@cookbook/solid-intl v0.1.2
Solid Intl
A universal internationalization (i18n) for Solid inspired by React Intl & FormatJS.
Quick start
Install it:
npm i @cookbook/solid-intl
# or
yarn add @cookbook/solid-intl
# or
pnpm add @cookbook/solid-intlWrap your application with <IntlProvider />
import { IntlProvider } from '@cookbook/solid-intl'
// usersLocale = 'en', 'en-US', 'pt-BR', so on...
// translationsForUsersLocale = translations for selected locale
render(() =>
<IntlProvider locale={usersLocale} messages={translationsForUsersLocale}>
<App />
</IntlProvider>,
document.getElementById("root")!);Create your message and use it with intl.formatMessage
import { defineMessages, useIntl } from '@cookbook/solid-intl';
const messages = defineMessages({
greeting: {
id: "app.greeting",
defaultMessage: "Hello world!",
},
});
function Component(props) {
const { intl } = useIntl();
return intl.formatMessage(messages.greeting) // outputs: Hello world!
}Table of contents
API
<IntlProvider />
Solid Intl uses the provider pattern to scope an i18n context to a tree of components. This allows configuration like the current locale and set of translated strings/messages to be provided at the root of a component tree and made available to the 'useIntl()' hook methods. This component is used to setup the i18n context for a tree. Usually, this component will wrap an app's root component so that the entire app will be within the configured i18n context. The following are the i18n configuration props that can be set:
interface IntlConfig {
locale: Locale;
timeZone?: string;
fallbackOnEmptyString?: boolean;
formats?: CustomFormats;
messages: Record<MessageIds, string> | Record<MessageIds, MessageFormatElement[]>;
defaultLocale?: string;
defaultFormats?: CustomFormats;
defaultRichTextElements?: Record<string, FormatXMLElementFn<JSXElement>>;
onError(err: MissingTranslationError | MessageFormatError | MissingDataError | InvalidConfigError | UnsupportedFormatterError | FormatError)?: void;
onWarn(warning: string)?: void;
}locale, formats, and messages
The user's current locale and what the app should be rendered in.
While defaultLocale and defaultFormats are for fallbacks or during development and represent the app's default.
Notice how there is no defaultMessages, that's because each Message Descriptor provides a defaultMessage.
defaultLocale and defaultFormats
Default locale & formats for when a message is not translated (missing from messages).
defaultLocale should be the locale that defaultMessages are declared in so that a sentence is coherent in a single locale. Without defaultLocale and/or if it's set incorrectly, you might run into scenario where a sentence is in English but embedded date/time is in Spanish.
defaultRichTextElements
A map of tag to rich text formatting function. This is meant to provide a centralized way to format common tags such as <b>, <p>... or enforcing certain Design System in the codebase (e.g standardized <a> or <button>...).
See https://github.com/formatjs/formatjs/issues/1752 for more context.
onError
Allows the user to provide a custom error handler.
By default, error messages are logged using console.error if NODE_ENV is not set to production.
onWarn
Allows the user to provide a custom warning handler.
By default, warning messages are logged using console.warning if NODE_ENV is not set to production.
useIntl() hook
formatDate
function formatDate(
value: number | Date,
options?: Intl.DateTimeFormatOptions & {format?: string}
): stringThis function will return a formatted date string.
It expects a value which can be parsed as a date (i.e., isFinite(new Date(value))), and accepts options that conform to DateTimeFormatOptions.
intl.formatDate(Date.now(), {
year: 'numeric',
month: 'numeric',
day: 'numeric',
})
// outputs: 04/01/2023formatTime
function formatTime(
value: number | Date,
options?: Intl.DateTimeFormatOptions & {format?: string}
): stringThis function will return a formatted date string, but it differs from formatDate by having the following default options:
{
hour: 'numeric',
minute: 'numeric',
}It expects a value which can be parsed as a date (i.e., isFinite(new Date(value))), and accepts options that conform to DateTimeFormatOptions.
intl.formatTime(Date.now())
// outputs: "4:03 PM"formatRelativeTime
BROWSER SUPPORT:
This requires Intl.RelativeTimeFormat which has limited browser support. Please use our polyfill if you plan to support them.
type Unit =
| 'second'
| 'minute'
| 'hour'
| 'day'
| 'week'
| 'month'
| 'quarter'
| 'year'
type RelativeTimeFormatOptions = {
numeric?: 'always' | 'auto'
style?: 'long' | 'short' | 'narrow'
}
function formatRelativeTime(
value: number,
unit: Unit,
options?: Intl.RelativeTimeFormatOptions & {
format?: string
}
): stringThis function will return a formatted relative time string (e.g., "1 hour ago"). It expects a value which is a number, a unit and options that conform to Intl.RelativeTimeFormatOptions.
intl.formatRelativeTime(0)
// outputs: in 0 secondsintl.formatRelativeTime(-24, 'hour', {style: 'narrow'})
// outputs: 24 hr agoformatNumber
This function uses Intl.NumberFormat options.
function formatNumber(
value: number,
options?: Intl.NumberFormatOptions & {format?: string}
): stringThis function will return a formatted number string.
