/DeepLy

PHP client for the DeepL.com translation API v2

Primary LanguagePHPMIT LicenseMIT

DeepLy

DeepL.com is a great, new translation service. It provides better translations compared to other popular translation engines. DeepLy is a PHP package that implements a client to interact with DeepL via their API, with optional Laravel integration.

Installation

Through Composer:

composer require octfx/deeply

From then on you may run composer update to get the latest version of this library.

It is possible to use this library without using Composer but then it is necessary to register an autoloader function.

This library requires PHP 7.2 or higher, the mbstring, and json extension.

If you do not have an API key, now is the time to get one. Vist DeepL-com to request an API key.

Usage Example

$deepLy = new Octfx\DeepLy\DeepLy('Your-API-Key');

$translatedText = $deepLy->translate('Hello world!', 'DE', 'EN');
    
echo $translatedText; // Prints "Hallo Welt!"

Sophisticated Example

use Octfx\DeepLy\DeepLy;

$apiKey = 'Your-API-Key';
$deepLy = new DeepLy($apiKey);

try {
    $translatedText = $deepLy->translate('Hello world!', DeepLy::LANG_EN, DeepLy::LANG_AUTO);
    
    echo $translatedText; // Prints "Hallo Welt!"
} catch (AuthenticationException $e) {
    // API Key invalid
    // Code 403
} catch (QuotaException $e) {
    // Quota exceeded
    // Code 456
} catch (RateLimitedException $e) {
    // Ratelimited
    // Code 429
} catch (TextLengthException $e) {
     // Textlength > 30000 chars
} catch (CallException $e) {
    // Other errors
    // See: https://www.deepl.com/docs-api.html?part=accessing
}

Always wrap calls of the translate method in a try-catch-block, because they might throw an exception if the arguments are invalid or the API call fails. Instead of using hardcoded strings as language arguments better use the language code constants of the DeepLy class. The class also offers methods such as supportsSourceLangCode($langCode) and supportsTargetLangCode($langCode).

Translation formality

The DeepL API allows to specify the formality of the translated text.
This feature currently works for all target languages except "ES" (Spanish), "JA" (Japanese) and "ZH" (Chinese).
Possible options are:

  • "default" (default)
  • "more" - for a more formal language
  • "less" - for a more informal language
$deepLy->formality('less');

// or
$translatedText = $deepLy->translate('Hello world!', DeepLy::LANG_EN, DeepLy::LANG_AUTO, 'more');

Auto-Detect Language

DeepLy has a method that uses the DeepL API to detect the language of a text:

$languageCode = $deepLy->detectLanguage('Hello world!');

This will return 'EN'. The language of the text has to be one of the supported languages or the result will be incorrect. If you do not need the code of the language but its name, you may call the $deepLy->getLangName($langCode) method.

The API in general can handle and completely translate texts that contain parts with different languages, if the language switch is not within a sentence. The detectLanguage() method will however only return the code of one language. It will throw an exception if it is unable to auto-detect the language. This will rarely happen, it is more likely that the API will return a "false positive": It will rather detect the wrong language than no language at all.

Supported Languages

All supported languages by DeepL can be requested by calling

$deepLy->getSupportedLanguages('source'); // all supported source languages
$deepLy->getSupportedLanguages('target'); // all supported target languages

DeepL(y) supports these source languages:

Code Language
auto Auto detect
DE German
EN English
FR French
IT Italian
JA Japanese
ES Spanish
NL Dutch
PL Polish
PT Portuguese
RU Russian
ZH Chinese

DeepL(y) supports these target languages:

Code Language
DE German
EN-GB English (British)
EN-US English (American)
EN DEPRECATED
FR French
IT Italian
JA Japanese
ES Spanish
NL Dutch
PL Polish
PT DEPRECATED
PT-PT Portuguese (all)
PT-BR Portuguese (Brazilian)
RU Russian
ZH Chinese

Note that auto detection only is possible for the source language.

Text Length Limit

According to the DeepL.com website, the length of the text that has to be translated is limited to 30000 characters. Per default DeepLy will throw an exception if the length limit is exceeded. You may call $deepLy->setValidateTextLength(false) to disable this validation.

HTTP Client

Per default DeepLy uses a HTTP client based on Guzzle. If you want to use a different HTTP client, create a class that implements the HttpClient\HttpClientInterface and makes use of the methods of the alternative HTTP client. Then use $deepLy->setHttpClient($yourHttpClient) to inject it.

Framework Integration

DeepLy comes with support for Laravel 5.x - 7.x and since it also supports package auto-discovery it will be auto-detected in Laravel 5.5 - 7.x

In Laravel 5.0-5.4 you manually have to register the service provider Octfx\DeepLy\Integrations\Laravel\DeepLyServiceProvider in the "providers" array and the facade Octfx\DeepLy\Integrations\Laravel\DeepLyFacade as an alias in the "aliases" array in your config/app.php config file.

DeepLy uses config('services.deepl.auth_key') to retrieve the API key, so you have to set it in the services.php settings.

return [
    'deepl' => [
        'auth_key' => env('DEEPL_AUTH_KEY'),
    ],
];

Request Limit

There is a request limit. The threshold of this limit is specified in the quota documentation delivered to you by DeepL.

Internals

The "core" of this library consists of these classes:

  • DeepLy - main class
  • HttpClient\GuzzleHttpClient - HTTP client class
  • Protocol\JsonProtocol - JSON is the protocol used by the DeepL API
  • ResponseBag\AbstractBag - base wrapper class for the responses of the DeepL API
  • ResponseBag\UsageBag - concrete class for API responses to usage statistics requests
  • ResponseBag\TranslationBag - concrete class for API responses to "translate" requests

There are also some exception classes, interfaces, an alternative HTTP client implementation that uses Guzzle and classes for the Laravel integration.

Disclaimer

This is not an official package. It is 100% open source and non-commercial.

DeepL is a product of DeepL GmbH. More info: deepl.com/publisher.html