#!/usr/bin/env python3
"""
Merge i18n translations from all pipeline modules into wwwroot/i18n.
Scans each module's i18n directory and merges into the app's centralized i18n.

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

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')

# Modules to scan for i18n
MODULES = [
    ('pipeline_core', os.path.join(ROOT_DIR, 'pipeline_core', 'i18n')),
    ('pipeline_ops', os.path.join(ROOT_DIR, 'pipeline_ops', 'i18n')),
    ('pipeline_dist', os.path.join(ROOT_DIR, 'pipeline_dist', 'i18n')),
    ('pipeline-sdlc', os.path.join(os.path.dirname(ROOT_DIR), 'pipeline-sdlc', 'i18n')),
    ('showcase', os.path.join(os.path.dirname(ROOT_DIR), 'showcase', 'i18n')),
]

LANGS = ['zh', 'en', 'ko', 'jp']


def parse_msg_txt(filepath):
    """Parse msg.txt: 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()
    return result


def merge():
    os.makedirs(I18N_OUT, exist_ok=True)
    stats = {}

    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)

        # Scan each module
        for mod_name, mod_i18n_dir in MODULES:
            lang_dir = os.path.join(mod_i18n_dir, lang)
            if not os.path.isdir(lang_dir):
                # Try app's internal i18n
                alt_dir = os.path.join(ROOT_DIR, 'pipeline_core', 'i18n', lang)
                if not os.path.isdir(alt_dir):
                    continue
                lang_dir = alt_dir

            # Load msg.txt
            msg_file = os.path.join(lang_dir, 'msg.txt')
            if os.path.exists(msg_file):
                msgs = parse_msg_txt(msg_file)
                for k, v in msgs.items():
                    if k not in merged:
                        merged[k] = v
                        stats.setdefault(mod_name, 0)
                        stats[mod_name] += 1

            # Load i18n.json
            mod_i18n = os.path.join(lang_dir, 'i18n.json')
            if os.path.exists(mod_i18n):
                with open(mod_i18n, 'r', encoding='utf-8') as f:
                    extra = json.load(f, object_pairs_hook=OrderedDict)
                for k, v in extra.items():
                    if k not in merged:
                        merged[k] = v

        # Write merged
        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')

    # Write msg.txt for the app root
    print(f'\nMerged i18n for {len(MODULES)} modules -> {I18N_OUT}')


if __name__ == '__main__':
    merge()