32 KiB
Internationalization (i18n)
Application internationalization is a many-faceted area of development, focused on making applications available and user-friendly to a worldwide audience. This page describes Angular's internationalization (i18n) tools, which can help you make your app available in multiple languages.
See the i18n Example for a simple example of an AOT-compiled app, translated into French.
{@a angular-i18n}
Angular and i18n
Internationalization is the process of designing and preparing your app to be usable in different languages. Localization is the process of translating your internationalized app into specific languages for particular locales.
Angular simplifies the following aspects of internationalization:
- Displaying dates, number, percentages, and currencies in a local format.
- Preparing text in component templates for translation.
- Handling plural forms of words.
- Handling alternative text.
For localization, you can use the Angular CLI to generate most of the boilerplate necessary to create files for translators, and to publish your app in multiple languages. After you have set up your app to use i18n, the CLI can help you with the following steps:
- Extracting localizable text into a file that you can send out to be translated.
- Building and serving the app for a given locale, using the translated text.
- Creating multiple language versions of your app.
{@a setting-up-cli}
Setting up localization with the Angular CLI
The first step to setting up localization when using the Angular CLI is to add the @angular/localize
package to the project. This will install the package within your project as well as
initialize the project to take advantage of Angular's localization features.
{@a setting-up-locale}
Setting up the locale of your app
A locale is an identifier (id) that refers to a set of user preferences that tend to be shared within a region of the world, such as country. This document refers to a locale identifier as a "locale" or "locale id".
A Unicode locale identifier is composed of a Unicode language identifier and (optionally) the
character -
followed by a locale extension. (For historical reasons the character _
is supported
as an alternative to -
.) For example, in the locale id fr-CA
the fr
refers to the French
language identifier, and the CA
refers to the locale extension Canada.
Angular follows the Unicode LDML convention that uses stable identifiers (Unicode locale identifiers) based on the norm BCP47. It is very important that you follow this convention when you define your locale, because the Angular i18n tools use this locale id to find the correct corresponding locale data.
By default, Angular uses the locale en-US
, which is English as spoken in the United States of America.
For more information about Unicode locale identifiers, see the CLDR core spec.
For a complete list of locales supported by Angular, see the Angular repository.
The locale identifiers used by CLDR and Angular are based on BCP47. These specifications change over time; the following table maps previous identifiers to current ones at time of writing:
Locale name | Old locale id | New locale id |
---|---|---|
Indonesian | in | id |
Hebrew | iw | he |
Romanian Moldova | mo | ro-MD |
Norwegian Bokmål | no, no-NO | nb |
Serbian Latin | sh | sr-Latn |
Filipino | tl | fil |
Portuguese Brazil | pt-BR | pt |
Chinese Simplified | zh-cn, zh-Hans-CN | zh-Hans |
Chinese Traditional | zh-tw, zh-Hant-TW | zh-Hant |
Chinese Traditional Hong Kong | zh-hk | zh-Hant-HK |
i18n pipes
Angular pipes can help you with internationalization: the DatePipe
, CurrencyPipe
, DecimalPipe
and PercentPipe
use locale data to format data based on the LOCALE_ID
dependency injection token.
By default, Angular only contains locale data for en-US
.
The CLI automatically includes the locale data and sets the LOCALE_ID
value for you when you use the
parameter --i18nLocale
or option i18nLocale
with ng serve
and ng build
.
However, if you manually set the value of LOCALE_ID
to another locale, you must import locale data for that new locale.
If you want to import locale data for other languages, you can do it manually:
The first parameter is an object containing the locale data imported from @angular/common/locales
.
By default, the imported locale data is registered with the locale id that is defined in the Angular
locale data itself.
If you want to register the imported locale data with another locale id, use the second parameter to
specify a custom locale id. For example, Angular's locale data defines the locale id for French as
"fr". You can use the second parameter to associate the imported French locale data with the custom
locale id "fr-FR" instead of "fr".
