I've been using django-modeltranslation to translate models in django for a while. It is really straightforward and it works really well on apps I've been developing, where all model translated content gets inserted with forms by the final user.
eg: inputs: content, content_en, content_pt, ...
I have to build an application where I need to translate 'built-in' model strings that are generated by django, like 'auth.permission.name' or 'contenttypes.contenttype.name' and add them to translation django.po files.
I came up with a solution that works fine,
which uses post_migration signals that create a file with lists of ugettext_lazy elements, so new strings, like a new contenttype.name for example, are added to 'django.po' dynamically and loaded to the database.
Yet, is a bit weird having to create a file with ugettext calls
in order to register the strings, but I didn't find another way of registering and adding them dynamically to the django.po file, so I need your help
Here's what I have done:
1. I created an app named 'tools', that is the last one on INSTALLED_APPS, so its migrations are naturally the last ones to be called. This app does not have any models, it just runs migrations, has the django-modeltranslation translation.py file and an application config with a post_migration signal call.
# translations.py
from modeltranslation.translator import translator, TranslationOptions
from django.contrib.auth.models import Permission
from django.contrib.contenttypes.models import ContentType
class PermissionTranslationOptions(TranslationOptions):
fields = ('name',)
class ContentTypeTranslationOptions(TranslationOptions):
fields = ('name',)
translator.register(Permission, PermissionTranslationOptions)
translator.register(ContentType, ContentTypeTranslationOptions)
2. Running 'manage.py makemigrations' creates the migrations on the 'auth' and 'contenttypes' applications with the extra 'name_*' fields.
3. the app has an application config that has a post_migrate signal
# __init__.py
default_app_config = 'apps.tools.config.SystemConfig'
# config.py
from django.apps import AppConfig
from django.db.models.signals import post_migrate
from apps.tools.translations.exporter import make_translations
from apps.tools.translations.importer import load_translations
def run_translations(sender, **kwargs):
# This creates the translations
make_translations()
# This loads the the translations to the db
load_translations()
class SystemConfig(AppConfig):
name = 'apps.tools'
verbose_name = 'Tools'
def ready(self):
# Call post migration operations
post_migrate.connect(run_translations, sender=self)
4. make_translations() is called after migrations and generates a file with lists of uggettext_lazy calls.
This is the bit I would like to change. Do I really need to create a file?
# exporter
import os
from django.contrib.auth.models import Permission
from django.contrib.contenttypes.models import ContentType
from django.utils import translation
from django.contrib.contenttypes.management import update_all_contenttypes
# TODO
# It has got to be another way
def make_translations():
# lets go default
translation.activate("en")
update_all_contenttypes()
try:
f = open(os.path.join(os.path.realpath(os.path.dirname(__file__)), 'translations.py'), 'w')
# Write file
f.write("from django.utils.translation import ugettext_lazy as _\n\n")
# All Permissions to lazy text
f.write('permissions = {\n')
for perm in Permission.objects.all().order_by('id'):
f.write(' "'+str(perm.id)+'": _("'+perm.name+'"),\n')
f.write('}\n\n')
# All Content types to lazy text
f.write('content_types = {\n')
for content in ContentType.objects.all().order_by('id'):
f.write(' "'+str(content.id)+'": _("'+content.name+'"),\n')
f.write('}\n\n')
# Closing file
f.close()
# Importing file to get it registered with ugettext_lazy
try:
from apps.tools.translations import translations
except:
print('Could not import file')
pass
except:
print('Could not create file')
pass
The above results in a file like this:
from django.utils.translation import ugettext_lazy as _
permissions = {
"1": _("Can add permission"),
"2": _("Can change permission"),
"3": _("Can delete permission"),
"4": _("Can add group"),
...
}
content_types = {
"1": _("group"),
"2": _("user"),
"3": _("permission"),
"4": _("content type"),
"5": _("session"),
...
}
5. Running 'makemessages' would add this strings to 'django.po' files, yet, the post_migration signal does not stop here, and loads the existing compiled strings in the database
# importer
from django.contrib.auth.models import Permission
from django.contrib.contenttypes.models import ContentType
from django.conf import settings
from django.utils import translation
def load_translations():
try:
from apps.tools.translations.translations import permissions, content_types
except:
# File does not exists
print('Translations could not be loaded')
return
# For each language
for lang in settings.LANGUAGES:
# Activate language
translation.activate(lang[0])
# Loading translated permissions
all_permissions = Permission.objects.all()
for permission in all_permissions:
permission.name = unicode(permissions[str(permission.id)])
permission.save()
# Loading translated content_types
all_contenttypes = ContentType.objects.all()
for contenttype in all_contenttypes:
contenttype.name = unicode(content_types[str(contenttype.id)])
contenttype.save()
How can I replace 'make_translations()' without creating a file and register those strings with ugettext_lazy?
Thanks for your help
