#!/usr/bin/env python3
"""
Merge i18n translations from all pipeline modules into wwwroot/i18n.

模块自动发现（不再手工维护清单——2026-09-12 i18n 治理：手工清单漏了 17 个模块，
词条永远进不了字典）：
  1. 主应用自身 i18n/
  2. pipeline_core / pipeline_ops（本地或 pkgs/）
  3. build.sh clone 清单里的全部业务模块（pkgs/<mod> 或同级仓库目录）
  4. pkgs/ 下任何带 i18n/ 目录的其它模块（兜底扫描）

每模块读 i18n/<lang>/msg.txt 与 i18n/<lang>/i18n.json，先到先得不覆盖。

Usage: python scripts/merge_i18n.py
"""
import os, json, re, sys
from collections import OrderedDict

SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
ROOT_DIR = os.path.dirname(SCRIPT_DIR)
WWWROOT = os.path.join(ROOT_DIR, 'wwwroot')
I18N_OUT = os.path.join(WWWROOT, 'i18n')
LANGS = ['zh', 'en', 'ko', 'jp']


def discover_modules():
    """返回 [(name, i18n_dir), ...]，顺序即合并优先级（主应用最先）。"""
    mods = []

    def add(name, i18n_dir):
        if os.path.isdir(i18n_dir) and not any(n == name for n, _ in mods):
            mods.append((name, i18n_dir))

    # 1. 主应用
    add('app', os.path.join(ROOT_DIR, 'i18n'))

    # 2/3. build.sh clone 清单 + 本地业务模块
    names = ['pipeline_core', 'pipeline_ops']
    build_sh = os.path.join(ROOT_DIR, 'build.sh')
    if os.path.exists(build_sh):
        content = open(build_sh, encoding='utf-8').read()
        for m in re.finditer(r'^for mod in ([^;]+); do\s*$', content, re.M):
            for n in m.group(1).split():
                if n not in names:
                    names.append(n)
    repos_parent = os.path.dirname(ROOT_DIR)
    for n in names:
        for base in (os.path.join(ROOT_DIR, 'pkgs'), ROOT_DIR, repos_parent,
                     os.path.join(ROOT_DIR, n)):
            cand = os.path.join(base, n, 'i18n') if not base.endswith(n) else os.path.join(base, 'i18n')
            if os.path.isdir(cand):
                add(n, cand)
                break

    # 4. 兜底：pkgs/ 下所有带 i18n/ 的目录
    pkgs = os.path.join(ROOT_DIR, 'pkgs')
    if os.path.isdir(pkgs):
        for n in sorted(os.listdir(pkgs)):
            add(n, os.path.join(pkgs, n, 'i18n'))
    return mods


def parse_msg_txt(filepath):
    """Parse msg.txt: 'key=value' or 'key: value' format."""
    result = OrderedDict()
    if not os.path.exists(filepath):
        return result
    with open(filepath, 'r', encoding='utf-8') as f:
        for line in f:
            line = line.strip()
            if not line or line.startswith('#'):
                continue
            if '=' in line:
                key, val = line.split('=', 1)
                result[key.strip()] = val.strip()
            elif ':' in line:
                key, val = line.split(':', 1)
                result[key.strip()] = val.strip()
    return result


def merge():
    modules = discover_modules()
    print(f'Discovered {len(modules)} i18n sources: {[n for n, _ in modules]}')
    os.makedirs(I18N_OUT, exist_ok=True)

    for lang in LANGS:
        merged = OrderedDict()
        out_dir = os.path.join(I18N_OUT, lang)
        os.makedirs(out_dir, exist_ok=True)

        # Load existing merged translations
        i18n_file = os.path.join(out_dir, 'i18n.json')
        if os.path.exists(i18n_file):
            with open(i18n_file, 'r', encoding='utf-8') as f:
                merged = json.load(f, object_pairs_hook=OrderedDict)

        for mod_name, mod_i18n_dir in modules:
            lang_dir = os.path.join(mod_i18n_dir, lang)
            if not os.path.isdir(lang_dir):
                continue
            # msg.txt
            msgs = parse_msg_txt(os.path.join(lang_dir, 'msg.txt'))
            for k, v in msgs.items():
                if k not in merged:
                    merged[k] = v
            # i18n.json
            mod_i18n = os.path.join(lang_dir, 'i18n.json')
            if os.path.exists(mod_i18n):
                try:
                    with open(mod_i18n, 'r', encoding='utf-8') as f:
                        extra = json.load(f, object_pairs_hook=OrderedDict)
                except Exception as e:
                    print(f'  !! {mod_name}/{lang}/i18n.json parse error: {e}', file=sys.stderr)
                    continue
                for k, v in extra.items():
                    if k not in merged:
                        merged[k] = v

        with open(i18n_file, 'w', encoding='utf-8') as f:
            json.dump(merged, f, ensure_ascii=False, indent=2)
        print(f'  {lang}: {len(merged)} keys')

    print(f'\nMerged i18n for {len(modules)} modules -> {I18N_OUT}')


if __name__ == '__main__':
    merge()