DazedTL/util/wolfdawn/names.py
2026-07-09 08:04:53 -05:00

610 lines
21 KiB
Python

"""WolfDawn ``names.json`` helpers for the WOLF translation workflow.
WolfDawn's ``names-extract`` produces ``names.json``: a project-wide glossary of
every value name the game references. Each entry has a ``source`` / ``text`` pair,
a ``note`` field (the WOLF database category it came from), and a static
``safety`` badge from WolfDawn's command-stream analysis:
* ``safe`` - display-only; no command references the string by name.
* ``refs`` - referenced by name (picture codes, switch labels, etc.); skipped.
* ``verify`` - also appears in indirect literals; left untranslated by default.
:mod:`modules.wolfdawn` translates only entries whose badge is ``safe``.
``refs``, ``verify``, and legacy entries without a badge keep ``text == source``
so injection is a no-op for them.
Phase 0 still harvests translated **short name-like** entries into ``vocab.txt``,
grouped by ``note``. Multi-line profile blurbs, dialogue, and other content-shaped
names are translated in ``names.json`` but omitted from the glossary.
"""
from __future__ import annotations
import json
import re
from dataclasses import dataclass, field
from pathlib import Path
from typing import Any, Union
PathLike = Union[str, Path]
SAFETY_SAFE = "safe"
SAFETY_REFS = "refs"
SAFETY_VERIFY = "verify"
_TRANSLATABLE_SAFETY = frozenset({SAFETY_SAFE})
# names.json ``note`` values that are content fields, not short display names.
_VOCAB_SKIP_NOTE_RE = re.compile(
r"プロフィール|説明|セリフ|台詞|会話|自己紹介|挨拶|モノローグ|独白",
)
# Short row labels and item/skill names fit; multi-line blurbs do not.
_VOCAB_HARVEST_MAX_LEN = 72
# WolfDawn labels standard database types bilingually in a document's group
# ``typeName`` (e.g. ``"Weapon · 武器"``). names.json ``note`` values are the JP
# half only, so this static map provides an English label for the common categories
# when the live DB labels are unavailable.
NOTE_EN: dict[str, str] = {
"武器": "Weapon",
"防具": "Armor",
"技能": "Skill",
"アイテム": "Item",
"状態設定": "State Setting",
"用語設定": "Term Setting",
"戦闘コマンド": "Battle Command",
"システム設定": "System Setting",
"属性名の設定": "Attribute",
"主人公ステータス": "Hero Status",
"敵グループ": "Enemy Group",
"敵キャラ個体データ": "Enemy Data",
}
_LABEL_SEP = " · "
def parse_names_entries(data: dict[str, Any]) -> list[dict[str, Any]]:
names = data.get("names", [])
if not isinstance(names, list):
return []
return [entry for entry in names if isinstance(entry, dict)]
def has_safety_metadata(data: dict[str, Any]) -> bool:
"""True when *data* includes WolfDawn per-name ``safety`` badges."""
return any("safety" in entry for entry in parse_names_entries(data))
def entry_safety(entry: dict[str, Any]) -> str:
"""Return the normalised ``safety`` badge for one names.json entry."""
return str(entry.get("safety", "")).strip().lower()
def is_name_translatable(entry: dict[str, Any]) -> bool:
"""True when WolfDawn marks *entry* as safe (per-name, not per-category)."""
return entry_safety(entry) in _TRANSLATABLE_SAFETY
def is_vocab_harvest_candidate(entry: dict[str, Any]) -> bool:
"""True when a translated names.json entry should seed ``vocab.txt``.
Translation uses :func:`is_name_translatable`; this is stricter. Profile text,
multi-line blurbs, and dialogue-shaped strings are still translated in
``names.json`` but are not glossary terms.
"""
if not is_name_translatable(entry):
return False
src = str(entry.get("source", ""))
if not src.strip():
return False
if "\n" in src or "\r" in src:
return False
if len(src) > _VOCAB_HARVEST_MAX_LEN:
return False
note = str(entry.get("note", ""))
if _VOCAB_SKIP_NOTE_RE.search(note):
return False
# Two or more sentence endings usually means dialogue, not a row label.
if len(re.findall(r"[。!?]", src)) >= 2:
return False
return True
def count_name_safety(data: dict[str, Any]) -> dict[str, int]:
"""Return per-badge counts for a names.json document."""
counts = {
"total": 0,
"safe": 0,
"refs": 0,
"verify": 0,
"unknown": 0,
"translatable": 0,
}
for entry in parse_names_entries(data):
counts["total"] += 1
safety = entry_safety(entry)
if safety == SAFETY_SAFE:
counts["safe"] += 1
elif safety == SAFETY_REFS:
counts["refs"] += 1
elif safety == SAFETY_VERIFY:
counts["verify"] += 1
else:
counts["unknown"] += 1
counts["translatable"] = counts["safe"]
return counts
def format_name_safety_summary(data: dict[str, Any]) -> str:
"""One-line summary for the workflow UI."""
counts = count_name_safety(data)
if not counts["total"]:
return "No names found in names.json."
if not has_safety_metadata(data):
return (
f"{counts['total']} name(s), but no WolfDawn safety badges - re-run names-extract "
"with a current WolfDawn build before Phase 0."
