decompyle_with_pycdc

Co-authored-by: Copilot <copilot@github.com>

decompyle_with_pycdc.py

@@ -0,0 +1,106 @@
+#!/usr/bin/env python3
+"""
+Recursively decompile every .pyc file inside an _extracted folder using pycdc,
+saving the results into a 'decompyle' folder with the same directory structure.
+
+Usage:
+    python decompyle_with_pycdc.py <extracted_dir> [--pycdc PATH] [--out OUT_DIR]
+
+Examples:
+    python decompyle_with_pycdc.py yourprogram.exe_extracted
+    python decompyle_with_pycdc.py yourprogram.exe_extracted --pycdc Release\\pycdc.exe --out decompyle
+"""
+
+import argparse
+import subprocess
+import sys
+from pathlib import Path
+
+
+def find_pycdc(user_path: str | None) -> str:
+    """Determine the path to the pycdc executable."""
+    if user_path:
+        if Path(user_path).is_file():
+            return user_path
+        sys.exit(f"[!] Could not find the specified pycdc: {user_path}")
+
+    # Common candidate locations (Windows / Linux)
+    candidates = [
+        "pycdc", "pycdc.exe",
+        "Release/pycdc.exe", "Release\\pycdc.exe",
+        "./pycdc",
+    ]
+    for c in candidates:
+        p = Path(c)
+        if p.is_file():
+            return str(p)
+    # It may be on PATH, so return as-is (validated at run time)
+    return "pycdc"
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="Recursively decompile .pyc files from an _extracted folder using pycdc"
+    )
+    parser.add_argument("extracted_dir", help="pyinstxtractor output folder (_extracted)")
+    parser.add_argument("--pycdc", default=None, help="Path to the pycdc executable")
+    parser.add_argument("--out", default="decompyle", help="Output folder (default: decompyle)")
+    args = parser.parse_args()
+
+    src_root = Path(args.extracted_dir).resolve()
+    if not src_root.is_dir():
+        sys.exit(f"[!] Not a directory: {src_root}")
+
+    out_root = Path(args.out).resolve()
+    out_root.mkdir(parents=True, exist_ok=True)
+
+    pycdc = find_pycdc(args.pycdc)
+
+    pyc_files = sorted(src_root.rglob("*.pyc"))
+    if not pyc_files:
+        sys.exit(f"[!] No .pyc files found in: {src_root}")
+
+    print(f"[*] pycdc       : {pycdc}")
+    print(f"[*] Input dir   : {src_root}")
+    print(f"[*] Output dir  : {out_root}")
+    print(f"[*] Target files: {len(pyc_files)}\n")
+
+    ok = 0
+    fail = 0
+    for pyc in pyc_files:
+        rel = pyc.relative_to(src_root)
+        out_path = (out_root / rel).with_suffix(".py")
+        out_path.parent.mkdir(parents=True, exist_ok=True)
+
+        try:
+            result = subprocess.run(
+                [pycdc, str(pyc)],
+                capture_output=True,
+                text=True,
+                encoding="utf-8",
+                errors="replace",
+            )
+        except FileNotFoundError:
+            sys.exit(f"[!] Failed to run pycdc. Check the path: {pycdc}")
+
+        # pycdc writes the decompiled output to stdout and warnings/errors to stderr.
+        out_path.write_text(result.stdout, encoding="utf-8")
+
+        if result.returncode == 0 and result.stdout.strip():
+            ok += 1
+            print(f"[OK]   {rel} -> {out_path.relative_to(out_root)}")
+        else:
+            fail += 1
+            print(f"[FAIL] {rel}")
+            err = (result.stderr or "").strip()
+            if err:
+                # Show only the first line for brevity
+                print(f"        {err.splitlines()[0]}")
+
+    print(f"\n[*] Done: {ok} succeeded, {fail} failed, {len(pyc_files)} total")
+    if fail:
+        print("[i] Failed files may be version mismatches or header-less entry-point .pyc files.")
+
+
+if __name__ == "__main__":
+    main()