Tools for using Babel with Django

This package contains various utilities for integration of Babel into the Django web framework:

• A message extraction plugin for Django templates.
• A middleware class that adds the Babel Locale object to requests.
• A set of template tags for date and number formatting.

Extracting Messages

Babel provides a message extraction framework similar to GNU xgettext, but more extensible and geared towards Python applications. While Django does provide wrapper scripts for making the use of xgettext more convenient, the extraction functionality is rather limited. For example, you can't use template files with an extension other than .html, and everything needs to be in your project package directory.

Extraction Method Mapping

So enmerkar comes with an extraction method plugin that can extract localizable messages from Django template files. Python is supported out of the box by Babel. To use this extraction functionality, create a file called babel.cfg in your project directory (the directory above your project package), with the content:

[django: templates/**.*]
[django: mypkg/*/templates/**.*]
[python: mypkg/**.py]

This instructs Babel to look for any files in the top-level templates directory, or any files in application templates directories, and use the extraction method named “django” to extract messages from those template files. You'll need to adjust those glob patterns to wherever you my be storing your templates.

Also, any files with the extension .py inside your package directory (replace “mypkg” with the actual name of your Django project package) are processed by the “python” extraction method.

If you don't use setuptools, or for some reason haven't installed enmerkar using setuptools/pip, you'll need to define what function the extraction method “django” maps to. This is done in an extra section at the top of the configuration file:

[extractors]
django = enmerkar.extract:extract_django

The encoding of the templates is assumed to be UTF-8. If you are using a different encoding, you will need to specify it in the configuration. For example:

[django: templates/**.*]
encoding = iso-8859-1

Running the Extraction Process

Once you've set up the configuration file, the actual extraction is performed by executing the command-line program pybabel which is installed alongside the Babel package:

$ cd projectdir
$ pybabel extract -F babel.cfg -o mypkg/locale/django.pot .

This creates the PO file template in mypkg/locale/django.pot.

Creating and Updating Translations Catalogs

If you don't already have translation catalogs, you need to create them. This is done using the pybabel init command:

$ pybabel init -D django -i mypkg/locale/django.pot -d mypkg/locale -l en_US
$ pybabel init -D django -i mypkg/locale/django.pot -d mypkg/locale -l de_DE

This should create two files: mypkg/locale/en_US/django.po and mypkg/locale/de_DE/django.po. These files are where you put the actual translations.

When you modify your Python source files or your templates, you genereally need to sync the translation catalogs. For that, you first perform a fresh extraction as described in the previous section, so that the django.pot file gets updated.

Then, you run the pybabel update command to merge the changes into the translation catalogs:

`bash $ pybabel update -D django -i mypkg/locale/django.pot -d mypkg/locale `

This will update all the .po files found in the mypkg/locale directory.

Compiling Translations Catalogs

Finally, you need to compile those .po files to binary .mo files. Use the pybabel compile command for that:

$ pybabel compile -D django -d mypkg/locale

Add the --statistics option to get information about the completeness of your translations:

$ pybabel compile -D django -d mypkg/locale --statistics

Using setup.py

Much of the above process can be automated if you add a setup.py script to your project and use the distutils/setuptools commands that come with Babel. This is described at Distutils/Setuptools Integration.

Using the Middleware

To use the enmerkar middleware, add it to the list of MIDDLEWARE_CLASSES in your settings module. If you're also using Django's own LocaleMiddleware to vary the locale based on user preference, the enmerkar middleware must be inserted after the Django one:

MIDDLEWARE_CLASSES = (
    ...
    'django.middleware.locale.LocaleMiddleware',
    'enmerkar.middleware.LocaleMiddleware',
    ...
)

This adds a locale attribute to the request object, which is an instance of the Babel Locale class. You can access the locale via request.locale when the request object is available, or otherwise use the django_babel.middleware.get_current_locale() function to get the current locale from a thread-local cache.

Using the Template Tags

The template filters provided by enmerkar allow formatting of date/time and number values in a locale-sensitive manner, providing much more powerful alternatives to the date, time, and floatformat filters that come with Django.

To make the template filters/tags available, you need to add enmerkar to the list of INSTALLED_APPS in your settings module:

INSTALLED_APPS = (
    ...
    'enmerkar',
    ...
)

And in every template you want to use the filters, you need to explicitly load the library:

{% load babel %}

General information on date/time and number formatting can be found at Date Formatting and Number Formatting.

The following filters are made available. The examples assume a locale of en_US.

datefmt

Renders a string representation of a date.

• Input: datetime.date, datetime.datetime, or a float/int timestamp
• Parameters: the format name or pattern (optional)

Assuming that book.pubdate returns a datetime.date or datetime.datetime object:

{{ book.pubdate|datefmt:"short" }}

would render: 4/1/07, and

{{ book.pubdate|datefmt:"E, MMM dd yyyy GGG" }}

would render: Sun, Apr 01 2007 AD

datetimefmt

Renders a string representation of a date and time.

• Input: datetime.datetime, or a float/int timestamp
• Parameters: the format name or pattern (optional)

Examples:

{{ book.pubdate|datetimefmt:"short" }}

would render: 4/1/07 3:30 PM, and

{{ book.pubdate|datetimefmt:"E, MMM dd yyyy GGG' - 'HH:mm:ss'" }}

would render: Sun, Apr 01 2007 AD - 15:30:00

timefmt

Renders a string representation of a time.

• Input: datetime.datetime, datetime.time, or a float/int timestamp
• Parameters: the format name or pattern (optional)

Examples:

{{ book.pubdate|timefmt:"short" }}

would render: 3:30 PM, and

{{ book.pubdate|timefmt:"h 'o''clock' a'" }}

would render: 3 o'clock PM

decimalfmt

Renders a string representation of a decimal number.

• Input: a Decimal object, or a float/int/long value
• Parameters: the format name or pattern (optional)

Examples:

{{ book.pagecount|decimalfmt }}

would render: 1,234, and

{{ book.pagecount|decimalfmt:"#,##0.00" }}

would render: 1,234.00

currencyfmt

Renders a number formatted as a currency value.

• Input: a Decimal object, or a float/int/long value
• Parameters: the currency code

Examples:

{{ book.price|currencyfmt:"USD" }}

would render: $49.90

percentfmt

Renders a string representation of a number as a percentage.

• Input: a Decimal object, or a float/int/long value
• Parameters: the format name or pattern (optional)

Examples:

Assuming book.rebate would return 0.15,

{{ book.rebate|percentfmt }}

would render 15%, and

{{ book.rebate|percentfmt:"#,##0.00%" }}

would render 15.00%.

scientificfmt

Renders a string representation of a number using scientific notation.

• Input: a Decimal object, or a float/int/long value
• Parameters: none

Examples:

Assuming book.numsold would return 1.000.000,

{{ book.numsold|scientificfmt }}

would render 10E5.