It expects a value which can be parsed as a number, and accepts options that conform to NumberFormatOptions.
intl.formatNumber(1000, {style: 'currency', currency: 'USD'})
// outputs: US$1,000.00Formatting Number using unit
Currently this is part of ES2020 NumberFormat.
We've provided a polyfill here and @formatjs/intl types allow users to pass in a sanctioned unit:
intl.formatNumber(1000, {
style: 'unit',
unit: 'kilobyte',
unitDisplay: 'narrow',
})
// outputs: 1,000kBintl.formatNumber(1000, {
unit: 'fahrenheit',
unitDisplay: 'long',
style: 'unit',
})
// outputs: 1,000 degrees FahrenheitformatPlural
type PluralFormatOptions = {
type?: 'cardinal' | 'ordinal' = 'cardinal'
}
function formatPlural(
value: number,
options?: Intl.PluralFormatOptions
): 'zero' | 'one' | 'two' | 'few' | 'many' | 'other'This function will return a plural category string: "zero", "one", "two", "few", "many", or "other".
It expects a value which can be parsed as a number, and accepts options that conform to PluralFormatOptions.
This is a low-level utility whose output could be provided to a switch statement to select a particular string to display.
intl.formatPlural(1)
// outputs: oneintl.formatPlural(3, {style: 'ordinal'})
// outputs: oneMULTIPLE LANGUAGE SUPPORT
This function should only be used in apps that only need to support one language. If your app supports multiple languages use formatMessage instead.
formatList
BROWSER SUPPORT
This requires Intl.ListFormat which has limited browser support. Please use our polyfill if you plan to support them.
This function allows you to join list of things together in an i18n-safe way. For example, when the locale is en:
intl.formatList(['Me', 'myself', 'I'], {type: 'conjunction'})
// outputs: Me, myself and Iintl.formatList(['5 hours', '3 minutes'], {type: 'unit'})
// outputs: 5 hours, 3 minutesformatDisplayName
BROWSER SUPPORT
This requires Intl.DisplayNames which has limited browser support. Please use our polyfill if you plan to support them.
type FormatDisplayNameOptions = {
style?: 'narrow' | 'short' | 'long'
type?: 'language' | 'region' | 'script' | 'currency'
fallback?: 'code' | 'none'
}
function formatDisplayName(
value: string | number | Record<string, unknown>,
options: FormatDisplayNameOptions
): string | undefinedUsage examples:
intl.formatDisplayName('zh-Hans-SG', {type: 'language'})
// outputs: Simplified Chinese (Singapore)// ISO-15924 four letters script code to localized display name
intl.formatDisplayName('Deva', {type: 'script'})
// outputs: Devanagari// ISO-4217 currency code to localized display name
intl.formatDisplayName('CNY', {type: 'currency'})
// outputs: Chinese Yuan// ISO-3166 two letters region code to localized display name
intl.formatDisplayName('UN', {type: 'region'})
// outputs: United NationsformatMessage
type MessageFormatPrimitiveValue = string | number | boolean | null | undefined
function formatMessage(
descriptor: MessageDescriptor,
values?: Record<string, MessageFormatPrimitiveValue>
): string
function formatMessage(
descriptor: MessageDescriptor,
values?: Record<
string,
MessageFormatPrimitiveValue | JSX.Element | FormatXMLElementFn
>
): string | JSX.Element[]This function will return a formatted message string.
It expects a MessageDescriptor with at least an id property, and accepts a shallow values object which are used to fill placeholders in the message.
If a translated message with the id has been passed to the <IntlProvider> via its messages prop it will be formatted, otherwise it will fallback to formatting defaultMessage.
See: Message Formatting Fallbacks for more details.
function () {
const messages = defineMessages({
greeting: {
id: 'app.greeting',
defaultMessage: 'Hello, {name}!',
},
})
return intl.formatMessage(messages.greeting, {name: 'Eric'})
// outputs: Hello, Eric!