The files in @angular/common/locales
contain most of the locale data that you
need, but some advanced formatting options might only be available in the extra dataset that you can
import from @angular/common/locales/extra
. An error message informs you when this is the case.
Ivy will automatically include the extra locale data if the locale was configured via i18nLocale
.
All locale data used by Angular are extracted from the Unicode Consortium's Common Locale Data Repository (CLDR).
Template translations
This document refers to a unit of translatable text as "text," a "message", or a "text message."
The i18n template translation process has four phases:
-
Mark static text messages in your component templates for translation.
-
Create a translation file: Use the Angular CLI
xi18n
command to extract the marked text into an industry-standard translation source file. -
Edit the generated translation file: Translate the extracted text into the target language.
-
Merge the completed translation file into the app. To do this, use the Angular CLI
build
command to compile the app, choosing a locale-specific configuration, or specifying the following command options.--i18nFile
=path to the translation file--i18nFormat
=format of the translation file (only required if using an earlier version of Angular, or if you disable Ivy)--i18nLocale
= locale id
The command replaces the original messages with translated text, and generates a new version of the app in the target language.
You need to build and deploy a separate version of the app for each supported language.
{@a i18n-attribute}
Mark text with the i18n attribute
The Angular i18n
attribute marks translatable content. Place it on every element tag whose fixed
text is to be translated.
In the example below, an <h1>
tag displays a simple English language greeting, "Hello i18n!"
To mark the greeting for translation, add the i18n
attribute to the <h1>
tag.
i18n
is a custom attribute, recognized by Angular tools and compilers.
After translation, the compiler removes it. It is not an Angular directive.
{@a help-translator}
Help the translator with a description and meaning
To translate a text message accurately, the translator may need additional information or context.
You can add a description of the text message as the value of the i18n
attribute, as shown in the
example below:
The translator may also need to know the meaning or intent of the text message within this particular app context.
You add context by beginning the i18n
attribute value with the meaning and
separating it from the description with the |
character: <meaning>|<description>
All occurrences of a text message that have the same meaning will have the same translation. A text message that is associated with different meanings can have different translations.
The Angular extraction tool preserves both the meaning and the description in the translation source file to facilitate contextually-specific translations, but only the combination of meaning and text message are used to generate the specific id of a translation. If you have two similar text messages with different meanings, they are extracted separately. If you have two similar text messages with different descriptions (not different meanings), then they are extracted only once.
{@a custom-id}
Set a custom id for persistence and maintenance
The angular i18n extractor tool generates a file with a translation unit entry for each i18n
attribute in a template. By default, it assigns each translation unit a unique id such as this one:
When you change the translatable text, the extractor tool generates a new id for that translation unit. You must then update the translation file with the new id.
Alternatively, you can specify a custom id in the i18n
attribute by using the prefix @@
.
The example below defines the custom id introductionHeader
:
When you specify a custom id, the extractor tool and compiler generate a translation unit with that custom id.
The custom id is persistent. The extractor tool does not change it when the translatable text changes. Therefore, you do not need to update the translation. This approach makes maintenance easier.
Use a custom id with a description
You can use a custom id in combination with a description by including both in the value of the
i18n
attribute. In the example below, the i18n
attribute value includes a description, followed
by the custom id
:
You also can add a meaning, as shown in this example:
Define unique custom ids
Be sure to define custom ids that are unique. If you use the same id for two different text messages, only the first one is extracted, and its translation is used in place of both original text messages.
In the example below the custom id myId
is used for two different messages:
<h3 i18n="@@myId">Hello</h3>
<!-- ... -->
<p i18n="@@myId">Good bye</p>
Consider this translation to French:
<trans-unit id="myId" datatype="html">
<source>Hello</source>
<target state="new">Bonjour</target>
</trans-unit>
Because the custom id is the same, both of the elements in the resulting translation contain
the same text, Bonjour
:
<h3>Bonjour</h3>
<!-- ... -->
<p>Bonjour</p>
{@a no-element}
Translate text without creating an element
If there is a section of text that you would like to translate, you can wrap it in a <span>
tag.
