#!/usr/bin/env python3
# 从 6 份手写 strings_*.dart 抽取 l10n 到单源 JSON。只读,不改 Dart。
import re, json, sys, os

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

# Dart 字符串字面量:单引号或双引号(各自允许 \ 转义);getter 与带参方法
STR = r"""(?:'((?:[^'\\]|\\.)*)'|"((?:[^"\\]|\\.)*)")"""
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*;")

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

getters = {}   # key -> {lang: val}
methods = {}   # key -> {args, tpl:{lang:val}}
counts = {}

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

# 完整性校验:每个 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 methods.items():
    for l in LANGS:
        if l not in d["tpl"]: missing.append(f"method {k}:{l}")

out = {"_note": "l10n 单源(唯一真相源)。codegen 生成 client/lib/l10n/strings_*.dart + app_text.dart + 原型读同源。勿手改生成物。",
       "getters": getters, "methods": methods}
os.makedirs("/Users/wangjia/code/pangolin/.claude/worktrees/configurable-proxy/design/i18n", exist_ok=True)
outpath = "/Users/wangjia/code/pangolin/.claude/worktrees/configurable-proxy/design/i18n/strings.json"
json.dump(out, open(outpath, "w", encoding="utf-8"), ensure_ascii=False, indent=2)

print("每语言抽取数:", counts)
print(f"合并后:getters={len(getters)} methods={len(methods)}")
print(f"缺失(应为0):{len(missing)}", missing[:5])
print("输出:", outpath)