#!/usr/bin/env python3 """Export the recovered monster HP-view stat table. The table is grounded by the monster HP dword and a +0x34 boundary marker in this exported HP-view. The actor initializer uses an 8-byte wider row prefix; see battle_actor_stat_layout_review for the actor struct mapping. The monster level lives in that prefix at actual row +0x06, which is HP-view -0x02 in this export. """ from __future__ import annotations import argparse import html import json from pathlib import Path from typing import Any ROOT = Path(__file__).resolve().parents[1] OUT = ROOT / "out" EXE = ROOT / "Hwanse2.exe" IMAGE_BASE_DELTA = 0x00402000 STAT_TABLE_FILE_OFFSET = 0x00055D48 STAT_TABLE_VA = STAT_TABLE_FILE_OFFSET + IMAGE_BASE_DELTA STAT_ROW_SIZE = 0x38 NAME_ID_BASE = 0x0205 NAME_POINTER_TABLE_VA = 0x004EC2C4 NAME_POINTER_MIN_VA = 0x004EC300 NAME_POINTER_MAX_VA = 0x004EC800 FRONT_NUMERIC_FIELD_OFFSETS = [0x04, 0x06, 0x08, 0x0A, 0x0C, 0x0E, 0x10, 0x12, 0x14] RATE_BYTE_START = 0x16 RATE_BYTE_END = 0x28 FIELD_LAYOUT = { "confirmed": [ {"offsetHex": "HP-view -0x02 / actual +0x06", "width": "u16", "label": "level"}, {"offsetHex": "+0x00", "width": "u32", "label": "HP"}, {"offsetHex": "+0x2e", "width": "u16", "label": "gold"}, {"offsetHex": "+0x34", "width": "u16-byte-swapped", "label": "name id boundary marker"}, ], "candidate": [ { "offsetRangeHex": "+0x04..+0x14", "width": "u16 x9", "label": "front numeric battle-stat candidates", }, { "offsetRangeHex": "+0x16..+0x27", "width": "u8 x18", "label": "rate/resistance-style byte candidates", }, { "offsetHex": "+0x36", "width": "u16", "label": "next-row prefix level; not current-row data", }, ], "reservedZero": [ {"offsetHex": "+0x02", "width": "u16"}, {"offsetRangeHex": "+0x28..+0x2c", "width": "u16 x3"}, {"offsetRangeHex": "+0x30..+0x32", "width": "u16 x2"}, ], } def va_to_file_offset(va: int) -> int: return va - IMAGE_BASE_DELTA def read_u16(data: bytes, offset: int) -> int: return int.from_bytes(data[offset:offset + 2], "little") def read_u32(data: bytes, offset: int) -> int: return int.from_bytes(data[offset:offset + 4], "little") def swap_u16(value: int) -> int: return ((value & 0xFF) << 8) | (value >> 8) def hex_va(value: int) -> str: return f"0x{value:08x}" def hex_off(value: int) -> str: return f"0x{value:06x}" def clean_name(value: str) -> str: return value.replace("\u3000", "").strip() def read_marker_string(data: bytes, va: int) -> str: offset = va_to_file_offset(va) if offset < 0 or offset >= len(data): return "" raw = data[offset:offset + 64] stop = len(raw) for marker in (b"@", b"\0"): index = raw.find(marker) if index >= 0: stop = min(stop, index) try: return raw[:stop].strip().decode("cp949") except UnicodeDecodeError: return "" def read_name_pointer_table(data: bytes) -> list[dict[str, Any]]: rows: list[dict[str, Any]] = [] table_offset = va_to_file_offset(NAME_POINTER_TABLE_VA) index = 0 while table_offset + index * 4 + 4 <= len(data): ptr = read_u32(data, table_offset + index * 4) if not (NAME_POINTER_MIN_VA <= ptr < NAME_POINTER_MAX_VA): break name = read_marker_string(data, ptr) if not name or not any("\uac00" <= char <= "\ud7a3" for char in name): break name_id = NAME_ID_BASE + index rows.append({ "index": index, "nameId": name_id, "nameIdHex": f"0x{name_id:04x}", "pointerTableVa": NAME_POINTER_TABLE_VA + index * 4, "pointerTableVaHex": hex_va(NAME_POINTER_TABLE_VA + index * 4), "nameVa": ptr, "nameVaHex": hex_va(ptr), "name": name, "cleanName": clean_name(name), }) index += 1 return rows def stat_words(data: bytes, offset: int) -> list[dict[str, Any]]: words = [] for word_offset in range(0, STAT_ROW_SIZE, 2): value = read_u16(data, offset + word_offset) words.append({ "offset": word_offset, "offsetHex": f"+0x{word_offset:02x}", "value": value, "hex": f"0x{value:04x}", }) return words def field_breakdown(data: bytes, offset: int) -> dict[str, Any]: blob = data[offset:offset + STAT_ROW_SIZE] front_fields = [ { "offset": field_offset, "offsetHex": f"+0x{field_offset:02x}", "key": f"u16_{field_offset:02x}", "label": f"u16 +0x{field_offset:02x}", "value": read_u16(data, offset + field_offset), "hex": f"0x{read_u16(data, offset + field_offset):04x}", "status": "candidate-front-numeric-field", } for field_offset in FRONT_NUMERIC_FIELD_OFFSETS ] rate_bytes = [ { "offset": field_offset, "offsetHex": f"+0x{field_offset:02x}", "key": f"rate_{field_offset - RATE_BYTE_START:02d}", "label": f"rate byte {field_offset - RATE_BYTE_START:02d}", "value": blob[field_offset], "hex": f"0x{blob[field_offset]:02x}", "status": "candidate-rate-or-resistance-byte", } for field_offset in range(RATE_BYTE_START, RATE_BYTE_END) ] tail_value = read_u16(data, offset + 0x36) level = read_u16(data, offset - 2) return { "confirmedFields": [ { "label": "level", "offsetHex": "HP-view -0x02 / actual +0x06", "width": "u16", "value": level, }, {"label": "HP", "offsetHex": "+0x00", "width": "u32", "value": read_u32(data, offset)}, {"label": "gold", "offsetHex": "+0x2e", "width": "u16", "value": read_u16(data, offset + 0x2E)}, { "label": "name id", "offsetHex": "+0x34", "width": "u16-byte-swapped", "value": swap_u16(read_u16(data, offset + 0x34)), "rawHex": f"0x{read_u16(data, offset + 0x34):04x}", }, ], "frontNumericFields": front_fields, "rateByteFields": rate_bytes, "tailCandidateField": { "offset": 0x36, "offsetHex": "+0x36", "key": "u16_36", "label": "next-row prefix level", "value": tail_value, "hex": f"0x{tail_value:04x}", "status": "next-row-boundary-field-not-current-row-level", }, "frontNumericSummary": " ".join(f"+{row['offset']:02x}={row['value']}" for row in front_fields), "rateByteSummary": " ".join(f"{row['value']}" for row in rate_bytes), "candidateFieldStatus": "structured-candidate-fields-not-promoted-to-formula-labels", } def read_stat_rows(data: bytes, names: list[dict[str, Any]]) -> list[dict[str, Any]]: name_by_id = {int(row["nameId"]): row for row in names} rows: list[dict[str, Any]] = [] index = 0 while STAT_TABLE_FILE_OFFSET + (index + 1) * STAT_ROW_SIZE <= len(data): offset = STAT_TABLE_FILE_OFFSET + index * STAT_ROW_SIZE actor_prefix_offset = offset - 8 raw_name_id = read_u16(data, offset + 0x34) name_id = swap_u16(raw_name_id) expected_id = NAME_ID_BASE + index if name_id != expected_id: break name_row = name_by_id.get(name_id) if not name_row: break hp = read_u32(data, offset) gold = read_u32(data, offset + 0x2E) level = read_u16(data, offset - 2) actor_kind_byte = data[offset - 4] actor_subtype_byte = data[offset - 3] words = stat_words(data, offset) breakdown = field_breakdown(data, offset) rows.append({ "index": index, "actorPrefixFileOffset": actor_prefix_offset, "actorPrefixFileOffsetHex": hex_off(actor_prefix_offset), "actorPrefixVa": actor_prefix_offset + IMAGE_BASE_DELTA, "actorPrefixVaHex": hex_va(actor_prefix_offset + IMAGE_BASE_DELTA), "actorKindByte": actor_kind_byte, "actorKindByteHex": f"0x{actor_kind_byte:02x}", "actorSubtypeByte": actor_subtype_byte, "actorSubtypeByteHex": f"0x{actor_subtype_byte:02x}", "fileOffset": offset, "fileOffsetHex": hex_off(offset), "va": offset + IMAGE_BASE_DELTA, "vaHex": hex_va(offset + IMAGE_BASE_DELTA), "rowSize": STAT_ROW_SIZE, "rowSizeHex": f"0x{STAT_ROW_SIZE:02x}", "name": name_row["name"], "cleanName": name_row["cleanName"], "nameId": name_id, "nameIdHex": f"0x{name_id:04x}", "rawNameId": raw_name_id, "rawNameIdHex": f"0x{raw_name_id:04x}", "nameVa": name_row["nameVa"], "nameVaHex": name_row["nameVaHex"], "namePointerTableVa": name_row["pointerTableVa"], "namePointerTableVaHex": name_row["pointerTableVaHex"], "hp": hp, "gold": gold, "level": level, "levelHex": f"0x{level:04x}", "levelSource": "actual stat row +0x06 / exported HP-view -0x02; copied to actor +0x06 and consumed by EXP formula 0x0042157b", "field2Hex": f"0x{read_u16(data, offset + 0x36):04x}", "hpViewPlus36NextRowLevelCandidate": read_u16(data, offset + 0x36), "hpViewPlus36NextRowLevelCandidateHex": f"0x{read_u16(data, offset + 0x36):04x}", "hpViewPlus36Meaning": "next actual row prefix +0x06 level; shifted by one row, not this monster level", "words": words, "wordHex": " ".join(word["hex"][2:] for word in words), **breakdown, }) index += 1 occurrence_counts: dict[str, int] = {} for row in rows: key = str(row["cleanName"]) occurrence_counts[key] = occurrence_counts.get(key, 0) + 1 row["nameOccurrence"] = occurrence_counts[key] total_by_name: dict[str, int] = {} for row in rows: key = str(row["cleanName"]) total_by_name[key] = total_by_name.get(key, 0) + 1 for row in rows: row["nameDuplicateCount"] = total_by_name[str(row["cleanName"])] return rows def trailing_candidate_block(data: bytes, index: int, name_row: dict[str, Any]) -> dict[str, Any] | None: offset = STAT_TABLE_FILE_OFFSET + index * STAT_ROW_SIZE if offset + STAT_ROW_SIZE > len(data): return None hp = read_u32(data, offset) if not (1 <= hp <= 10000): return None raw_name_id = read_u16(data, offset + 0x34) swapped_name_id = swap_u16(raw_name_id) words = stat_words(data, offset) return { "status": "trailing-raw-block-candidate", "reason": "The block starts where the next 0x38 stat row would start and has a plausible HP-like first value, but it lacks the row-local byte-swapped name id at +0x34.", "index": index, "fileOffset": offset, "fileOffsetHex": hex_off(offset), "va": offset + IMAGE_BASE_DELTA, "vaHex": hex_va(offset + IMAGE_BASE_DELTA), "name": name_row["name"], "cleanName": name_row["cleanName"], "nameId": name_row["nameId"], "nameIdHex": name_row["nameIdHex"], "hpCandidate": hp, "hpCandidateHex": f"0x{hp:08x}", "u16At2e": read_u16(data, offset + 0x2E), "u16At2eHex": f"0x{read_u16(data, offset + 0x2E):04x}", "rawNameIdAt34": raw_name_id, "rawNameIdAt34Hex": f"0x{raw_name_id:04x}", "swappedNameIdAt34": swapped_name_id, "swappedNameIdAt34Hex": f"0x{swapped_name_id:04x}", "field36Hex": f"0x{read_u16(data, offset + 0x36):04x}", "words": words, "wordHex": " ".join(word["hex"][2:] for word in words), } def attach_name_only_candidates(data: bytes, rows: list[dict[str, Any]], name_only: list[dict[str, Any]]) -> None: for offset_index, row in enumerate(name_only): candidate = trailing_candidate_block(data, len(rows) + offset_index, row) if candidate: row["trailingCandidateBlock"] = candidate def build_summary(data: bytes) -> dict[str, Any]: names = read_name_pointer_table(data) rows = read_stat_rows(data, names) name_only = names[len(rows):] attach_name_only_candidates(data, rows, name_only) duplicate_names = sorted({ row["cleanName"] for row in rows if row.get("nameDuplicateCount", 0) > 1 }) return { "scope": "Monster stat HP-view rows recovered from Hwanse2.exe by HP/gold fields and the +0x34 name-id boundary marker.", "source": ["Hwanse2.exe"], "promotionStatus": "enemy-stat-table-name-hp-gold-grounded", "table": { "statTableFileOffset": STAT_TABLE_FILE_OFFSET, "statTableFileOffsetHex": hex_off(STAT_TABLE_FILE_OFFSET), "statTableVa": STAT_TABLE_VA, "statTableVaHex": hex_va(STAT_TABLE_VA), "rowSize": STAT_ROW_SIZE, "rowSizeHex": f"0x{STAT_ROW_SIZE:02x}", "nameIdBase": NAME_ID_BASE, "nameIdBaseHex": f"0x{NAME_ID_BASE:04x}", "namePointerTableVa": NAME_POINTER_TABLE_VA, "namePointerTableVaHex": hex_va(NAME_POINTER_TABLE_VA), }, "rowCount": len(rows), "namePointerCount": len(names), "nameOnlyCount": len(name_only), "duplicateNames": duplicate_names, "fieldLayout": FIELD_LAYOUT, "confirmedFieldCountPerRow": 