translate.py

import os

CONFIG_FILE_NAME = 'l10n.yaml'

def find_config_file_path(top: str = os.path.dirname(__file__)) -> str | None:
    for dirpath, dirnames, filenames in os.walk(topdown=True, top=top):
        for filename in filenames:
            if filename == CONFIG_FILE_NAME:
                return os.path.join(dirpath, filename)

    parent = os.path.dirname(top)

    if os.path.exists(parent):
        return find_config_file_path(parent)

    return None

config_file_path = find_config_file_path()

import yaml

with open(config_file_path, 'r', encoding='utf8') as config_file:
    config = yaml.safe_load(config_file)
    template_arb_file_name = config['template-arb-file']
    arb_dir_path = os.path.realpath(os.path.join(os.path.dirname(config_file_path), config['arb-dir']))
    arb_template_file_path = os.path.join(arb_dir_path, template_arb_file_name)

import json
from typing import Any
from gtranslate import translate
from hashlib import md5

def read_file_as_json(file_path: str) -> dict[str, Any]:
    with open(file_path, 'r', encoding='utf8') as file:
        return json.loads(file.read())

def write_file_as_json(file_path: str, data: dict[str, Any]) -> None:
    with open(file_path, 'w', encoding='utf8') as file:
        json.dump(data, file, indent=2, ensure_ascii=False)

base_i18n_data = read_file_as_json(arb_template_file_path)

def get_nested_value(nested_dict: dict[str, Any], *keys) -> Any:
    if len(keys) == 1:
        return nested_dict.get(keys[0])
    else:
        key = keys[0]
        if key in nested_dict:
            return get_nested_value(nested_dict[key], *keys[1:])
        else:
            return None

def set_nested_value(nested_dict, value, *keys):
    key = keys[0]
    if len(keys) == 1:
        nested_dict[key] = value
    else:
        if key not in nested_dict:
            nested_dict[key] = {}
        set_nested_value(nested_dict[key], value, *keys[1:])

for dirpath, dirnames, filenames in os.walk(arb_dir_path):
    cached = 0
    noncached = 0

    for filename in filenames:
        if filename == template_arb_file_name: 
            continue

        target_lang_file_path = os.path.join(dirpath, filename)

        target_lang_file_data = read_file_as_json(target_lang_file_path)

        target_locale = target_lang_file_data['@@locale']

        for k, v in base_i18n_data.items():
            def is_reserved_key() -> bool:
                RESERVED_PREFIXES = '_', '@'
                for prefix in RESERVED_PREFIXES:
                    if k.startswith(prefix):
                        # Not a translation key
                        # pass
                        return True
                return False

            if is_reserved_key():
                continue

            print(f'Key: {k}')
            print(f'Lang: {target_locale}')
            print(f'Original: {v}')

            def is_same_source_value():
                has_translation = k in target_lang_file_data and target_lang_file_data[k] is not None

                translation_source_value = get_nested_value(target_lang_file_data, f'@{k}', 'info', 'source')
                same_source_value = has_translation and translation_source_value == v

                return same_source_value

            try:
                if is_same_source_value():
                    cached += 1
                    print(f'Translated (cached): {translated}')
                else:
                    noncached += 1
                    translated = translate(v, translate_to=target_locale)
                    print(f'Translated: {translated}')

                    target_lang_file_data[k] = translated
                    set_nested_value(target_lang_file_data, v, f'@{k}', 'info', 'source')
                    write_file_as_json(target_lang_file_path, target_lang_file_data)
            except:
                print('Translated: [Translation not available, skipping...]')
                continue
            finally:
                print('-' * 30)

        print('*' * 30)
        print(f'Finished translating to {target_locale}')
        print('*' * 30)

    print('=' * 30)
    print(f'Loaded {noncached} non-cached results')
    print(f'Loaded {cached} cached results')
    print('=' * 30)