#!/usr/bin/env python3
# 完整抽取 l10n → 单源 JSON:getters(String get) + lists(List<String> get) + methods(带参)。
# 只读 Dart,不改。与已提交 strings.json 的 getters/methods 校验无回归,新增 lists 段。
import re, json, os

BASE = "/Users/wangjia/code/pangolin/.claude/worktrees/configurable-proxy/client/lib/l10n"
OUT = "/Users/wangjia/code/pangolin/.claude/worktrees/configurable-proxy/design/i18n/strings.json"
LANGS = ["zh", "en", "ja", "ko", "ru", "es"]

STR = r"""(?:'((?:[^'\\]|\\.)*)'|"((?:[^"\\]|\\.)*)")"""
STR_RE = re.compile(STR)
GET_RE = re.compile(r"String\s+get\s+(\w+)\s*=>\s*" + STR + r"\s*;")
MET_RE = re.compile(r"String\s+(\w+)\(([^)]*)\)\s*=>\s*" + STR + r"\s*;")
# List<String> get name => const [ ... ];  —— 跨行,惰性到第一个 '];'
LIST_RE = re.compile(r"List<String>\s+get\s+(\w+)\s*=>\s*const\s*\[(.*?)\]\s*;", re.DOTALL)


def unescape(s):
    return (s.replace("\\'", "'").replace('\\"', '"')
             .replace('\\n', '\n').replace('\\$', '$').replace('\\\\', '\\'))


getters, lists, methods, counts = {}, {}, {}, {}
for lang in LANGS:
    src = open(os.path.join(BASE, f"strings_{lang}.dart"), encoding="utf-8").read()
    g = m = ln = 0
    for mt in GET_RE.finditer(src):
        k = mt.group(1)
        v = unescape(mt.group(2) if mt.group(2) is not None else mt.group(3))
        getters.setdefault(k, {})[lang] = v; g += 1
    for mt in MET_RE.finditer(src):
        k = mt.group(1); args = mt.group(2).strip()
        v = unescape(mt.group(3) if mt.group(3) is not None else mt.group(4))
        methods.setdefault(k, {"args": args, "tpl": {}})
        methods[k]["tpl"][lang] = v; m += 1
    for mt in LIST_RE.finditer(src):
        k = mt.group(1); body = mt.group(2)
        items = [unescape(a if a is not None else b) for a, b in STR_RE.findall(body)]
        lists.setdefault(k, {})[lang] = items; ln += 1
    counts[lang] = {"getters": g, "methods": m, "lists": ln}

# 完整性:每 key 六语齐全
missing = []
for k, d in getters.items():
    for l in LANGS:
        if l not in d: missing.append(f"getter {k}:{l}")
for k, d in lists.items():
    for l in LANGS:
        if l not in d: missing.append(f"list {k}:{l}")
for k, d in methods.items():
    for l in LANGS:
        if l not in d["tpl"]: missing.append(f"method {k}:{l}")

# 与已提交 JSON 的 getters/methods 校验无回归
old = json.load(open(OUT, encoding="utf-8")) if os.path.exists(OUT) else {}
regress = []
if old:
    if set(old.get("getters", {})) != set(getters):
        regress.append("getter 键集变化")
    for k, d in old.get("getters", {}).items():
        if getters.get(k) != d: regress.append(f"getter 值变化:{k}")
    if set(old.get("methods", {})) != set(methods):
        regress.append("method 键集变化")

out = {"_note": "l10n 单源(唯一真相源)。design/codegen/gen_l10n_dart.mjs 生成 client/lib/l10n/strings_*.dart;"
                 "原型 serve.mjs 读同源。勿手改生成物;abstract 声明由 app_text.dart 手写(编译器强制 concrete⊆abstract)。",
       "getters": getters, "lists": lists, "methods": methods}
json.dump(out, open(OUT, "w", encoding="utf-8"), ensure_ascii=False, indent=2)

print("每语言:", counts)
print(f"合并:getters={len(getters)} lists={len(lists)} methods={len(methods)}")
print(f"缺失(应0):{len(missing)}", missing[:5])
print(f"回归(应0):{len(regress)}", regress[:5])
print("lists 键:", list(lists))
print("输出:", OUT)