However, if you don't want to create a new DOM element merely to facilitate translation,
you can wrap the text in an <ng-container>
element.
The <ng-container>
is transformed into an html comment:
{@a translate-attributes}
Translate attributes
Displayed text is sometimes supplied as the value of an attribute, rather than the content of tag.
For example, if your template has an image with a title
attribute, the text value of the title
attribute needs to be translated.
To mark an attribute for translation, add an attribute in the form of i18n-x
,
where x
is the name of the attribute to translate. The following example shows how to mark the
title
attribute for translation by adding the i18n-title
attribute on the img
tag:
This technique works for any attribute of any element.
You also can assign a meaning, description, and id with the i18n-x="<meaning>|<description>@@<id>"
syntax.
Regular expressions for plurals and selections
Different languages have different pluralization rules and grammatical constructions that add
complexity to the translation task.
You can use regular expressions with the plural
and select
clauses to provide patterns that aid translation in these cases.
{@a plural-ICU}
Pluralization
Suppose that you want to say that something was "updated x minutes ago". In English, depending upon the number of minutes, you could display "just now", "one minute ago", or "x minutes ago" (with x being the actual number). Other languages might express the cardinality differently.
The example below shows how to use a plural
ICU expression to display one of those three options
based on when the update occurred:
- The first parameter is the key. It is bound to the component property (
minutes
), which determines the number of minutes. - The second parameter identifies this as a
plural
translation type. - The third parameter defines a pluralization pattern consisting of pluralization categories and their matching values.
This syntax conforms to the ICU Message Format as specified in the CLDR pluralization rules.
Pluralization categories include (depending on the language):
- =0 (or any other number)
- zero
- one
- two
- few
- many
- other
After the pluralization category, put the default English text in braces ({}
).
In the example above, the three options are specified according to that pluralization pattern. For
talking about zero minutes, you use =0 {just now}
. For one minute, you use =1 {one minute}
.
Any unmatched cardinality uses other {{{minutes}} minutes ago}
. You could choose to add patterns
for two, three, or any other number if the pluralization rules were different. For the example of
"minute", only these three patterns are necessary in English.
You can use interpolations and html markup inside of your translations.
{@a select-ICU}
Select among alternative text messages
If your template needs to display different text messages depending on the value of a variable, you need to translate all of those alternative text messages.
You can handle this with a select
ICU expression. It is similar to the plural
expressions
except that you choose among alternative translations based on a string value instead of a number,
and you define those string values.
The following format message in the component template binds to the component's gender
property,
which outputs one of the following string values: "male", "female" or "other".
The message maps those values to the appropriate translations:
{@a nesting-ICUS}
Nesting plural and select ICU expressions
You can also nest different ICU expressions together, as shown in this example:
{@a ng-xi18n} {@a ng-xi18n-options}
Create a translation source file
When your app is ready, you can use the Angular CLI to extract the text messages marked with i18n
and attributes marked with i18n-x
into a translation source file.
Open a terminal window at the root of the app project and run the CLI command xi18n
.
By default, the command creates a file named messages.xlf
in your project's root directory.
{@a other-formats}
Output options
You can supply command options to change the format, the name, the location, and the source locale of the extracted file.
For example, to create a file in the src/locale
folder, specify the output path:
By default, the xi18n
command generates a translation file named messages.xlf
in the
XML Localization Interchange File Format
(XLIFF, version 1.2).
The command can read and write files in three translation formats:
- XLIFF 1.2 (default)
- XLIFF 2
- XML Message Bundle (XMB)
You can specify the translation format explicitly with the --format
command option, as illustrated in
these example commands:
The sample in this guide uses the default XLIFF 1.2 format.