)
parts = []
if counts["safe"]:
parts.append(f"{counts['safe']} safe")
if counts["refs"]:
parts.append(f"{counts['refs']} refs")
if counts["verify"]:
parts.append(f"{counts['verify']} verify")
badge_text = ", ".join(parts) if parts else "no badges"
return (
f"{counts['translatable']} of {counts['total']} name(s) will translate "
f"({badge_text}). Refs and verify names are skipped."
)
def derive_db_labels(files_dir) -> dict[str, str]:
"""Build a ``{japanese_note: english}`` map from staged WolfDawn ``db`` files."""
from pathlib import Path
labels: dict[str, str] = {}
try:
base = Path(files_dir)
if not base.is_dir():
return labels
for path in sorted(base.glob("*.project.json")):
try:
doc = json.loads(path.read_text(encoding="utf-8-sig"))
except Exception:
continue
if doc.get("kind") != "db":
continue
for group in doc.get("groups") or []:
type_name = group.get("typeName") or ""
if _LABEL_SEP in type_name:
english, _, japanese = type_name.partition(_LABEL_SEP)
english, japanese = english.strip(), japanese.strip()
if english and japanese and japanese not in labels:
labels[japanese] = english
except Exception:
pass
return labels
def note_header(note: str, db_labels: dict[str, str] | None = None) -> str:
"""Return a bilingual vocab section header for a names.json ``note``."""
english = (db_labels or {}).get(note) or NOTE_EN.get(note)
if english and english != note:
return f"{english}{_LABEL_SEP}{note}"
return note
def format_note_category_summary(
data: dict[str, Any],
*,
top_n: int = 8,
db_labels: dict[str, str] | None = None,
) -> str:
"""Multi-line breakdown of top ``note`` categories for the Names workflow step."""
from collections import Counter
counts: Counter[str] = Counter()
for entry in parse_names_entries(data):
note = str(entry.get("note", "")).strip()
if note:
counts[note] += 1
if not counts:
return "No note categories found."
lines = [f"Top categories in this game ({len(counts)} unique):"]
for note, count in counts.most_common(top_n):
header = note_header(note, db_labels)
lines.append(f" {count:4d} {header}")
remaining = len(counts) - min(top_n, len(counts))
if remaining > 0:
lines.append(f" … and {remaining} more categor{'y' if remaining == 1 else 'ies'}")
return "\n".join(lines)
def collect_name_notes(data: dict[str, Any]) -> list[str]:
"""Return sorted unique ``note`` values from a names.json document."""
notes: set[str] = set()
for entry in parse_names_entries(data):
note = str(entry.get("note", "")).strip()
if note:
notes.add(note)
return sorted(notes)
def apply_names_wrap(
names_path: PathLike,
*,
note: str | None = None,
width: int | None = None,
lines: int | None = None,
dry_run: bool = False,
log_fn=None,
) -> tuple[bool, str]:
"""Run ``wolf names-wrap`` on one names.json file.
When *note* is set, only that category (``names[].note``) is processed.
Omit *width* / *lines* to measure slot geometry from JP text (recommended).
Returns ``(ok, summary)``. *ok* is False only when the CLI exits with a hard
error; overflow warnings still count as success.
"""
from util import wolfdawn
emit = log_fn or (lambda _msg: None)
path = Path(names_path)
if not path.is_file():
return False, f"names.json not found: {path}"
res = wolfdawn.names_wrap(
path,
note=note,
width=width,
lines=lines,
dry_run=dry_run,
log_fn=None,
)
for line in (res.stdout or "").splitlines():
line = line.strip()
if line:
emit(line)
for line in (res.stderr or "").splitlines():
line = line.strip()
if line:
emit(line)
if not res.ok and res.returncode not in (0, 2):
return False, f"names-wrap exited {res.returncode}"
count = wolfdawn.parse_names_wrap_counts(res.stdout, res.stderr)
shrunk = wolfdawn.parse_names_wrap_shrink_count(res.stdout, res.stderr)
scope = f"category {note}" if note else "names.json"
parts: list[str] = []
if dry_run:
if count is not None:
parts.append(
f"dry run: {count} entr{'y' if count == 1 else 'ies'} in {scope} "
"would be re-wrapped"
)
else:
parts.append(f"dry run complete for {scope}")
elif count is not None:
parts.append(
f"re-wrapped {count} entr{'y' if count == 1 else 'ies'} in {scope}"
)
else:
parts.append(f"names-wrap complete for {scope}")
if width is not None and int(width) > 0:
parts.append(f"width={int(width)}")
if lines is not None and int(lines) > 0:
parts.append(f"lines={int(lines)}")
if shrunk:
parts.append(f"{shrunk} shrunk with \\f[N]")
return True, "; ".join(parts)
def roles_json_path_for_names(names_path: PathLike) -> Path:
"""Return ``wolfdawn-roles.json`` beside *names_path* (WolfDawn looks there)."""
