Sylpheed/tools/re-capture/isl_dialogue.py

#!/usr/bin/env python3
"""Render a mission script's dialogue: every `request_script_message` in order,
with the caption text it plays.

    isl_dialogue.py <StageNN.ssb> [GP_MAIN_GAME_E.pak]

Built-in 64 takes a symbol-table-1 (type 6) message name; the caption text lives
in the pack's IXUD blocks under `<name>_<page>_<line>`. All 2683 call sites
across the 28 stages resolve, so this is a total mapping rather than a sample. 356 of the 1338 distinct
message names span more than one page; every page is printed.
"""
import collections
import struct
import sys

import isl

NO_NAME = 0xFFFFFFFF

# A caption id is keyed `<id>_<page>_<line>`. A page is one subtitle box of up
# to four wrapped lines; successive pages are successive utterances. Measured
# maxima on the English pack: 8 pages, 4 lines.
MAX_PAGES = 16
MAX_LINES = 8


def be32(b, o):
    return struct.unpack_from('>I', b, o)[0]


def wstr(b, pool, off):
    base = pool + 2 * off
    e = base
    while b[e:e + 2] != b'\0\0':
        e += 2
    return ''.join(chr(c) for c in struct.unpack('>%dH' % ((e - base) // 2), b[base:e]))


def captions(pak_entries):
    """`MSG_*` key -> text, read from the IXUD record/field table."""
    out = {}
    for b in pak_entries.values():
        if b[:4] != b'IXUD':
            continue
        n = be32(b, 4)
        o = 8 + 16 * n
        m = be32(b, o)
        o += 4
        fields = [struct.unpack_from('>III', b, o + 12 * i) for i in range(m)]
        o += 12 * m
        pool = o + 4
        for (_k, no, vo) in fields:
            if no == NO_NAME:
                continue
            try:
                key, val = wstr(b, pool, no), wstr(b, pool, vo)
            except Exception:
                continue
            if key.startswith('MSG_') and val.strip():
                out.setdefault(key, val)
    return out


def walk(b, base, nxt, sink):
    staged, pending, off = {}, None, base
    while off + 4 <= nxt:
        w = be32(b, off)
        op, ln = w & 0xFF, (w >> 8) & 0xFF
        k1, k0 = (w >> 24) & 0xFF, (w >> 16) & 0xFF
        if ln < 4 or ln % 4:
            return
        ws = [be32(b, off + i) for i in range(4, max(ln, 4), 4) if off + i + 4 <= len(b)]
        if op in (0, 1) and len(ws) >= 2:
            if k0 == 2 and k1 == 1:
                pending = ws[1]
            elif k0 == 3 and k1 == 2 and pending is not None:
                staged[ws[0]] = pending
            elif k0 == 3 and k1 == 1:
                staged[ws[0]] = ws[1]
        if op == 19 and ws:
            sink(off, ws[0], dict(staged))
            staged = {}
        off += ln


def main():
    ssb = sys.argv[1]
    pak = sys.argv[2] if len(sys.argv) > 2 else \
        '/work/sylph_extract/dat/GP_MAIN_GAME_E.pak'
    sys.path.insert(0, sys.path[0])
    from unitgroup import read_entry  # noqa: F401  (kept for parity)
    import zlib
    import os

    def pak_entries(path):
        d = open(path, 'rb').read()
        assert d[:4] == b'IPFB'
        segs, base, i = b'', path[:-4], 0
        while os.path.exists('%s.p%02d' % (base, i)):
            segs += open('%s.p%02d' % (base, i), 'rb').read()
            i += 1
        out = {}
        for k in range(be32(d, 4)):
            h, off, cs = struct.unpack_from('>III', d, 0x10 + 12 * k)
            raw = segs[off:off + cs]
            if raw[:2] == b'Z1':
                raw = zlib.decompress(raw[10:])
            out[h] = raw
        return out

    cap = captions(pak_entries(pak))
    b = isl.load(ssb)
    s1 = isl.symbols(b, 1)
    end = be32(b, 0x0C)
    lines = []

    def sink(off, bid, st):
        if bid != 64:
            return
        v = st.get(0)
        if not isinstance(v, int) or v not in s1:
            return
        name = s1[v][1]
        pages = []
        for page in range(MAX_PAGES):
            box = [cap[k] for k in
                   ('%s_%03d_%02d' % (name, page, i) for i in range(MAX_LINES))
                   if k in cap]
            if not box:
                break
            pages.append(' '.join(box))
        lines.append((off, name, pages))

    bases = isl.phase_bases(b) + [end]
    for i in range(len(bases) - 1):
        walk(b, bases[i], bases[i + 1], sink)

    name = ssb.replace('\\', '/').split('/')[-1]
    print('# %s — dialogue, in script order' % name)
    print()
    print('Generated by `tools/re-capture/isl_dialogue.py`. Each line is one')
    print('`request_script_message` (built-in 64) and the caption it plays.')
    print()
    multi = sum(1 for _, _, p in lines if len(p) > 1)
    print('%d message calls, %d with text, %d spanning more than one page.'
          % (len(lines), sum(1 for _, _, p in lines if p), multi))
    print()
    print('A page is one subtitle box (measured maximum 4 wrapped lines).')
    print('Successive pages are successive utterances and may be different')
    print('speakers, so a multi-page id is a whole exchange, not one line.')
    print()
    seen = set()
    for off, nm, pages in lines:
        if off in seen:
            continue
        seen.add(off)
        print('%06X  %-22s %s' % (off, nm, pages[0] if pages else ''))
        for extra in pages[1:]:
            print('%6s  %-22s %s' % ('', '', extra))


if __name__ == '__main__':
    main()