Files
d2r-mobile/dm/tools/dmx.py
T
2026-10-12 16:51:08 +08:00

118 lines
4.2 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
#!/usr/bin/env python3
"""Switch D2R RomFS 存储最小提取器(只读)。
存储布局(实测):
build_manifest.psv 每行 = FileId|KeyLayout|Chunk|Archive|CKey|CSize|EKey|ESize|Encoding|Path|Flags
数据文件 = <Chunk:02d>-<Archive:08d>.data
偏移 = EKey 末 5 字节(大端 40 位;key-layout-0 = 8 8 40 1 → EKey[9]=Chunk, EKey[10]=Archive, EKey[11:16]=偏移)
编码 n → 原样 CSize 字节
z → 一整段 zlib(无 BLTE 头)
b:..*=.. → 标准 BLTE 容器('BLTE' + 头长 + 0x0F + 块数 + {esize,csize,md5}×N,每块首字节 N/Z)
校验 md5(解出内容) == CKey
用法:
dmx.py get <Path> <out> 取单个文件(Path 用清单 Path 列,或去掉 'data:' 前缀,大小写不敏感)
dmx.py list [glob] 列 Path\tCSize
Minimal extractor for Switch D2R RomFS storage (read-only).
Storage layout (measured):
build_manifest.psv each line = FileId|KeyLayout|Chunk|Archive|CKey|CSize|EKey|ESize|Encoding|Path|Flags
data file = <Chunk:02d>-<Archive:08d>.data
offset = last 5 bytes of EKey (big-endian 40 bits; key-layout-0 = 8 8 40 1 → EKey[9]=Chunk, EKey[10]=Archive, EKey[11:16]=offset)
encoding n → CSize bytes as is
z → one whole zlib segment (no BLTE header)
b:..*=.. → standard BLTE container ('BLTE' + header length + 0x0F + block count + {esize,csize,md5}×N, first byte of each block N/Z)
check md5(decoded content) == CKey
Usage:
dmx.py get <Path> <out> fetch a single file (Path uses the manifest's Path column, or with the 'data:' prefix removed; case-insensitive)
dmx.py list [glob] list Path\\tCSize
"""
import fnmatch, hashlib, os, struct, sys, zlib
ROOT = os.environ.get("DMX_ROOT", "~/Diablo II Resurrected - 提取内容/Update/RomFS/data")
_IDX = None
_FH = {}
def index():
global _IDX
if _IDX is None:
_IDX = {}
with open(os.path.join(ROOT, "build_manifest.psv"), encoding="utf-8") as f:
next(f)
for line in f:
if line.startswith("#"):
continue
c = line.rstrip("\n").split("|")
if len(c) < 11:
continue
_IDX[c[9].lower()] = c
return _IDX
def lookup(path):
idx = index()
k = path.lower().replace("\\", "/")
for cand in (k, "data:" + k, "data:data/" + k):
if cand in idx:
return idx[cand]
raise KeyError(path)
def _fh(chunk, arch):
key = (chunk, arch)
if key not in _FH:
_FH[key] = open(os.path.join(ROOT, f"{chunk:02d}-{arch:08d}.data"), "rb")
return _FH[key]
def _block(b):
m = b[:1]
if m == b"N":
return b[1:]
if m == b"Z":
return zlib.decompress(b[1:])
raise ValueError(f"未知 BLTE 块模式 {m!r}")
def read_row(c):
chunk, arch, csize, esize, enc = int(c[2]), int(c[3]), int(c[5]), int(c[7]), c[8]
ekey = bytes.fromhex(c[6])
off = int.from_bytes(ekey[11:16], "big")
assert ekey[9] == chunk and ekey[10] == arch, (c[9], ekey.hex())
f = _fh(chunk, arch)
f.seek(off)
raw = f.read(esize)
if raw[:4] == b"BLTE":
hlen = struct.unpack(">I", raw[4:8])[0]
n = int.from_bytes(raw[9:12], "big")
pos, out = hlen, []
for i in range(n):
es, cs = struct.unpack(">II", raw[12 + 24 * i:20 + 24 * i])
out.append(_block(raw[pos:pos + es]))
pos += es
data = b"".join(out)
elif enc == "n":
data = raw
elif enc == "z":
data = zlib.decompress(raw)
else:
raise ValueError(f"未知编码 {enc}")
if len(data) != csize or hashlib.md5(data).hexdigest() != c[4]:
raise ValueError(f"校验失败 {c[9]} len={len(data)} csize={csize}")
return data
def get(path):
return read_row(lookup(path))
if __name__ == "__main__":
if sys.argv[1] == "get":
d = get(sys.argv[2])
open(sys.argv[3], "wb").write(d)
print(len(d), "bytes, md5 ok")
elif sys.argv[1] == "list":
pat = sys.argv[2].lower() if len(sys.argv) > 2 else "*"
for k, c in index().items():
if fnmatch.fnmatch(k, pat):
print(f"{c[9]}\t{c[5]}")