From 25fdbab10d6047fbbfc4e5ad15b288796e648bb3 Mon Sep 17 00:00:00 2001
From: lifegpc <root@lifegpc.com>
Date: Sat, 20 Sep 2025 14:19:55 +0800
Subject: [PATCH] Add script to unpack softpal text.dat

---
 unpack_softpal_text_dat.py | 73 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 73 insertions(+)
 create mode 100644 unpack_softpal_text_dat.py

diff --git a/unpack_softpal_text_dat.py b/unpack_softpal_text_dat.py
new file mode 100644
index 0000000..2f990e1
--- /dev/null
+++ b/unpack_softpal_text_dat.py
@@ -0,0 +1,73 @@
+import struct
+from io import BytesIO
+import json
+from os.path import splitext
+
+
+def read_u32(f):
+    return struct.unpack('<I', f.read(4))[0]
+
+
+def read_cstring(f):
+    chars = []
+    while True:
+        c = f.read(1)
+        if c == b'\x00' or c == b'':
+            break
+        chars.append(c)
+    return b''.join(chars)
+
+
+def rotate_left(val, rshift):
+    return ((val << rshift) & 0xFF) | (val >> (8 - rshift))
+
+
+def decrypt(data: bytes) -> BytesIO:
+    pos = 0x10
+    shift = 4
+    decrypted = bytearray(data)
+    while pos + 4 <= len(decrypted):
+        data = decrypted[pos:pos+4]
+        data[0] = rotate_left(data[0], shift)
+        shift = (shift + 1) % 8
+        data = struct.unpack('<I', data)[0]
+        data ^= 0x084DF873 ^ 0xFF987DEE
+        decrypted[pos:pos+4] = struct.pack('<I', data)
+        pos += 4
+    return BytesIO(decrypted)
+
+
+def load_text_data(f, e):
+    f.read(12)  # Skip header
+    count = read_u32(f)
+    first_idx = read_u32(f)
+    if first_idx != 0:
+        f.seek(0)
+        f = decrypt(f.read())
+        f.read(16)
+        first_idx = read_u32(f)
+    f.seek(16)
+    table = {}
+    for j in range(count):
+        cur = read_u32(f)
+        s = read_cstring(f).decode(e)
+        table[cur] = s
+    return table
+
+
+if __name__ == '__main__':
+    import sys
+
+    if len(sys.argv) < 2:
+        print("Usage: python unpack_softpal_text_dat.py <file> [<output>] [<encoding>]")
+        sys.exit(1)
+
+    filename = sys.argv[1]
+    output = sys.argv[2] if len(sys.argv) > 2 else splitext(filename)[0] + '.json'
+    encoding = sys.argv[3] if len(sys.argv) > 3 else 'cp932'
+
+    with open(filename, 'rb') as f:
+        table = load_text_data(f, encoding)
+    with open(output, 'w', encoding='utf-8') as f:
+        json.dump(table, f, ensure_ascii=False, indent=4)
+