Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 79 |
|
0.00% |
0 / 8 |
CRAP | |
0.00% |
0 / 1 |
| MicrosoftWebService | |
0.00% |
0 / 79 |
|
0.00% |
0 / 8 |
306 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| getType | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| mapCode | |
0.00% |
0 / 9 |
|
0.00% |
0 / 1 |
2 | |||
| doPairs | |
0.00% |
0 / 27 |
|
0.00% |
0 / 1 |
42 | |||
| getQuery | |
0.00% |
0 / 26 |
|
0.00% |
0 / 1 |
20 | |||
| parseResponse | |
0.00% |
0 / 9 |
|
0.00% |
0 / 1 |
6 | |||
| wrapUntranslatable | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| unwrapUntranslatable | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | declare( strict_types = 1 ); |
| 3 | |
| 4 | namespace MediaWiki\Extension\Translate\WebService; |
| 5 | |
| 6 | use MediaWiki\Http\HttpRequestFactory; |
| 7 | |
| 8 | /** |
| 9 | * Implements support for Microsoft translation api v3. |
| 10 | * @author Niklas Laxström |
| 11 | * @author Ulrich Strauss |
| 12 | * @license GPL-2.0-or-later |
| 13 | * @since 2013-01-01 |
| 14 | * @see https://docs.microsoft.com/fi-fi/azure/cognitive-services/Translator/reference/v3-0-reference |
| 15 | * @ingroup TranslationWebService |
| 16 | */ |
| 17 | class MicrosoftWebService extends TranslationWebService { |
| 18 | private HttpRequestFactory $httpRequestFactory; |
| 19 | |
| 20 | public function __construct( |
| 21 | HttpRequestFactory $httpRequestFactory, |
| 22 | string $serviceName, |
| 23 | array $config |
| 24 | ) { |
| 25 | parent::__construct( $serviceName, $config ); |
| 26 | $this->httpRequestFactory = $httpRequestFactory; |
| 27 | } |
| 28 | |
| 29 | /** @inheritDoc */ |
| 30 | public function getType(): string { |
| 31 | return 'mt'; |
| 32 | } |
| 33 | |
| 34 | /** @inheritDoc */ |
| 35 | protected function mapCode( string $code ): string { |
| 36 | $map = [ |
| 37 | 'tl' => 'fil', |
| 38 | 'zh-hant' => 'zh-Hant', |
| 39 | 'zh-hans' => 'zh-Hans', |
| 40 | 'sr-ec' => 'sr-Cyrl', |
| 41 | 'sr-el' => 'sr-Latn', |
| 42 | 'pt-br' => 'pt', |
| 43 | ]; |
| 44 | |
| 45 | return $map[$code] ?? $code; |
| 46 | } |
| 47 | |
| 48 | /** @inheritDoc */ |
| 49 | protected function doPairs(): array { |
| 50 | if ( !isset( $this->config['key'] ) ) { |
| 51 | throw new TranslationWebServiceConfigurationException( 'key is not set' ); |
| 52 | } |
| 53 | |
| 54 | $key = $this->config['key']; |
| 55 | |
| 56 | $options = []; |
| 57 | $options['method'] = 'GET'; |
| 58 | $options['timeout'] = $this->config['timeout']; |
| 59 | |
| 60 | $url = $this->config['url'] . '/languages?api-version=3.0'; |
| 61 | |
| 62 | $req = $this->httpRequestFactory->create( $url, $options, __METHOD__ ); |
| 63 | $req->setHeader( 'Ocp-Apim-Subscription-Key', $key ); |
| 64 | |
| 65 | $status = $req->execute(); |
| 66 | if ( !$status->isOK() ) { |
| 67 | $error = $req->getContent(); |
| 68 | // Most likely a timeout or other general error |
| 69 | throw new TranslationWebServiceException( |
| 70 | 'HttpRequestFactory::get failed:' . serialize( $error ) . serialize( $status ) |
| 71 | ); |
| 72 | } |
| 73 | |
| 74 | $json = $req->getContent(); |
| 75 | $response = json_decode( $json, true ); |
| 76 | if ( !isset( $response[ 'translation' ] ) ) { |
