Files
dragon-hd/tools/build_cjk.py

269 lines
9.2 KiB
Python

"""Build the Chinese UI font atlas + code tables for The I of the Dragon.
Approach
--------
The engine normalizes `Fonts.dat` char rects by the *actual* D3D texture size
(AddChar -> vtable+0x38 surface desc). So if we replace the UI font texture
`UI\\Overhead\\Shrift_gb_Germany.TGA` (256x128) with a bigger atlas that still
contains the original ASCII artwork at pixel (0,0), every existing ASCII rect
keeps working unchanged.
CJK glyphs are rendered as full-width cells at the bottom/right of the atlas and
described by the side tables consumed by the `.cjk` code caves:
cjk_glyph : 65536 * float4 (u1,v1,u2,v2)
cjk_metric : 65536 * int2 (advance, height)
`key = (lead_byte << 8) | trail_byte` -- exactly the code the cave builds.
Outputs (into --out):
Shrift_gb_Germany.TGA new texture (TGA type 2, 32bpp, bottom-up, same desc)
cjk_glyph.bin
cjk_metric.bin
cells.json code -> cell / unicode, for diagnostics
atlas_preview.png downscaled preview (not deployed)
"""
import argparse
import json
import os
import struct
import sys
from PIL import Image, ImageDraw, ImageFont
DEFAULT_GAME = r"F:\steam\steamapps\common\The I of the Dragon"
DEFAULT_OUT = r"E:\DragonHD\build\cjk"
TEX_NAME = r"UI\Overhead\Shrift_gb_Germany.TGA"
FONT_CANDIDATES = [
(r"C:\Windows\Fonts\simsun.ttc", 0),
(r"C:\Windows\Fonts\simhei.ttf", 0),
(r"C:\Windows\Fonts\msyh.ttc", 0),
]
# ---------------------------------------------------------------- res reader
def read_res_entry(res_path, want_name):
data = open(res_path, "rb").read()
count = struct.unpack_from("<I", data, 0)[0]
p = 4
for _ in range(count):
ln = struct.unpack_from("<I", data, p)[0]
p += 4
name = data[p:p + ln].decode("latin1")
p += ln
off, size = struct.unpack_from("<II", data, p)
p += 8
if name == want_name:
return data[off:off + size]
return None
def load_original_tga(game):
loose = os.path.join(game, "Data", "Textures", "Ui", "Overhead",
os.path.basename(TEX_NAME.replace("\\", os.sep)))
if os.path.exists(loose):
return open(loose, "rb").read()
blob = read_res_entry(os.path.join(game, "Data", "Textures.res"), TEX_NAME)
if blob is None:
raise SystemExit("cannot locate " + TEX_NAME)
return blob
# ---------------------------------------------------------------- TGA <-> PIL
def tga_geometry(raw):
if len(raw) < 18:
raise SystemExit("not a TGA")
idlen = raw[0]
imgtype = raw[2]
w = struct.unpack_from("<H", raw, 12)[0]
h = struct.unpack_from("<H", raw, 14)[0]
bpp = raw[16]
desc = raw[17]
return idlen, imgtype, w, h, bpp, desc
def tga_to_image(raw):
idlen, imgtype, w, h, bpp, desc = tga_geometry(raw)
if imgtype != 2 or bpp != 32:
raise SystemExit("unsupported TGA type=%d bpp=%d" % (imgtype, bpp))
off = 18 + idlen
bottom_up = not (desc & 0x20)
img = Image.new("RGBA", (w, h))
dst = img.load()
row_bytes = w * 4
for y in range(h):
src_row = (h - 1 - y) if bottom_up else y
base = off + src_row * row_bytes
row = raw[base:base + row_bytes]
for x in range(w):
b, g, r, a = row[x * 4:x * 4 + 4]
dst[x, y] = (r, g, b, a)
return img, desc
def image_to_tga(img, desc=0x08, footer=b""):
w, h = img.size
rgba = img.convert("RGBA")
r, g, b, a = rgba.split()
bgra = Image.merge("RGBA", (b, g, r, a))
bottom_up = not (desc & 0x20)
if bottom_up:
bgra = bgra.transpose(Image.FLIP_TOP_BOTTOM)
body = bgra.tobytes()
hdr = (struct.pack("<BBB", 0, 0, 2)
+ struct.pack("<HHB", 0, 0, 0)
+ struct.pack("<HHHHBB", 0, 0, w, h, 32, desc))
return hdr + body + footer
# ---------------------------------------------------------------- charset
def iter_gb2312():
"""code(int) -> unicode char, full GB2312 double byte space."""
out = {}
for hi in range(0xA1, 0xF8):
for lo in range(0xA1, 0xFF):
try:
u = bytes([hi, lo]).decode("gb2312")
except Exception:
continue
out[(hi << 8) | lo] = u
return out
def chars_from_text(path, base):
"""merge chars found in a utf-8 text file into the base set."""
