*.rpyc", "*.rpymc"):
+        scripts.extend(game_directory.rglob(suffix))
+    return sorted(scripts)
+
+
+def object_dump_output_path(output_root: Path, extracted_root: Path, game_root: Path, script: Path) -> Path:
+    namespace = game_root.relative_to(extracted_root)
+    relative_script = script.relative_to(game_root)
+    return output_root / "ast-raw" / namespace / relative_script.with_suffix(relative_script.suffix + ".json")
+
+
+def run_renpy_object_dump(game_root: Path, script: Path, output: Path) -> None:
+    exporter = Path(__file__).with_name("renpy_object_dump.py").resolve()
+    bundled_python = game_root / "lib" / "linux-x86_64" / "python"
+    output.parent.mkdir(parents=True, exist_ok=True)
+    command = [
+        "steam-run",
+        str(bundled_python),
+        str(exporter),
+        "--game-root",
+        str(game_root),
+        "--input",
+        str(script),
+        "--output",
+        str(output),
+    ]
+    LOG.info("dumping Ren'Py object graph: %s", script)
+    subprocess.run(command, check=True)
+
+
+def dump_renpy_objects(output_root: Path, overwrite: bool) -> None:
+    extracted_root = output_root / "extracted"
+    if not extracted_root.is_dir():
+        LOG.critical("extracted output directory does not exist: %s", extracted_root)
         raise SystemExit(2)
 
-    extracted = get_output_directory(output_root, "extracted")
-    extract_game_files(game_root, extracted)
+    game_roots = find_game_roots(extracted_root)
+    if not game_roots:
+        LOG.warning("no Ren'Py game roots found under %s", extracted_root)
+        return
+
+    for game_root in game_roots:
+        scripts = find_compiled_scripts(game_root)
+        LOG.info("found %d compiled Ren'Py scripts under %s", len(scripts), game_root)
+        for script in scripts:
+            output = object_dump_output_path(output_root, extracted_root, game_root, script)
+            if output.exists() and not overwrite:
+                LOG.info("skipping existing Ren'Py object dump: %s", output)
+                continue
+            run_renpy_object_dump(game_root, script, output)
+
+
+def run(args: argparse.Namespace) -> None:
+    output_root = resolve_output_root(args.output_root)
+
+    if args.game_root is not None:
+        game_root = args.game_root.resolve()
+        if not game_root.is_dir():
+            LOG.critical(f"game root is not a directory: {game_root}")
+            raise SystemExit(2)
+
+        extracted = get_output_directory(output_root, "extracted")
+        extract_game_files(game_root, extracted)
+    elif not args.dump_renpy_objects:
+        LOG.critical("missing --game-root unless --dump-renpy-objects is set")
+        raise SystemExit(2)
+
+    if args.dump_renpy_objects:
+        dump_renpy_objects(output_root, args.overwrite)
 
 
 if __name__ == "__main__":