| 77 | throw new TranslationWebServiceException( |
| 78 | 'Unable to fetch list of available languages: ' . $json |
| 79 | ); |
| 80 | } |
| 81 | |
| 82 | $languages = array_keys( $response[ 'translation' ] ); |
| 83 | |
| 84 | // Let's make a cartesian product, assuming we can translate from any language to any language |
| 85 | $pairs = []; |
| 86 | foreach ( $languages as $from ) { |
| 87 | foreach ( $languages as $to ) { |
| 88 | $pairs[$from][$to] = true; |
| 89 | } |
| 90 | } |
| 91 | |
| 92 | return $pairs; |
| 93 | } |
| 94 | |
| 95 | /** @inheritDoc */ |
| 96 | protected function getQuery( string $text, string $sourceLanguage, string $targetLanguage ): TranslationQuery { |
| 97 | if ( !isset( $this->config['key'] ) ) { |
| 98 | throw new TranslationWebServiceConfigurationException( 'key is not set' ); |
| 99 | } |
| 100 | |
| 101 | $key = $this->config['key']; |
| 102 | $text = trim( $text ); |
| 103 | $text = $this->wrapUntranslatable( $text ); |
| 104 | |
| 105 | $url = $this->config['url'] . '/translate'; |
| 106 | $params = [ |
| 107 | 'api-version' => '3.0', |
| 108 | 'from' => $sourceLanguage, |
| 109 | 'to' => $targetLanguage, |
| 110 | 'textType' => 'html', |
| 111 | ]; |
| 112 | $headers = [ |
| 113 | 'Ocp-Apim-Subscription-Key' => $key, |
| 114 | 'Content-Type' => 'application/json', |
| 115 | ]; |
| 116 | $body = json_encode( [ [ 'Text' => $text ] ] ); |
| 117 | |
| 118 | if ( $body === false ) { |
| 119 | throw new TranslationWebServiceInvalidInputException( 'Could not JSON encode source text' ); |
| 120 | } |
| 121 | |
| 122 | if ( strlen( $body ) > 5000 ) { |
| 123 | throw new TranslationWebServiceInvalidInputException( 'Source text too long' ); |
| 124 | } |
| 125 | |
| 126 | return TranslationQuery::factory( $url ) |
| 127 | ->timeout( intval( $this->config['timeout'] ) ) |
| 128 | ->queryParameters( $params ) |
| 129 | ->queryHeaders( $headers ) |
| 130 | ->postWithData( $body ); |
| 131 | } |
| 132 | |
| 133 | /** @inheritDoc */ |
| 134 | protected function parseResponse( TranslationQueryResponse $reply ): string { |
| 135 | $body = $reply->getBody(); |
| 136 | |
| 137 | $response = json_decode( $body, true ); |
| 138 | if ( !isset( $response[ 0 ][ 'translations' ][ 0 ][ 'text' ] ) ) { |
| 139 | throw new TranslationWebServiceException( |
| 140 | 'Unable to parse translation response: ' . $body |
| 141 | ); |
| 142 | } |
| 143 | |
| 144 | $text = $response[ 0 ][ 'translations' ][ 0 ][ 'text' ]; |
| 145 | $text = $this->unwrapUntranslatable( $text ); |
| 146 | |
| 147 | return $text; |
| 148 | } |
| 149 | |
| 150 | /** @inheritDoc */ |
| 151 | protected function wrapUntranslatable( string $text ): string { |
| 152 | $pattern = '~%[^% ]+%|\$\d|{VAR:[^}]+}|{?{(PLURAL|GRAMMAR|GENDER):[^|]+\||%(\d\$)?[sd]~'; |
| 153 | $wrap = '<span class="notranslate">\0</span>'; |
| 154 | return preg_replace( $pattern, $wrap, $text ); |
| 155 | } |
| 156 | |
| 157 | /** @inheritDoc */ |
| 158 | protected function unwrapUntranslatable( string $text ): string { |
| 159 | $pattern = '~<span class="notranslate">\s*(.*?)\s*</span>~'; |
| 160 | return preg_replace( $pattern, '\1', $text ); |
| 161 | } |
| 162 | } |