}With JSX.Element
function () {
const messages = defineMessages({
greeting: {
id: 'app.greeting',
defaultMessage: 'Hello, {name}!',
},
})
return intl.formatMessage(messages.greeting, {name: <b>Eric</b>})
// outputs: Hello, <b>Eric</b>!
}With rich text formatting
function () {
const messages = defineMessages({
greeting: {
id: 'app.greeting',
defaultMessage: 'Hello, {name}!',
},
})
return intl.formatMessage(messages.greeting, {
name: 'Eric',
bold: str => <b>{str}</b>,
})
// outputs: Hello, <b>Eric</b>!
}defineMessages/defineMessage
interface MessageDescriptor {
id?: string
description?: string | object
defaultMessage?: string
}
function defineMessages(
messageDescriptors: Record<string, MessageDescriptor>
): Record<string, MessageDescriptor>
function defineMessage(messageDescriptor: MessageDescriptor): MessageDescriptorThese functions are re-exported from the @formatjs/intl package and are simply a hook for our CLI & babel/TS plugin to use when compiling default messages defined in JavaScript source files.
This function simply returns the Message Descriptor map object that's passed-in.
import {defineMessages, defineMessage} from '@cookbook/solid-intl'
const messages = defineMessages({
greeting: {
id: 'app.home.greeting',
defaultMessage: 'Hello, {name}!',
},
})
const msg = defineMessage({
id: 'single',
description: 'header',
})Message Extraction
Now that you've declared some messages, it's time to extract them.
npm i @formatjs/cli
# or
yarn add @formatjs/cli
# or
pnpm add @formatjs/cliExtraction
Add the following command to your package.json scripts:
{
"scripts": {
"extract": "formatjs extract"
}
}And execute it:
npm run extract -- 'src/**/*.ts*' --ignore='**/*.d.ts' --out-file lang/en.json'Given a file that has the following messages:
import { useIntl } from '@cookbook/solid-intl'
export function List(props) {
const intl = useIntl();
return (
<section>
<header>
{intl.formatMessage({
id: 'settings.list.header',
defaultMessage: 'Control Panel',
})}
</header>
<ul>
<li>
<button>
{intl.formatMessage({
id: 'settings.list.button.delete',
defaultMessage: 'Delete user {name}',
}, {
name: props.name,
})}
</button>
</li>
<PasswordChange />
</ul>
</section>
)
}running the above command will create a file called `lang/en.json:
{
"settings.list.header": "Control Panel",
"settings.list.button.delete": "Delete user {name}"
}Basic Internationalization Principles
What Is Internationalization and Why Does It Matter?
Internationalized software supports the languages and cultural customs of people throughout the world. The Web reaches all parts of the world. Internationalized web apps provide a great user experience for people everywhere.
Localized software adapts to a specific language and culture by translating text into the user's language and formatting data in accordance with the user's expectations. An app is typically localized for a small set of locales.
The ECMA-402 JavaScript internationalization specification has an excellent overview.
Locales: Language and Region
A "locale" refers to the lingual and cultural expectations for a region. It is represented using a "locale code" defined in UTS LDML.
This code is comprised of several parts separated by hyphens (-). The first part is a short string representing the language. The second, optional, part is a short string representing the region. Additionally, various extensions and variants can be specified.
Typically, web apps are localized to just the language or language-region combination. Examples of such locale codes are:
enfor Englishen-USfor English as spoken in the United Statesen-GBfor English as spoken in the United Kingdomes-ARfor Spanish as spoken in Argentinaar-001for Arabic as spoken throughout the worldar-AEfor Arabic as spoken in United Arab EmiratesMost internationalized apps only support a small list of locales.
Translating Strings
You likely have some text in your application that is in a natural language such as English or Japanese. In order to support other locales, you will need to translate these strings.
FormatJS provides a mechanism to let you write the core "software" of your application without special code for different translations. The considerations for each locale are encapsulated in your translated strings and our libraries.
const messages = {
en: {
GREETING: 'Hello {name}',
},
fr: {
GREETING: 'Bonjour {name}',
},
}We use the ICU Message syntax which is also used in Java, C, PHP and various other platforms.
Message Syntax
String/Message formatting is a paramount feature of Solid Intl and it builds on ICU Message Formatting by using the ICU Message Syntax. This message syntax allows for simple to complex messages to be defined, translated, and then formatted at runtime.
