App\Domain\Publisher\ExternMapper

Complexity

Total Complexity

77

Size/Duplication

Total Lines	307
Duplicated Lines	0 %

Importance

Changes

0

17 Methods

Rating	Name	Duplication	Size	Complexity
A	clean()	0	9	2
A	processJsonMapping()	0	13	5
A	processMapping()	0	22	6
B	authorsEtAl()	0	16	7
A	wikifyPressAgency()	0	19	3
A	isAnArticle()	0	7	2
A	checkJSONLD()	0	3	2
A	convertInstitutionnel()	0	9	5
A	process()	0	5	2
A	__construct()	0	3	1
A	convertOGtype2format()	0	14	4
A	convertDate()	0	20	4
B	convertURLaccess()	0	14	7
A	convertDCpage()	0	12	3
A	postProcess()	0	13	5
C	convertAuteur()	0	35	16
A	convertLangue()	0	11	3

<?php
/**
 * This file is part of dispositif/wikibot application
 * 2019 © Philippe M. <dispositif@gmail.com>
 * For the full copyright and MIT license information, please view the LICENSE file.
 */

declare(strict_types=1);

namespace App\Domain\Publisher;

use App\Domain\Enums\Language;
use App\Domain\Utils\ArrayProcessTrait;
use DateTime;
use Exception;
use Psr\Log\LoggerInterface;

/**
 * Generic mapper for press/revue article on web.
 * Using JSON-LD and meta tags to obtain {article} data.
 * Generic mapper for web pages URL to wiki-template references.
 * Converting to {article}, {lien web} or {lien brisé}
 * Using JSON-LD, Open Graph and Dublin Core meta extracted from HTML.
 * Class ExternMapper
 *
 * @package App\Domain\Publisher
 */
class ExternMapper implements MapperInterface
{
    use ArrayProcessTrait, ExternOGMapperTrait, ExternLDMapperTrait;

    /**
     * @var LoggerInterface
     */
    private $log;

    public function __construct(LoggerInterface $log)
    {
        $this->log = $log;
    }

    public function process($data): array
    {
        $dat = $this->processMapping($data);

        return (!empty($dat)) ? $this->postProcess($dat) : [];
    }

    protected function processMapping($data): array
    {
        $mapJson = [];
        $mapMeta = [];
        if (!empty($data['JSON-LD'])) {
            $mapJson = $this->processJsonMapping($data['JSON-LD']);
        }
        if (!empty($data['meta'])) {
            $mapMeta = $this->mapLienwebFromMeta($data['meta']);
        }

        // langue absente JSON-LD mais array_merge risqué (doublon)
        if (!empty($mapJson)) {
            if (!isset($mapJson['langue']) && isset($mapMeta['langue'])) {
                $mapJson['langue'] = $mapMeta['langue'];
                $mapJson['DATA-TYPE'] = 'JSON-LD+META';
            }

            return $mapJson;
        }

        return $mapMeta;
    }

    /**
     * todo move to mapper ?
     *
     * @param array $LDdata
     *
     * @return array
     */
    private function processJsonMapping(array $LDdata): array
    {
        if ($this->checkJSONLD($LDdata)) {
            return $this->mapArticleDataFromJSONLD($LDdata);
        }
        // gestion des multiples objets comme Figaro
        foreach ($LDdata as $dat) {
            if (is_array($dat) && $this->checkJSONLD($dat)) {
                return $this->mapArticleDataFromJSONLD($dat);
            }
        }

        return [];
    }

    protected function checkJSONLD(array $jsonLD): bool
    {
        return isset($jsonLD['headline']) && isset($jsonLD['@type']);
    }

    /**
     * todo Refac/move domain special mapping
     * todo Config parameter for post-process
     *
     * @param array $dat
     *
     * @return array
     */
    protected function postProcess(array $dat): array
    {
        $dat = $this->deleteEmptyValueArray($dat);
        if (isset($dat['langue']) && 'fr' === $dat['langue']) {
            unset($dat['langue']);
        }

        // Ça m'énerve ! Gallica met "vidéo" pour livre numérisé
        if (isset($dat['site']) && $dat['site'] === 'Gallica') {
            unset($dat['format']);
        }

        return $dat;
    }

    protected function isAnArticle(?string $str): bool
    {
        if (in_array($str, ['article', 'journalArticle'])) {
            return true;
        }

        return false;
    }

    protected function convertURLaccess($data): ?string
    {
        // NYT, Figaro
        if (isset($data['isAccessibleForFree'])) {
            return $data['isAccessibleForFree'] ? 'ouvert' : 'limité';
        }
        if (isset($data['DC.rights'])) {
            return (in_array($data['DC.rights'], ['free', 'public domain'])) ? 'ouvert' : 'limité';
        }
        if (isset($data['og:article:content_tier'])) {
            return ($data['og:article:content_tier'] === 'free') ? 'ouvert' : 'limité';
        }