return Path(names_path).resolve().parent / "wolfdawn-roles.json"
def load_name_wrap_roles(roles_path: PathLike) -> dict[str, Any]:
"""Load ``wolfdawn-roles.json``, or an empty document."""
path = Path(roles_path)
if not path.is_file():
return {}
try:
data = json.loads(path.read_text(encoding="utf-8-sig"))
return data if isinstance(data, dict) else {}
except Exception:
return {}
def get_name_wrap_role(roles: dict[str, Any], note: str) -> dict[str, Any] | None:
"""Return the ``nameWraps`` entry matching *note*, if any."""
wraps = roles.get("nameWraps")
if not isinstance(wraps, list):
return None
for entry in wraps:
if not isinstance(entry, dict):
continue
entry_note = str(entry.get("note") or "")
if entry_note == note or (note and note in entry_note) or (entry_note and entry_note in note):
return entry
return None
def upsert_name_wrap_role(
roles_path: PathLike,
note: str,
*,
width: int | None = None,
max_lines: int | None = None,
font: int | None = None,
min_font: int | None = None,
) -> Path:
"""Create/update a ``nameWraps`` rule for *note* in ``wolfdawn-roles.json``.
WolfDawn reads this file from the same directory as ``names.json``.
"""
path = Path(roles_path)
data = load_name_wrap_roles(path)
wraps = data.get("nameWraps")
if not isinstance(wraps, list):
wraps = []
data["nameWraps"] = wraps
entry = None
for item in wraps:
if isinstance(item, dict) and str(item.get("note") or "") == note:
entry = item
break
if entry is None:
entry = {"note": note}
wraps.append(entry)
if width is not None and int(width) > 0:
entry["width"] = int(width)
if max_lines is not None and int(max_lines) > 0:
entry["maxLines"] = int(max_lines)
if font is not None and int(font) > 0:
entry["font"] = int(font)
if min_font is not None and int(min_font) > 0:
entry["minFont"] = int(min_font)
path.parent.mkdir(parents=True, exist_ok=True)
path.write_text(
json.dumps(data, ensure_ascii=False, indent=2) + "\n",
encoding="utf-8",
)
return path
# ── Name-consistency reconcile (names.json wins) ─────────────────────────────
@dataclass
class NameReconcileChange:
"""One extract line rewritten to match the chosen canonical translation."""
json_file: str
source: str
old_text: str
new_text: str
reason: str # glossary | majority
@dataclass
class NameReconcileReport:
changes: list[NameReconcileChange] = field(default_factory=list)
skipped_identity_glossary: list[str] = field(default_factory=list)
unresolved: list[str] = field(default_factory=list)
files_written: list[str] = field(default_factory=list)
@property
def changed(self) -> int:
return len(self.changes)
def glossary_map(names_doc: dict[str, Any]) -> dict[str, str]:
"""``source -> text`` from names.json (last entry wins if duplicates)."""
out: dict[str, str] = {}
for entry in parse_names_entries(names_doc):
src = entry.get("source")
txt = entry.get("text")
if isinstance(src, str) and src and isinstance(txt, str):
out[src] = txt
return out
def _walk_source_text_leaves(obj: Any, visit) -> None:
"""Call ``visit(leaf_dict)`` for every object with string source+text."""
if isinstance(obj, dict):
src, txt = obj.get("source"), obj.get("text")
if isinstance(src, str) and isinstance(txt, str):
visit(obj)
for value in obj.values():
_walk_source_text_leaves(value, visit)
elif isinstance(obj, list):
for item in obj:
_walk_source_text_leaves(item, visit)
def _collect_extract_variants(
extract_docs: dict[str, dict[str, Any]],
glossary_sources: set[str],
) -> dict[str, dict[str, int]]:
"""source -> {text: count} for non-identity glossary-name lines in extracts."""
from collections import Counter
variants: dict[str, Counter[str]] = {}
for doc in extract_docs.values():
def visit(leaf: dict[str, Any], _doc=doc) -> None:
src = leaf["source"]
txt = leaf["text"]
if src not in glossary_sources or txt == src:
return
variants.setdefault(src, Counter())[txt] += 1
_walk_source_text_leaves(doc, visit)
return {src: dict(counts) for src, counts in variants.items()}
def choose_canonical(
source: str,
glossary_text: str,
extract_variants: dict[str, int],
) -> tuple[str | None, str]:
"""Pick the canonical English for *source*.