4, "candidateFrontNumericFieldCountPerRow": len(FRONT_NUMERIC_FIELD_OFFSETS), "candidateRateByteFieldCountPerRow": RATE_BYTE_END - RATE_BYTE_START, "rows": rows, "namePointers": names, "nameOnlyEntries": name_only, "checks": { "continuousStatRowsRecovered": len(rows) == 80, "namePointersRecovered": len(names) == 81, "rowStrideIs0x38": STAT_ROW_SIZE == 0x38, "hpOffsetHex": "+0x00", "levelOffsetHex": "HP-view -0x02 / actual +0x06 / actor +0x06", "goldOffsetHex": "+0x2e", "nameIdOffsetHex": "+0x34", "hpViewPlus36IsNextRowPrefixLevel": True, "candidateFrontNumericRangeHex": "+0x04..+0x14", "candidateRateByteRangeHex": "+0x16..+0x27", "firstRowName": rows[0]["cleanName"] if rows else "", "lastContinuousRowName": rows[-1]["cleanName"] if rows else "", "trailingNameOnlyEntries": [row["cleanName"] for row in name_only], "trailingCandidateBlocks": [ { "name": row["cleanName"], "nameIdHex": row["nameIdHex"], "vaHex": row["trailingCandidateBlock"]["vaHex"], "hpCandidate": row["trailingCandidateBlock"]["hpCandidate"], } for row in name_only if row.get("trailingCandidateBlock") ], }, "notes": [ "HP and gold are grounded from the recovered stat row layout.", "In this exported HP-view, the name id boundary marker is stored byte-swapped at +0x34, e.g. bytes 26 02 mean logical id 0x0226. The actor initializer uses an 8-byte wider row prefix, so +0x34 is not a current actor struct field.", "Monster level is recovered from the actor row prefix at actual row +0x06. Because this export starts at actual row +0x08(HP), that is HP-view -0x02.", "HP-view +0x36 crosses into the next actual row prefix and reads the next monster's level. It is kept only as a boundary diagnostic and must not be used as the current monster level.", "One trailing name pointer, 폭호 id 0x0255, is present after the continuous stat rows. The next raw 0x38-sized block starts with 2000, matching the known special 폭호 HP variant, but it is kept as a candidate because +0x34 is not the expected byte-swapped name id.", "Sprite binding, formation, encounter, reward routing, and combat formulas are separate tables and are not promoted by this report.", "+0x04..+0x14 and +0x16..+0x27 are structured as candidate numeric/rate fields in this HP-view. Use battle_actor_stat_layout_review to map them to actor offsets.", ], } def js_json(value: Any) -> str: return json.dumps(value, ensure_ascii=False, separators=(",", ":")) def html_page(summary: dict[str, Any]) -> str: table = summary["table"] check_rows = "".join( "" f"{html.escape(key)}" f"{html.escape(str(value))}" "" for key, value in summary["checks"].items() ) row_html = "".join( "" f"{row['index']}" f"{html.escape(row['cleanName'])}" f"{row['level']}" f"{row['hp']}" f"{row['gold']}" f"{html.escape(row['vaHex'])}" f"{html.escape(row['nameIdHex'])}" f"{html.escape(row['frontNumericSummary'])}" f"{html.escape(row['rateByteSummary'])}" f"{html.escape(row['field2Hex'])}" f"{html.escape(row['wordHex'])}" "" for row in summary["rows"] ) name_only_rows = "".join( "" f"{html.escape(row['nameIdHex'])}" f"{html.escape(row['cleanName'])}" f"{html.escape(row['nameVaHex'])}" f"{html.escape('HP candidate ' + str(row.get('trailingCandidateBlock', {}).get('hpCandidate')) if row.get('trailingCandidateBlock') else '')}" "" for row in summary["nameOnlyEntries"] ) return "\n".join([ "", '', "", ' ', ' ', " Enemy Stat Table", " ", "", "", "