txt = open(path, encoding="utf-8").read()
used = {}
for ch in txt:
if ord(ch) < 0x80:
continue
try:
enc = ch.encode("gb2312")
except Exception:
print(" ! cannot gb2312-encode %r" % ch, file=sys.stderr)
continue
if len(enc) != 2:
continue
used[(enc[0] << 8) | enc[1]] = ch
return used
# ---------------------------------------------------------------- build
def pick_font(size):
for path, idx in FONT_CANDIDATES:
if os.path.exists(path):
return ImageFont.truetype(path, size, index=idx), path
raise SystemExit("no CJK ttf found")
def lower_full_stop(tile):
"""Anchor the Chinese full stop at the bottom of its full-width cell."""
bbox = tile.getchannel('A').getbbox()
if bbox is None:
return tile.copy()
result = Image.new('RGBA', tile.size, (0, 0, 0, 0))
result.paste(tile.crop(bbox), (bbox[0], tile.height - (bbox[3] - bbox[1])))
return result
def main():
ap = argparse.ArgumentParser()
ap.add_argument("--game", default=DEFAULT_GAME)
ap.add_argument("--out", default=DEFAULT_OUT)
ap.add_argument("--size", type=int, default=12,
help="CJK pixel size (12 matches the UI pixel font)")
ap.add_argument("--cell", type=int, default=12)
ap.add_argument("--atlas-w", type=int, default=2048)
ap.add_argument("--atlas-h", type=int, default=1024)
ap.add_argument("--cjk-y0", type=int, default=128)
ap.add_argument("--draw-dy", type=int, default=-1,
help="vertical offset of the glyph inside the cell")
ap.add_argument("--charset", default="gb2312",
choices=["gb2312", "text"],
help="gb2312 = every double byte code; text = only chars in --text")
ap.add_argument("--text", default=None)
args = ap.parse_args()
os.makedirs(args.out, exist_ok=True)
orig_raw = load_original_tga(args.game)
orig_img, desc = tga_to_image(orig_raw)
ow, oh = orig_img.size
footer = orig_raw[-26:] if len(orig_raw) >= 26 and orig_raw[-18:-1] == b"TRUEVISION-XFILE." else b""
print("original %dx%d desc=0x%02x footer=%d" % (ow, oh, desc, len(footer)))
base = iter_gb2312()
if args.charset == "text":
base = chars_from_text(args.text, base)
print("charset size: %d" % len(base))
cell = args.cell
cols = args.atlas_w // cell
rows = (args.atlas_h - args.cjk_y0) // cell
if len(base) > cols * rows:
raise SystemExit("charset %d does not fit in %dx%d (%d cells)"
% (len(base), cols, rows, cols * rows))
atlas = Image.new("RGBA", (args.atlas_w, args.atlas_h), (0, 0, 0, 0))
atlas.paste(orig_img, (0, 0))
font, fontpath = pick_font(args.size)
print("font: %s @ %dpx" % (fontpath, args.size))
d = ImageDraw.Draw(atlas)
cells = {}
clipped = 0
for i, code in enumerate(sorted(base)):
uni = base[code]
col = i % cols
row = i // cols
x = col * cell
y = args.cjk_y0 + row * cell
d.text((x, y + args.draw_dy), uni, font=font, fill=(255, 255, 255, 255))
if uni == '\u3002':
atlas.paste(lower_full_stop(atlas.crop((x, y, x + cell, y + cell))), (x, y))
cells[code] = {"x": x, "y": y, "uni": uni}
# sanity: ink must fit the cell
bb = atlas.crop((x, y, x + cell, y + cell)).getbbox()
if bb is None:
clipped += 1
print("placed %d glyphs (%d empty), grid %dx%d" % (len(cells), clipped, cols, rows))
# tables
glyph = bytearray(65536 * 16)
metric = bytearray(65536 * 8)
aw, ah = args.atlas_w, args.atlas_h
for code, c in cells.items():
x, y = c["x"], c["y"]
struct.pack_into("<4f", glyph, code * 16,
x / aw, y / ah, (x + cell) / aw, (y + cell) / ah)
struct.pack_into("<2i", metric, code * 8, cell, cell)
open(os.path.join(args.out, "cjk_glyph.bin"), "wb").write(glyph)
open(os.path.join(args.out, "cjk_metric.bin"), "wb").write(metric)
tga = image_to_tga(atlas, desc=desc, footer=footer)
open(os.path.join(args.out, "Shrift_gb_Germany.TGA"), "wb").write(tga)
with open(os.path.join(args.out, "cells.json"), "w", encoding="utf-8") as f:
json.dump({"cell": cell, "atlas": [aw, ah], "cjk_y0": args.cjk_y0,
"size": args.size, "font": fontpath, "cells": cells}, f)
meta = {"cell": cell, "atlas": [aw, ah], "cjk_y0": args.cjk_y0,
"size": args.size, "font": fontpath, "count": len(cells),
"advance": cell, "height": cell}
with open(os.path.join(args.out, "atlas_meta.json"), "w", encoding="utf-8") as f:
json.dump(meta, f, indent=2)
# preview
prev = atlas.resize((args.atlas_w // 4, args.atlas_h // 4), Image.NEAREST)
prev.save(os.path.join(args.out, "atlas_preview.png"))
print("atlas TGA %d bytes, glyph/probe tables written to %s" % (len(tga), args.out))
if __name__ == "__main__":
main()