Simple Message:
Hello everyoneSimple Argument
You can use a {key} argument for placing a value into the message.
The key is looked up in the input data, and the string is interpolated with its value.
Hello, {name}Complex Message:
Hello, {name}, you have {itemCount, plural,
=0 {no items}
one {# item}
other {# items}
}.Message Descriptor
Solid Intl has a Message Descriptor concept which is used to define your app's default messages/strings. The Message Descriptors work very well for providing the data necessary for having the strings/messages translated, and they contain the following properties:
id: A unique, stable identifier for the messagedescription: Context for the translator about how it's used in the UIdefaultMessage: The default message (probably in English)
interface MessageDescriptor {
id?: MessageIds;
description?: string | object;
defaultMessage?: string | MessageFormatElement[];
}Message Formatting Fallbacks
The message formatting APIs go the extra mile to provide fallbacks for the common situations where formatting fails; at the very least a non-empty string should always be returned.
Here's the message formatting fallback algorithm:
- Lookup and format the translated message at id, passed to .
- Fallback to formatting the defaultMessage.
- Fallback to translated message at id's source.
- Fallback to defaultMessage source.
- Fallback to the literal message id.
Formatting Message Argument
Values can also be formatted based on their type. You use a {key, type, format} argument to do that.
The elements of the argument are:
keyis where in the input data to find the datatypeis how to interpret the value (see below)formatis optional, and is a further refinement on how to display that type of data
I have {numCats, number} cats.number formatting
This type is used to format numbers in a way that is sensitive to the locale.
It understands the following values for the optional format element of the argument:
I have {numCats, number} cats.
Almost {pctBlack, number, ::percent} of them are black.Internally it uses the Intl.NumberFormat API. You can define custom values for the format element, which are passed to the Intl.NumberFormat constructor.
Sometimes embedding how the number will be formatted provides great context to translators. We also support ICU Number Skeletons using the same syntax:
The price of this bagel is {num, number, ::sign-always compact-short currency/GBP}You can read more about this here.
date formatting
This type is used to format dates in a way that is sensitive to the locale. It understands the following values for the optional format element of the argument:
shortis used to format dates in the shortest possible waymediumis used to format dates with short textual representation of the monthlongis used to format dates with long textual representation of the monthfullis used to format dates with the most detail
Sale begins {start, date, medium}Internally it uses the Intl.DateTimeFormat API. You can define custom values for the format element, which are passed to the Intl.DateTimeFormat constructor.
time formatting
This type is used to format times in a way that is sensitive to the locale. It understands the following values for the optional format element of the argument:
shortis used to format times with hours and minutesmediumis used to format times with hours, minutes, and secondslongis used to format times with hours, minutes, seconds, and timezonefullis the same as long
Coupon expires at {expires, time, short}Internally it uses the Intl.DateTimeFormat API. You can define custom values for the format element, which are passed to the Intl.DateTimeFormat constructor.
Supported DateTime Skeleton
Similar to number type, we also support ICU DateTime skeleton.
ICU provides a wide array of pattern to customize date time format.
However, not all of them are available via ECMA402's Intl API.
Therefore, we only support the following patterns
| Symbol | Meaning | Notes |
|---|---|---|
| G | Era designator | |
| y | year | |
| M | month in year | |
| L | stand-alone month in year | |
| d | day in month | |
| E | day of week | |
| e | local day of week | e..eee is not supported |
| c | stand-alone local day of week | c..ccc is not supported |
| a | AM/PM marker | |
| h | Hour 1-12 | |
| H | Hour 0-23 | |
| K | Hour 0-11 | |
| k | Hour 1-24 | |
| m | Minute | |
| s | Second | |
| z | Time Zone |
{select} Format
The {key, select, matches} is used to choose output by matching a value to one of many choices.
(It is similar to the switch statement available in some programming languages.)
The key is looked up in the input data.
The corresponding value is matched to one of matches and the corresponding output is returned.
The key argument must follow Unicode Pattern_Syntax.
The matches is a space-separated list of matches.
The format of a match is match {output}.
(A match is similar to the case statement of the switch found in some programming languages.)
The match is a literal value.
If it is the same as the value for key then the corresponding output will be used.
output is itself a message, so it can be a literal string or also have more arguments nested inside of it.
The other match is special and is used if nothing else matches. (This is similar to the default case of the switch found in some programming languages.)
DANGER
otheris required as pericu4jimplementation. We will throw an error ifselectis used withoutother.
{gender, select,
male {He}
female {She}
other {They}
} will respond shortly.Here's an example of nested arguments.