Priority:
1. Translated ``names.json`` entry (``glossary_text != source``)
2. Majority among extract variants (tie → lexicographically first)
3. None if glossary is still JP and there are no extract variants
"""
if glossary_text != source:
return glossary_text, "glossary"
if not extract_variants:
return None, "identity_glossary"
best_count = max(extract_variants.values())
candidates = sorted(t for t, n in extract_variants.items() if n == best_count)
return candidates[0], "majority"
def reconcile_names_to_glossary(
names_doc: dict[str, Any],
extract_docs: dict[str, dict[str, Any]],
*,
dry_run: bool = False,
) -> NameReconcileReport:
"""Rewrite extract lines so each glossary name uses one canonical translation.
*names_doc* is not modified. Only exact full-string ``source`` matches are
touched (same rule as WolfDawn ``names-check``).
"""
report = NameReconcileReport()
gloss = glossary_map(names_doc)
if not gloss:
return report
variants = _collect_extract_variants(extract_docs, set(gloss))
canonical: dict[str, tuple[str, str]] = {}
for source, gtext in gloss.items():
extract_texts = variants.get(source, {})
chosen, reason = choose_canonical(source, gtext, extract_texts)
if chosen is None:
continue
if reason == "majority" and len(extract_texts) < 2 and gtext == source:
# Glossary still JP and extracts already agree - nothing to reconcile.
continue
if reason == "glossary":
# Always apply glossary EN onto extracts (identity or wrong EN).
canonical[source] = (chosen, reason)
elif len(extract_texts) >= 2:
# Divergent extracts, glossary still JP - unify to majority.
canonical[source] = (chosen, reason)
# else: glossary JP, single extract variant - leave alone
for json_name, doc in extract_docs.items():
changed_here = 0
def visit(leaf: dict[str, Any], _name=json_name) -> None:
nonlocal changed_here
src = leaf["source"]
if src not in canonical:
return
chosen, reason = canonical[src]
old = leaf["text"]
if old == chosen:
return
# When unifying via majority (glossary still JP), do not invent EN onto
# identity lines - only collapse divergent translations.
if reason == "majority" and old == src:
return
report.changes.append(
NameReconcileChange(
json_file=_name,
source=src,
old_text=old,
new_text=chosen,
reason=reason,
)
)
if not dry_run:
leaf["text"] = chosen
changed_here += 1
_walk_source_text_leaves(doc, visit)
if changed_here and not dry_run:
report.files_written.append(json_name)
return report
def format_reconcile_summary(report: NameReconcileReport) -> str:
"""Short human summary for the GUI / log."""
if not report.changes:
return "Name reconcile: nothing to change."
parts = [f"Name reconcile: {report.changed} line(s) updated"]
by_reason: dict[str, int] = {}
for c in report.changes:
by_reason[c.reason] = by_reason.get(c.reason, 0) + 1
if by_reason:
detail = ", ".join(f"{n} via {r}" for r, n in sorted(by_reason.items()))
parts.append(f"({detail})")
if report.files_written:
parts.append(f"in {len(report.files_written)} file(s)")
majority_sources = sorted({c.source for c in report.changes if c.reason == "majority"})
if majority_sources:
parts.append(
f"; {len(majority_sources)} name(s) used majority because names.json "
"is still Japanese (refs/verify) - translate them in names.json to lock the winner"
)
return " ".join(parts)
def reconcile_translated_dir(
translated_dir: PathLike,
*,
dry_run: bool = False,
) -> NameReconcileReport:
"""Load ``translated/names.json`` + other JSON extracts and reconcile in place."""
base = Path(translated_dir)
names_path = base / "names.json"
if not names_path.is_file():
report = NameReconcileReport()
report.unresolved.append("names.json missing")
return report
names_doc = json.loads(names_path.read_text(encoding="utf-8-sig"))
if not isinstance(names_doc, dict):
report = NameReconcileReport()
report.unresolved.append("names.json invalid")
return report
extract_docs: dict[str, dict[str, Any]] = {}
for path in sorted(base.glob("*.json")):
if path.name in ("names.json", "wolfdawn-roles.json", "wrap_profile.json"):
continue
try:
data = json.loads(path.read_text(encoding="utf-8-sig"))
except Exception:
continue
if isinstance(data, dict) and data.get("kind") in (
"db",
"map",
"common",
"gamedat",
"txt",
"txt-dir",
):
extract_docs[path.name] = data
report = reconcile_names_to_glossary(names_doc, extract_docs, dry_run=dry_run)
if not dry_run:
for name in report.files_written:
path = base / name
path.write_text(
json.dumps(extract_docs[name], ensure_ascii=False, indent=4) + "\n",
encoding="utf-8",
)
return report