XLIFF files have the extension .xlf. The XMB format generates .xmb source files but uses .xtb (XML Translation Bundle: XTB) translation files.
You can change the name of the translation source file that is generated by the extraction tool with
the --outFile
command option:
ng xi18n --out-file source.xlf
You can specify the base locale of your app with the--i18n-locale
command option:
ng xi18n --i18n-locale fr
The extraction tool uses the locale to add the app locale information into your translation source file. This information is not used by Angular, but external translation tools may need it.
{@a translate}
Translate the source text
By default, the ng xi18n
command generates a translation source file named messages.xlf
in the project src
folder.
The next step is to translate the display strings in this source file into language-specific
translation files. The example in this guide creates a French translation file.
{@a localization-folder}
Create a localization folder
Most apps are translated into more than one other language. For this reason, it is standard practice for the project structure to reflect the entire internationalization effort.
One approach is to dedicate a folder to localization and store related assets, such as internationalization files, there.
Localization and internationalization are different but closely related terms.
This guide follows that approach. It has a locale
folder under src/
.
Assets within that folder have a filename extension that matches their associated locale.
Create the translation files
For each translation source file, there must be at least one language translation file for the resulting translation.
For this example:
- Make a copy of the
messages.xlf
file. - Put the copy in the
locale
folder. - Rename the copy to
messages.fr.xlf
for the French language translation.
If you were translating to other languages, you would repeat these steps for each target language.
{@a translate-text-nodes}
Translate text nodes
In a large translation project, you would send the messages.fr.xlf
file to a French translator who
would enter the translations using an XLIFF file editor.
This sample file is easy to translate without a special editor or knowledge of French.
- Open
messages.fr.xlf
and find the first<trans-unit>
section:
This XML element represents the translation of the
<h1>
greeting tag that you marked with thei18n
attribute earlier in this guide.
Note that the translation unit
id=introductionHeader
is derived from the customid
that you set earlier, but without the@@
prefix required in the source HTML.
- Duplicate the
<source/>
tag, rename ittarget
, and then replace its content with the French greeting. If you were working with a more complex translation, you could use the information and context provided by the source, description, and meaning elements to guide your selection of the appropriate French translation.
- Translate the other text nodes the same way:
The Angular i18n tools generated the ids for these translation units. Don't change them.
Each id
depends upon the content of the template text and its assigned meaning.
If you change either the text or the meaning, then the id
changes.
For more information, see the translation file maintenance discussion.
{@a translate-plural-select}
Translating plural and select expressions
The plural and select ICU expressions are extracted separately, so they require special attention when preparing for translation.
Look for these expressions in relation to other translation units that you recognize from
elsewhere in the source template. In this example, you know the translation unit for the select
must be just below the translation unit for the logo.
{@a translate-plural}
Translate plural
To translate a plural
, translate its ICU format match values:
You can add or remove plural cases, with each language having its own cardinality. (See CLDR plural rules.)
{@a translate-select}
Translate select
Below is the content of our example select
ICU expression in the component template:
The extraction tool broke that into two translation units because ICU expressions are extracted separately.
The first unit contains the text that was outside of the select
.
In place of the select
is a placeholder, <x id="ICU">
, that represents the select
message.
Translate the text and move around the placeholder if necessary, but don't remove it. If you remove
the placeholder, the ICU expression will not be present in your translated app.
The second translation unit, immediately below the first one, contains the select
message.
Translate that as well.
Here they are together, after translation:
{@a translate-nested}
Translate a nested expression
A nested expression is similar to the previous examples. As in the previous example, there are two translation units. The first one contains the text outside of the nested expression:
The second unit contains the complete nested expression:
And both together:
The entire template translation is complete. The next section describes how to load that translation into the app.
{@a app-pre-translation}
The app and its translation file
The sample app and its translation file are now as follows:
{@a merge}
Merge the completed translation file into the app
To merge the translated text into component templates, compile the app with the completed translation file.