Enemy Stat Table

", f"

{html.escape(summary['scope'])}

", f"

Table {html.escape(table['statTableVaHex'])}, row size {html.escape(table['rowSizeHex'])}, rows {summary['rowCount']}.

", "

Checks

", f" {check_rows}
checkvalue
", "

Rows

", f" {row_html}
#nameLvHPgoldrow VAname idcandidate front fieldscandidate rate bytes+0x36 next Lvraw words
", "

Name-Only Entries

", f" {name_only_rows}
name idnamename VAcandidate
", "", "", "", ]) def write_outputs(summary: dict[str, Any], out_dir: Path) -> None: out_dir.mkdir(parents=True, exist_ok=True) (out_dir / "enemy_stat_table.json").write_text( json.dumps(summary, ensure_ascii=False, separators=(",", ":")) + "\n", encoding="utf-8", ) (out_dir / "enemy_stat_table.html").write_text(html_page(summary), encoding="utf-8") (out_dir / "enemy_stat_table.js").write_text( "window.HWANSE_ENEMY_STAT_TABLE = " + js_json({ "promotionStatus": summary["promotionStatus"], "scope": summary["scope"], "table": summary["table"], "rowCount": summary["rowCount"], "namePointerCount": summary["namePointerCount"], "nameOnlyCount": summary["nameOnlyCount"], "duplicateNames": summary["duplicateNames"], "fieldLayout": summary["fieldLayout"], "confirmedFieldCountPerRow": summary["confirmedFieldCountPerRow"], "candidateFrontNumericFieldCountPerRow": summary["candidateFrontNumericFieldCountPerRow"], "candidateRateByteFieldCountPerRow": summary["candidateRateByteFieldCountPerRow"], "checks": summary["checks"], "rows": summary["rows"], "nameOnlyEntries": summary["nameOnlyEntries"], "notes": summary["notes"], }) + ";\n", encoding="utf-8", ) def main() -> None: parser = argparse.ArgumentParser(description=__doc__) parser.add_argument("--exe", type=Path, default=EXE) parser.add_argument("--out-dir", type=Path, default=OUT) args = parser.parse_args() data = args.exe.read_bytes() summary = build_summary(data) write_outputs(summary, args.out_dir) print(f"wrote enemy stat table -> {args.out_dir / 'enemy_stat_table.html'}") if __name__ == "__main__": main()