专名本土化(信达雅):冻结译名表 + 渲染层,并修掉校验器的正则盲点

- 新增 dfannals/glossary.py(译名表)与 dfannals/localize.py(渲染层)
  人名=音译+绰号意译(Ral Fastenhatchets→拉尔·缚斧、朗古德·商熊「剖胎之寒」),
  地名整体意译(Fordedwinds→涉风渡、The Volcano Of Ravens→渡鸦火山)
- 生成链路本土化:素材渲染、人物表、扩写白名单全部走译名;已有产物做确定性替换
- 修复 factcheck 的 \b 盲点:Unicode 下汉字也算词字符,紧贴汉字的英文名不被检查
  (曾因此漏掉编造专名)。修好后抓出并定点修掉长版第 1 集的 2 处编造
- 修正 title_form 大小写(YETI→Yeti),此前 Yeti/Goblin 一直漏译
- 新增 tests/test_localize.py(13 项回归)
This commit is contained in:
Chen Yi
2026-10-05 22:39:53 +08:00
parent 8630dcad55
commit 16b476167e
14 changed files with 924 additions and 350 deletions
+208
View File
@@ -0,0 +1,208 @@
"""专名译名表(冻结)。
为什么要有这张表:正文里满屏 `Ral Fastenhatchets`、`Fordedwinds` 对中文读者很难读。
但如果交给模型每次现译,同一个角色会在第 3 集变成另一个名字,所以译名必须冻结在这里。
翻译策略(信达雅):
1. **人名 = 名字音译 + 姓氏/绰号意译**,用「·」连接。
这与《魔戒》把 Thorin Oakenshield 译作「索林·橡木盾」是同一套惯例,
读起来像中国人名,又保住了原名的意思。
2. **地名整体意译**。DF 地名同样是词表拼接(Fordedwinds = forded + winds),
直译成「涉风渡」比音译成「福德温兹」更像地名。
3. **称号(the X of Y)单独译**,DF 给传奇人物的长称号自带画面感,直译最好。
DF 的名字由词表拼接而成,所以同一批词根会在不同世界反复出现:
这张表按「整词」而非「词根」收词,避免把 Fateddie 硬拆成"命定+死"这类生硬堆叠。
遇到表里没有的名字,localize 会原样保留并登记进待译清单,而不是硬造一个。
"""
# ------------------------------------------------------------------ 名字词干(音译)
STEMS: dict[str, str] = {
"alath": "阿拉斯",
"amxu": "阿姆苏",
"ana": "安娜",
"artuk": "阿图克",
"asmel": "阿斯梅尔",
"astri": "阿斯特里",
"domas": "多马斯",
"dostngosp": "多斯特恩戈斯普",
"ducim": "杜西姆",
"espu": "埃斯普",
"evala": "埃瓦拉",
"guspu": "古斯普",
"id": "伊德",
"ineth": "伊内斯",
"ingish": "英吉什",
"langgud": "朗古德",
"lolor": "洛洛尔",
"meng": "蒙",
"moldath": "莫尔达斯",
"momuz": "莫穆兹",
"mosus": "莫苏斯",
"mudi": "穆迪",
"ngokang": "恩戈康",
"ngom": "恩戈姆",
"niri": "尼里",
"nithe": "尼瑟",
"ongu": "翁古",
"onul": "奥努尔",
"pari": "帕里",
"perom": "佩罗姆",
"ral": "拉尔",
"rigoth": "里戈斯",
"sathrel": "萨瑟雷尔",
"sayrsnis": "赛尔斯尼斯",
"shosa": "绍萨",
"sina": "西娜",
"someg": "索梅格",
"song": "松",
"stinthad": "斯廷哈德",
"tekkud": "泰库德",
"thrun": "斯伦",
"tirist": "提里斯特",
"tulon": "图隆",
"uknol": "乌克诺尔",
"usel": "乌塞尔",
"utes": "乌特斯",
"zuntir": "尊提尔",
"zutthan": "祖坦",
}
# ------------------------------------------------------------------ 姓氏/绰号(意译)
# 键是姓名后半段的拼接结果(小写、无空格)
SURNAMES: dict[str, str] = {
"baldedattic": "秃阁",
"basementrampart": "地垒",
"blazedbrushes": "焰刷",
"blottedmine": "污矿",
"bonestands": "骨立",
"bronzescorched": "铜焦",
"churchbrass": "堂铜",
"clappedbrass": "拍铜",
"crazeowned": "狂据",
"decentpage": "端页",
"dyesong": "染歌",
"fastenhatchets": "缚斧",
"fateddie": "命终",
"fencefinders": "篱寻",
"fisherfair": "渔市",
"flayerline": "剥血",
"frillyknots": "褶结",
"glimmeredhelps": "光援",
"glowroofs": "光顶",