        return null;
    }

    /**
     * Réduit le nombre d'auteurs si > 3.
     * En $modeEtAll=true vérification pour "et al.=oui".
     * TODO : wikifyPressAgency()
     *
     * @param string|null $authors
     * @param bool        $modeEtAl
     *
     * @return string|null
     */
    protected function authorsEtAl(?string $authors, $modeEtAl = false): ?string
    {
        if (empty($authors)) {
            return null;
        }
        // conserve juste les 3 premiers auteurs TODO : refactor
        // Bob, Martin ; Yul, Bar ; ... ; ...
        if (preg_match('#([^;]+;[^;]+);[^;]+;.+#', $authors, $matches)) {
            return ($modeEtAl) ? 'oui' : $matches[1];
        }
        // Bob Martin, Yul Bar, ..., ...,...
        if (preg_match('#([^,]+,[^,]+),[^,]+,.+#', $authors, $matches)) {
            return ($modeEtAl) ? 'oui' : $matches[1];
        }

        return ($modeEtAl) ? null : $authors;
    }

    protected function convertDCpage(array $meta): ?string
    {
        if (isset($meta['citation_firstpage'])) {
            $page = $meta['citation_firstpage'];
            if (isset($meta['citation_lastpage'])) {
                $page .= '–'.$meta['citation_lastpage'];
            }

            return (string)$page;
        }

        return null;
    }

    // TODO encodage + normalizer
    public function clean(?string $str = null): ?string
    {
        if ($str === null) {
            return null;
        }
        $str = str_replace(['&#39;','&#039;', '&apos;', "\n", "&#10;", "|", "&eacute;"], ["'","'", "'", '', ' ', '/',
                                                                                          "é"], $str);

        return html_entity_decode($str);
    }

    protected function convertOGtype2format(?string $ogType): ?string
    {
        if (empty($ogType)) {
            return null;
        }
        // og:type = default: website / video.movie / video.tv_show video.other / article, book, profile
        if (strpos($ogType, 'video') !== false) {
            return 'vidéo';
        }
        if (strpos($ogType, 'book') !== false) {
            return 'livre';
        }

        return null;
    }

    /**
     * https://developers.facebook.com/docs/internationalization#locales
     */
    protected function convertLangue(?string $lang = null): ?string
    {
        if (empty($lang)) {
            return null;
        }
        // en_GB
        if (preg_match('#^([a-z]{2})_[A-Z]{2}$#', $lang, $matches)) {
            return $matches[1];
        }

        return Language::all2wiki($lang);
    }

    protected function convertAuteur($data, $indice)
    {
        // author=Bob
        if (isset($data['author']) && is_string($data['author']) && $indice === 1) {
            return html_entity_decode($data['author']);
        }

        // author ['name'=>'Bob','@type'=>'Person']
        if (0 === $indice
            && isset($data['author'])
            && isset($data['author']['name'])
            && (!isset($data['author']['@type'])
                || 'Person' === $data['author']['@type'])
        ) {
            if (is_string($data['author']['name'])) {
                return html_entity_decode($data['author']['name']);
            }

            return html_entity_decode($data['author']['name'][0]);
        }

        // author [ 0 => ['name'=>'Bob'], 1=> ...]
        if (isset($data['author']) && isset($data['author'][$indice])
            && (!isset($data['author'][$indice]['@type'])
                || 'Person' === $data['author'][$indice]['@type'])
        ) {
            if (isset($data['author'][$indice]['name']) && is_string($data['author'][$indice]['name'])) {
                return html_entity_decode($data['author'][$indice]['name']);
            }

            // "author" => [ "@type" => "Person", "name" => [] ]
            return html_entity_decode($data['author'][$indice]['name'][0]);
        }

        return null;
    }

    protected function convertInstitutionnel($data)
    {
        if (isset($data['author']) && isset($data['author'][0]) && isset($data['author'][0]['@type'])
            && 'Person' !== $data['author'][0]['@type']
        ) {
            return html_entity_decode($data['author'][0]['name']);
        }

        return null;
    }

    /**
     * @param string $str
     *
     * @return string
     */
    protected function convertDate(?string $str): ?string
    {
        if (empty($str)) {
            return null;
        }

        // "2012"
        if (preg_match('#^[12][0-9]{3}$#', $str)) {
            return $str;
        }

        try {
            $date = new DateTime($str);
        } catch (Exception $e) {
            dump('EXCEPTION DATE');

            return $str;
        }

        return $date->format('d-m-Y');
    }

    /**
     * Wikification des noms/acronymes d'agences de presse.
     *
     * @param string $str
     *
     * @return string
     */
    protected function wikifyPressAgency(?string $str): ?string
    {
        if (empty($str)) {
            return null;
        }
        // skip potential wikilinks
        if (strpos($str, '[') !== false) {
            return $str;
        }
        $str = preg_replace('#\b(AFP)\b#i', '[[Agence France-Presse|AFP]]', $str);
        $str = str_replace('Reuters', '[[Reuters]]', $str);
        $str = str_replace('Associated Press', '[[Associated Press]]', $str);
        $str = preg_replace('#\b(PA)\b#', '[[Press Association|PA]]', $str);
        $str = preg_replace('#\b(AP)\b#', '[[Associated Press|AP]]', $str);
        $str = str_replace('Xinhua', '[[Xinhua]]', $str);
        $str = preg_replace('#\b(ATS)\b#', '[[Agence télégraphique suisse|ATS]]', $str);
        $str = preg_replace('#\b(PC|CP)\b#', '[[La Presse canadienne|PC]]', $str);

        return $str;
    }
}