===========
Translation
===========

In order to make you bot translatable you have to add a minimal number of hooks to your Python code.

These hooks are called translation strings.

The aiogram translation utils is build on top of `GNU gettext Python module <https://docs.python.org/3/library/gettext.html>`_
and `Babel library <http://babel.pocoo.org/en/latest/>`_.

Installation
============

Babel is required to make simple way to extract translation strings from your code

Can be installed from pip directly:

.. code-block:: bash

    pip install Babel


or as `aiogram` extra dependency:

.. code-block:: bash

    pip install aiogram[i18n]


Make messages translatable
==========================

In order to gettext need to know what the strings should be translated you will need to write translation strings.

For example:

.. code-block:: python
    :emphasize-lines: 6-8

    from aiogram import html
    from aiogram.utils.i18n import gettext as _

    async def my_handler(message: Message) -> None:
        await message.answer(
            _("Hello, {name}!").format(
                name=html.quote(message.from_user.full_name)
            )
        )


.. danger::

    f-strings can't be used as translations string because any dynamic variables should be added to message after getting translated message


Also if you want to use translated string in keyword- or magic- filters you will need to use lazy gettext calls:


.. code-block:: python
    :emphasize-lines: 4

    from aiogram import F
    from aiogram.utils.i18n import lazy_gettext as __

    @router.message(F.text == __("My menu entry"))
    ...


.. danger::

    Lazy gettext calls should always be used when the current language is not know at the moment


.. danger::

    Lazy gettext can't be used as value for API methods or any Telegram Object (like :class:`aiogram.types.inline_keyboard_button.InlineKeyboardButton` or etc.)

**Working with plural forms**

The `gettext` from `aiogram.utils.i18n` is the one alias for two functions _gettext_ and _ngettext_
of `GNU gettext Python module <https://docs.python.org/3/library/gettext.html>`_. Therefore, the wrapper for message
strings is the same `_()`. You need to pass three parameters to the function:
a singular string, a plural string, and a value.

.. code-block:: python
    :emphasize-lines: 6, 10
    from aiogram import html
    from aiogram.utils.i18n import gettext as _

    async def my_handler(message: Message) -> None:
        try:
            n = int(message.text)
        except ValueError:
            n = 1
        await message.answer(
            _("You ordered {n} piece.", "You ordered {n} pieces.",n).format(n)
            )
        )

Configuring engine
==================

After you messages is already done to use gettext your bot should know how to detect user language

On top of your application the instance of :class:`aiogram.utils.i18n.I18n` should be created


.. code-block::

    i18n = I18n(path="locales", default_locale="en", domain="messages")


After that you will need to choose one of builtin I18n middleware or write your own.

Builtin middlewares:


SimpleI18nMiddleware
~~~~~~~~~~~~~~~~~~~~

.. autoclass:: aiogram.utils.i18n.middleware.SimpleI18nMiddleware
    :members: __init__

ConstI18nMiddleware
~~~~~~~~~~~~~~~~~~~

.. autoclass:: aiogram.utils.i18n.middleware.ConstI18nMiddleware
    :members: __init__

FSMI18nMiddleware
~~~~~~~~~~~~~~~~~

.. autoclass:: aiogram.utils.i18n.middleware.FSMI18nMiddleware
    :members: __init__,set_locale


I18nMiddleware
~~~~~~~~~~~~~~

or define you own based on abstract I18nMiddleware middleware:

.. autoclass:: aiogram.utils.i18n.middleware.I18nMiddleware
    :members: __init__,setup,get_locale


Deal with Babel
===============

Step 1 Extract messages
~~~~~~~~~~~~~~~~~~~~~~~

.. code-block:: bash

    pybabel extract --input-dirs=. -o locales/messages.pot


Here is :code:`--input-dirs=.` - path to code and the :code:`locales/messages.pot`
is template where messages will be extracted and `messages` is translation domain.

**Working with plural forms**

Extracting with Pybabel all strings options:

- :code:`-k _:1,1t -k _:1,2` - for both singular and plural
- :code:`-k __` - for lazy strings

.. code-block:: bash

    pybabel extract -k _:1,1t -k _:1,2 -k __ --input-dirs=. -o locales/messages.pot

.. note::

    Some useful options:

    - Add comments for translators, you can use another tag if you want (TR) :code:`--add-comments=NOTE`
    - Contact email for bugreport :code:`--msgid-bugs-address=EMAIL`
    - Disable comments with string location in code :code:`--no-location`
    - Copyrights :code:`--copyright-holder=AUTHOR`
    - Set project name :code:`--project=MySuperBot`
    - Set version :code:`--version=2.2`


Step 2: Init language
~~~~~~~~~~~~~~~~~~~~~

.. code-block:: bash

    pybabel init -i locales/messages.pot -d locales -D messages -l en

- :code:`-i locales/messages.pot` - pre-generated template
- :code:`-d locales` - translations directory
- :code:`-D messages` - translations domain
- :code:`-l en` - language. Can be changed to any other valid language code (For example :code:`-l uk` for ukrainian language)


Step 3: Translate texts
~~~~~~~~~~~~~~~~~~~~~~~

To open .po file you can use basic text editor or any PO editor, e.g. `Poedit <https://poedit.net/>`_

Just open the file named :code:`locales/{language}/LC_MESSAGES/messages.po` and write translations

Step 4: Compile translations
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. code-block:: bash

    pybabel compile -d locales -D messages


Step 5: Updating messages
~~~~~~~~~~~~~~~~~~~~~~~~~~

When you change the code of your bot you need to update po & mo files

- Step 5.1: regenerate pot file: command from step 1
- Step 5.2: update po files
    .. code-block::

        pybabel update -d locales -D messages -i locales/messages.pot

- Step 5.3: update your translations: location and tools you know from step 3
- Step 5.4: compile mo files: command from step 4