"gleamcircled": "光环",
"growllie": "吼卧",
"hammershake": "锤摇",
"huggedfastened": "抱固",
"incenselyrics": "香诗",
"jackalbrims": "豺帽",
"matchedmenaces": "匹凶",
"menacedslinks": "危潜",
"mirroredfloors": "镜层",
"oilshove": "油推",
"palmsaved": "掌救",
"playfills": "戏填",
"ruinedhorrors": "毁怖",
"searcheddated": "寻期",
"squirtlocks": "喷锁",
"steelgrouped": "钢聚",
"stilledearthen": "静土",
"stolendipped": "窃浸",
"strikepaints": "击彩",
"talonhollows": "爪穴",
"theatercrested": "剧冠",
"thornticks": "棘蜱",
"tongswall": "钳墙",
"tradebears": "商熊",
"walkclasped": "行扣",
"watchedleads": "守铅",
"wheeleddrinks": "轮饮",
"wisdomskins": "智皮",
"yellplank": "啸板",
}
# ------------------------------------------------------------------ 地名(整体意译)
PLACES: dict[str, str] = {
"clampokes": "钳戳",
"fordedwinds": "涉风渡",
"fuchsiarumor": "紫谣",
"gloryholes": "荣耀穴",
"lashheroes": "鞭雄",
"leapdrinks": "跃饮",
"liecats": "卧猫",
"mortalpears": "殇梨",
"paperhushed": "纸寂",
"rainseduced": "雨诱",
"tomessoldiers": "典兵",
"toolfainted": "器昏",
"walkedbirds": "行鸟",
}
# ------------------------------------------------------------------ 称号(the X of Y)
EPITHETS: dict[str, str] = {
"the eviscerated chill of pregnancies": "剖胎之寒",
"the trickery of mornings": "晨之诡计",
}
# ------------------------------------------------------------------ 势力/组织
GROUPS: dict[str, str] = {
"the volcano of ravens": "渡鸦火山",
}
# ------------------------------------------------------------------ 种族
RACES: dict[str, str] = {
"AARDVARK_MAN": "土豚人",
"ALLIGATOR": "鳄鱼",
"ALLIGATOR_MAN": "鳄人",
"BAT_GIANT": "巨蝠",
"BEAR_BLACK": "黑熊",
"BEAR_GRIZZLY": "灰熊",
"BEAR_GRIZZLY_MAN": "灰熊人",
"BEAR_POLAR": "白熊",
"BIRD_OWL_GREAT_HORNED": "大雕鸮",
"BIRD_ROC": "大鹏",
"BIRD_WREN": "鹩鹪",
"BLIND_CAVE_BEAR": "盲穴熊",
"BLIND_CAVE_OGRE": "盲穴食人魔",
"BOBCAT": "山猫",
"CAMEL_1_HUMP_MAN": "单峰驼人",
"CARDINAL_MAN": "红衣主教雀人",
"CAVE_DRAGON": "穴龙",
"CHEETAH": "猎豹",
"CHINCHILLA_MAN": "绒鼠人",
"COLOSSUS_BRONZE": "青铜巨像",
"COUGAR": "美洲狮",
"CYCLOPS": "独眼巨人",
"DOG": "犬",
"DWARF": "矮人",
"EAGLE_MAN": "鹰人",
"ELEPHANT": "象",
"ELF": "精灵",
"FALCON": "隼族",
"GOAT_MOUNTAIN_MAN": "山羊山人",
"GOBLIN": "哥布林",
"HUMAN": "人类",
"HYDRA": "九头蛇",
"JAGUAR": "美洲豹",
"KOBOLD": "狗头人",
"LEOPARD": "豹",
"LION": "狮",
"TIGER": "虎",
"TROLL": "巨魔",
"WOLF": "狼",
"YETI": "雪人",
}
# ------------------------------------------------------------------ 地点类型
SITE_TYPES: dict[str, str] = {
"cave": "洞穴",
"dark fortress": "黑堡",
"dark pits": "幽坑",
"forest retreat": "林居",
"fortress": "要塞",
"hamlet": "村落",
"hillocks": "丘屯",
"tomb": "陵墓",
"town": "城镇",
}