Add support to modify BGI scenario strings

2025-04-19 12:42:41 +08:00
parent 5a5b61ece5
commit 7c2f3b2778
2 changed files with 220 additions and 0 deletions
--- a/extract_bgi.py
+++ b/extract_bgi.py
@@ -0,0 +1,76 @@
+import struct
+import json
+import sys
+
+
+class BGIScript:
+    def __init__(self, path: str):
+        self.path = path
+        with open(path, 'rb') as f:
+            self.data = bytearray(f.read())
+        self.read_header()
+        self.get_instr_end()
+
+    def read_header(self):
+        self.header_size = struct.unpack('<L', self.data[28:32])[0]
+        self.iPos = 28 + self.header_size
+
+    def get_instr_end(self):
+        finder = self.iPos
+        self.len = len(self.data)
+        self.last_pos = self.len
+        while finder + 4 < self.len:
+            d = struct.unpack('<L', self.data[finder:finder + 4])[0]
+            if d == 0x1b:
+                self.last_pos = finder
+            finder += 4
+        print(self.len)
+
+    def extract_string(self):
+        offset = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0]
+        self.iPos += 4
+        start_pos = 28 + self.header_size + offset
+        if start_pos < self.last_pos:
+            return None
+        pos = start_pos
+        while True:
+            if self.data[pos] == 0x00:
+                break
+            pos += 1
+        data = self.data[start_pos:pos]
+        try:
+            data = data.decode('cp932')
+        except UnicodeDecodeError:
+            data = data.decode('utf-8')
+            data = "utf8:" + data
+        return start_pos, data
+    
+    def extract_value(self):
+        offset = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0]
+        self.iPos += 4
+        return offset
+
+    def extract_values(self):
+        strings = []
+        while self.iPos < self.last_pos:
+            ins = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0]
+            self.iPos += 4
+            if ins == 0x03:
+                string = self.extract_string()
+                if string is not None:
+                    strings.append(string)
+            elif ins in [0x00, 0x01, 0x02, 0x08, 0x09, 0x0a, 0x19, 0x3f]:
+                val = self.extract_value()
+        return dict(strings)
+
+
+base = sys.argv[1]
+json_f = f"{base}.json"
+if len(sys.argv) > 2:
+    json_f = sys.argv[2]
+scr = BGIScript(base)
+print(scr.header_size)
+print(scr.iPos)
+print(scr.last_pos)
+with open(json_f, "w", encoding="utf-8") as f:
+    json.dump(scr.extract_values(), f, ensure_ascii=False, indent=2)
--- a/patch_bgi.py
+++ b/patch_bgi.py
@@ -0,0 +1,144 @@
+import struct
+import json
+import sys
+
+
+class BGIScript:
+    def __init__(self, path: str):
+        self.path = path
+        with open(path, 'rb') as f:
+            self.data = bytearray(f.read())
+        self.read_header()
+        self.get_instr_end()
+
+    def read_header(self):
+        self.header_size = struct.unpack('<L', self.data[28:32])[0]
+        self.iPos = 28 + self.header_size
+
+    def get_instr_end(self):
+        finder = self.iPos
+        self.len = len(self.data)
+        self.last_pos = self.len
+        while finder + 4 < self.len:
+            d = struct.unpack('<L', self.data[finder:finder + 4])[0]
+            if d == 0x1b:
+                self.last_pos = finder
+            finder += 4
+        print(self.len)
+
+    def extract_string(self):
+        offset = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0]
+        self.iPos += 4
+        start_pos = 28 + self.header_size + offset
+        if start_pos < self.last_pos:
+            return None
+        pos = start_pos
+        while True:
+            if self.data[pos] == 0x00:
+                break
+            pos += 1
+        data = self.data[start_pos:pos]
+        try:
+            data = data.decode('cp932')
+        except UnicodeDecodeError:
+            data = data.decode('utf-8')
+            data = "utf8:" + data
+        return start_pos, pos - start_pos
+    
+    def extract_value(self):
+        offset = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0]
+        self.iPos += 4
+        return offset
+
+    def extract_values(self):
+        strings = []
+        while self.iPos < self.last_pos:
+            ins = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0]
+            self.iPos += 4
+            if ins == 0x03:
+                string = self.extract_string()
+                if string is not None:
+                    strings.append(string)
+            elif ins in [0x00, 0x01, 0x02, 0x08, 0x09, 0x0a, 0x19, 0x3f]:
+                val = self.extract_value()
+        return dict(strings)
+
+    def setup_values(self, values: dict):
+        data = bytearray(self.data)
+        offsets = self.extract_values()
+        self.iPos = 28 + self.header_size
+        all_keys = list(int(i) for i in values.keys())
+        all_keys.sort()
+        new_offsets = {}
+        for key in all_keys:
+            new_offsets[key] = key
+        last_len = 0
+        last_key = all_keys[-1]
+        for i in range(len(all_keys)):
+            key = all_keys[i]
+            ori_len = offsets[key] + 1
+            v = values[f"{key}"]
+            if v.startswith("utf8:"):
+                v = v[5:]
+                v = v.encode("utf-8")
+            else:
+                v = v.encode("cp932")
+            v += b'\0'
+            new_len = len(v)
+            last_len = new_len
+            if new_len == ori_len:
+                continue
+            for j in range(i + 1, len(all_keys)):
+                off = all_keys[j]
+                new_offsets[off] += new_len - ori_len
+        old_len = last_key + last_len
+        new_len = new_offsets[last_key] + last_len
+        print(new_len, old_len)
+        data = bytearray(data[:all_keys[0]] + b'\0' * (new_len - old_len) + data[all_keys[0] + last_len:])
+        print(len(data))
+        for i in all_keys:
+            start = new_offsets[i]
+            v = values[f"{i}"]
+            if v.startswith("utf8:"):
+                v = v[5:]
+                v = v.encode("utf-8")
+            else:
+                v = v.encode("cp932")
+            v += b'\0'
+            end = start + len(v)
+            data[start:end] = v
+            print(data[start:end])
+        while self.iPos < self.last_pos:
+            ins = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0]
+            self.iPos += 4
+            if ins == 0x03:
+                off = struct.unpack('<L', self.data[self.iPos:self.iPos + 4])[0] + 28 + self.header_size
+                new_off = new_offsets[int(off)]
+                data[self.iPos:self.iPos + 4] = struct.pack('<L', new_off - 28 - self.header_size)
+                self.iPos += 4
+            elif ins in [0x00, 0x01, 0x02, 0x08, 0x09, 0x0a, 0x19, 0x3f]:
+                val = self.extract_value()
+        print(len(data))
+        print(new_offsets)
+        self.data = data
+
+    def save_as(self, path: str):
+        with open(path, 'wb') as f:
+            f.write(self.data)
+
+
+base = sys.argv[1]
+json_f = f"{base}.json"
+if len(sys.argv) >= 3:
+    json_f = sys.argv[2]
+output = f"../patched/{base}"
+if len(sys.argv) >= 4:
+    output = sys.argv[3]
+scr = BGIScript(base)
+print(scr.header_size)
+print(scr.iPos)
+print(scr.last_pos)
+with open(json_f, "r", encoding="utf-8") as f:
+    data = json.load(f)
+scr.setup_values(data)
+scr.save_as(output)