lucky_shop/scripts/i18n_build_zhcn.py

"""
i18n_build_zhcn.py - 用真实中文 (docs/i18n-real-zh.json) 重建 lang/zh-cn.js

- key -> 真实中文 (来自 f258987 .vue 硬编码)
- 保留现有 pageTitles (用户说 pageTitles 用现在值即可)
- en-us 端留空 stub 保持对称
"""
import re
import json
import sys
from pathlib import Path
from collections import defaultdict

ROOT = Path(__file__).resolve().parent.parent
EN_FILE = ROOT / 'lang' / 'en-us.js'
ZH_FILE = ROOT / 'lang' / 'zh-cn.js'
REAL_ZH = ROOT / 'docs' / 'i18n-real-zh.json'
EXISTING_ZH = ROOT / 'docs' / 'i18n-existing-zh.json'


def parse_existing_keys(path):
    """读现有 lang/<file>.js 的所有 key -> value 映射"""
    if not path.exists():
        return {}
    text = path.read_bytes().decode('utf-8', errors='ignore')
    out = {}
    indent_stack = [(-1, '')]
    for ln in text.split('\n'):
        if ln.strip().startswith('//') or not ln.strip(): continue
        indent = len(ln) - len(ln.lstrip())
        m = re.match(r"""^['\"]?([A-Za-z_][A-Za-z0-9_.]*)['\"]?\s*:\s*(.*)$""", ln.lstrip())
        if not m: continue
        key, rest = m.group(1), m.group(2).strip().rstrip(',')
        while len(indent_stack) > 1 and indent_stack[-1][0] >= indent: indent_stack.pop()
        parent = indent_stack[-1][1]
        full_key = f'{parent}.{key}' if parent else key
        if rest.endswith('{'): indent_stack.append((indent, full_key)); continue
        v = rest
        if (v.startswith('"') and v.endswith('"')) or (v.startswith("'") and v.endswith("'")):
            if len(v) >= 2: v = v[1:-1]
        out[full_key] = v
    return out


def write_lang(path, d, locale):
    """写带双引号 key 的 lang/<locale>.js, UTF-8"""
    grouped = defaultdict(dict)
    top_level = {}
    for k, v in d.items():
        parts = k.split('.')
        if len(parts) == 1:
            top_level[k] = v
        else:
            ns = parts[0]
            sub = '.'.join(parts[1:])
            grouped[ns][sub] = v

    lines = [
        f'// Auto-generated by i18n_build_zhcn.py',
        f'// Single authoritative lang source. Do not edit manually.',
        f'// locale: {locale}',
        f'// Generated: 2026-06-23',
        '',
        'export const lang = {',
    ]
    written = set()
    nested_ns = {'pageTitles', 'common', 'tabBar', 'customer'}
    for ns in sorted(set(grouped.keys()) | set(top_level.keys())):
        if ns in nested_ns and ns in grouped:
            lines.append(f'\t"{ns}": {{')
            for sub_k in sorted(grouped[ns].keys()):
                sub_v = grouped[ns][sub_k]
                escaped = sub_v.replace('\\', '\\\\').replace('"', '\\"')
                lines.append(f'\t\t"{sub_k}": "{escaped}",')
            lines.append('\t},')
            written.add(ns)
    for ns in sorted(grouped.keys()):
        if ns in written: continue
        lines.append(f'\t"{ns}": {{')
        for sub_k in sorted(grouped[ns].keys()):
            sub_v = grouped[ns][sub_k]
            escaped = sub_v.replace('\\', '\\\\').replace('"', '\\"')
            lines.append(f'\t\t"{sub_k}": "{escaped}",')
        lines.append('\t},')
        written.add(ns)
    for k in sorted(top_level.keys()):
        if k in written: continue
        v = top_level[k]
        escaped = v.replace('\\', '\\\\').replace('"', '\\"')
        lines.append(f'\t"{k}": "{escaped}",')

    lines.append('};')
    lines.append('')
    content = '\n'.join(lines)
    path.write_bytes(content.encode('utf-8'))


def main():
    # 1. 真实中文 (从 f258987 .vue 硬编码提取)
    real_zh = json.loads(REAL_ZH.read_text(encoding='utf-8'))
    print(f'真实中文: {len(real_zh)} 条')

    # 2. 现有 pageTitles 保留 (用户要求)
    # 解析当前 zh-cn.js, 取 pageTitles 部分
    current_zh = parse_existing_keys(ZH_FILE)
    page_titles = {k: v for k, v in current_zh.items() if k.startswith('pageTitles.')}
    print(f'pageTitles 保留: {len(page_titles)} 条')

    # 3. 现有 en-us 端英文 (用作 en-us 基础, 保留已有的英文翻译)
    current_en = parse_existing_keys(EN_FILE)
    print(f'现有 en-us keys: {len(current_en)}')

    # 4. 构造新的 zh-cn
    new_zh = {}
    # 先 pageTitles
    for k, v in page_titles.items():
        new_zh[k] = v
    # 再真实中文 (新 key) - 来自 f258987 硬编码
    for text, new_key in real_zh.items():
        new_zh[new_key] = text
    # 再补充: f258987 时 .vue 已有的 $lang 调用 (key 直接用原始 key)
    if EXISTING_ZH.exists():
        existing = json.loads(EXISTING_ZH.read_text(encoding='utf-8'))
        for k, v in existing.items():
            if k not in new_zh:
                new_zh[k] = v
        print(f'已补 {len(existing)} 个 f258987 已有的 $lang 值')
    # 保留现有 zh-cn 里非 pageTitles 的 key
    for k, v in current_zh.items():
        if k.startswith('pageTitles.'): continue
        if k not in new_zh:
            new_zh[k] = v
    print(f'new zh-cn keys: {len(new_zh)}')

    # 5. 构造新的 en-us: 保留已有, 缺中文对应的留空
    new_en = {}
    for k, v in current_en.items():
        if k.startswith('pageTitles.'): continue
        new_en[k] = v
    # 真实中文对应的 key, 英文侧: 优先保留已有, 缺则空
    for new_key in real_zh.values():
        if new_key not in new_en:
            new_en[new_key] = ''
    # 已有 $lang 对应的 key 留空 (中文侧已有, en 端缺)
    if EXISTING_ZH.exists():
        existing = json.loads(EXISTING_ZH.read_text(encoding='utf-8'))
        for k in existing.keys():
            if k not in new_en:
                new_en[k] = ''
    # pageTitles 保留
    for k, v in page_titles.items():
        new_en[k] = v
    print(f'new en-us keys: {len(new_en)}')

    # 6. 写
    apply = '--apply' in sys.argv
    if not apply:
        print('\n[DRY-RUN] 加 --apply 真写文件')
        return 0

    write_lang(ZH_FILE, new_zh, 'zh-cn')
    print(f'[OK] {ZH_FILE.relative_to(ROOT)}: {len(new_zh)} keys')
    write_lang(EN_FILE, new_en, 'en-us')
    print(f'[OK] {EN_FILE.relative_to(ROOT)}: {len(new_en)} keys')
    return 0


if __name__ == '__main__':
    sys.exit(main())