Provide the Angular compiler with three translation-specific pieces of information:
- The translation file.
- The translation file format.
- The locale (
fr
oren-US
for instance).
The compilation process is the same whether the translation file is in .xlf
format or in another
format that Angular understands, such as .xtb
.
{@a merge-aot} The AOT compiler is part of a build process that produces a small, fast, ready-to-run application package, with Ivy it is now used for both production and development.
Ivy does not support merging i18n translations when using JIT mode. If you disable Ivy and are using JIT mode, additional information regarding translation merging can be found here.
When you internationalize with the AOT compiler, you must pre-build a separate application package for each language and serve the appropriate package based on either server-side language detection or URL parameters.
To instruct the AOT compiler to use your translation configuration, set the three "i18n" build configuration options in your CLI configuration file, angular.json
.
i18nFile
: the path to the translation file.i18nFormat
: the format of the translation file (only required if using an earlier version of Angular, or if you disable Ivy).i18nLocale
: the locale id.
You should also direct the output to a locale-specific folder to keep it separate from other locale versions of your app, by setting the outputPath
configuration option.
"build": {
...
"configurations": {
...
"fr": {
"aot": true,
"outputPath": "dist/my-project-fr/",
"i18nFile": "src/locale/messages.fr.xlf",
"i18nFormat": "xlf",
"i18nLocale": "fr",
...
}
}
},
"serve": {
...
"configurations": {
...
"fr": {
"browserTarget": "*project-name*:build:fr"
}
}
}
You can then pass this configuration to the ng serve
or ng build
commands.
The example below shows how to serve the French language file created in previous
sections of this guide:
For production builds, you can use configuration composition to execute both configurations:
ng build --configuration=production,fr...
"architect": {
"build": {
"builder": "@angular-devkit/build-angular:browser",
"options": { ... },
"configurations": {
"fr": {
"aot": true,
"outputPath": "dist/my-project-fr/",
"i18nFile": "src/locale/messages.fr.xlf",
"i18nFormat": "xlf",
"i18nLocale": "fr",
"i18nMissingTranslation": "error",
}
}
},
...
"serve": {
"builder": "@angular-devkit/build-angular:dev-server",
"options": {
"browserTarget": "my-project:build"
},
"configurations": {
"production": {
"browserTarget": "my-project:build:production"
},
"fr": {
"browserTarget": "my-project:build:fr"
}
}
}
}
The same configuration options can also be provided through the CLI with your existing production
configuration.
{@a missing-translation}
Report missing translations
By default, when a translation is missing, the build succeeds but generates a warning such as
Missing translation for message "foo"
. You can configure the level of warning that is generated by
the Angular compiler:
- Error: throw an error. If you are using AOT compilation, the build will fail. If you are using JIT compilation, the app will fail to load.
- Warning (default): show a 'Missing translation' warning in the console or shell.
- Ignore: do nothing.
You specify the warning level in the configurations
section of your Angular CLI configuration file, angular.json
. The example below shows how to set the warning level to error.
"configurations": {
...
"fr": {
...
"i18nMissingTranslation": "error"
}
}
Build for multiple locales
When you use the CLI build
or serve
command to build your application for different locales, change the output path using the --outputPath
command option (along with the i18n-specific command options), so that the translation files are saved to different locations.
When you are serving a locale-specific version from a subdirectory, you can also change the base URL used by your app by specifying the --baseHref
option.
For example, if the French version of your application is served from https://example.com/fr/, configure the build for the French version as follows.
"configurations": {
"fr": {
"aot": true,
"outputPath": "dist/my-project-fr/",
"baseHref": "/fr/",
"i18nFile": "src/locale/messages.fr.xlf",
"i18nFormat": "xlf",
"i18nLocale": "fr",
"i18nMissingTranslation": "error",
}
For more details about how to create scripts to generate an app in multiple languages and how to set up Apache 2 and NGINX to serve them from different subdirectories, read this tutorial by Philippe Martin.