{taxableArea, select,
yes {An additional {taxRate, number, percent} tax will be collected.}
other {No taxes apply.}
}{plural} Format
The {key, plural, matches} is used to choose output based on the pluralization rules of the current locale. It is very similar to the {select} format above except that the value is expected to be a number and is mapped to a plural category.
The match is a literal value and is matched to one of these plural categories. Not all languages use all plural categories.
zero: This category is used for languages that have grammar specialized specifically for zero number of items. (Examples are Arabic and Latvian.)one: This category is used for languages that have grammar specialized specifically for one item. Many languages, but not all, use this plural category. (Many popular Asian languages, such as Chinese and Japanese, do not use this category.)two: This category is used for languages that have grammar specialized specifically for two items. (Examples are Arabic and Welsh.)few: This category is used for languages that have grammar specialized specifically for a small number of items. For some languages this is used for 2-4 items, for some 3-10 items, and other languages have even more complex rules.many: This category is used for languages that have grammar specialized specifically for a larger number of items. (Examples are Arabic, Polish, and Russian.)other: This category is used if the value doesn't match one of the other plural categories. Note that this is used for "plural" for languages (such as English) that have a simple "singular" versus "plural" dichotomy.=value: This is used to match a specific value regardless of the plural categories of the current locale.
DANGER
otheris required as pericu4jimplementation. We will throw an error ifselectis used withoutother.
Cart: {itemCount} {itemCount, plural,
one {item}
other {items}
}You have {itemCount, plural,
=0 {no items}
one {1 item}
other {{itemCount} items}
}.In the output of the match, the # special token can be used as a placeholder for the numeric value and will be formatted as if it were {key, number}.
{selectordinal} Format
The {key, selectordinal, matches} is used to choose output based on the ordinal pluralization rules (1st, 2nd, 3rd, etc.) of the current locale. It is very similar to the {plural} format above except that the value is mapped to an ordinal plural category.
The match is a literal value and is matched to one of these plural categories. Not all languages use all plural categories.
zero: This category is used for languages that have grammar specialized specifically for zero number of items. (Examples are Arabic and Latvian.)one: This category is used for languages that have grammar specialized specifically for one item. Many languages, but not all, use this plural category. (Many popular Asian languages, such as Chinese and Japanese, do not use this category.)two: This category is used for languages that have grammar specialized specifically for two items. (Examples are Arabic and Welsh.)few: This category is used for languages that have grammar specialized specifically for a small number of items. For some languages this is used for 2-4 items, for some 3-10 items, and other languages have even more complex rules.many: This category is used for languages that have grammar specialized specifically for a larger number of items. (Examples are Arabic, Polish, and Russian.)other: This category is used if the value doesn't match one of the other plural categories. Note that this is used for "plural" for languages (such as English) that have a simple "singular" versus "plural" dichotomy.=value: This is used to match a specific value regardless of the plural categories of the current locale.
DANGER
otheris required as pericu4jimplementation. We will throw an error ifselectis used withoutother.
In the output of the match, the # special token can be used as a placeholder for the numeric value and will be formatted as if it were {key, number}.
It's my cat's {year, selectordinal,
one {#st}
two {#nd}
few {#rd}
other {#th}
} birthday!Rich Text Formatting
We also support embedded rich text formatting in our message using tags. This allows developers to embed as much text as possible so sentences don't have to be broken up into chunks
NOTE: This is not XML/HTML tag
Our price is <boldThis>{price, number, ::currency/USD precision-integer}</boldThis>
with <link>{pct, number, ::percent} discount</link>Quoting / Escaping
The ASCII apostrophe ' (U+0027) can be used to escape syntax characters in the text portion of the message. A single apostrophe escapes one syntax character immediately following it; a pair of apostrophes escape every syntax characters wrapped between them:
"This is not an interpolation: '{word}"
//→ "This is not an interpolation: {word}"
"These are not interpolations: '{word1} {word2}'"
//→ "These are not interpolations: {word1} {word2}"
"'<notATag>"
//→ "<notATag>"
"'<notATag>hello</notATag>'"
//→ "<notATag>hello</notATag>"Two consecutive ASCII apostrophes represents one ASCII apostrophe, similar to %% in printf represents one %. However, we recommend using curly apostrophe ’ (U+2019) for human-readable strings and only use ASCII apostrophe ' (U+0027) in ICU message syntax.
"This '{isn''t}' obvious."
//→ "This {isn't} obvious."