From c964053c91e0a0cc8343c06bd4b2c2ee32955bac Mon Sep 17 00:00:00 2001
From: Joshua Ramon Enslin <joshua@museum-digital.de>
Date: Thu, 18 Mar 2021 01:23:45 +0100
Subject: [PATCH] Add function for reading Wikidata ID from a Wikipedia page

---
 src/NodaWikidataFetcher.php | 55 +++++++++++++++++++++++++++++++++----
 1 file changed, 50 insertions(+), 5 deletions(-)

diff --git a/src/NodaWikidataFetcher.php b/src/NodaWikidataFetcher.php
index 8c7f6c9..aa6e853 100644
--- a/src/NodaWikidataFetcher.php
+++ b/src/NodaWikidataFetcher.php
@@ -10,7 +10,7 @@ declare(strict_types = 1);
 /**
  * Helps fetching information from Wikidata.
  */
-class NodaWikidataFetcher {
+final class NodaWikidataFetcher {
 
     const LANGUAGES_MAIN_DESC = ['de', 'da', 'en', 'es', 'fr', 'hu', 'it', 'jp', 'nl', 'pt', 'ru', 'sv', 'zh'];
     const LANGUAGES_TO_CHECK = ['ar', 'bn', 'cs', 'da', 'de', 'el', 'en', 'es', 'fa', 'fi', 'fr', 'ha', 'he', 'hi', 'hu', 'id', 'it', 'ja', 'ka', 'ko', 'nl', 'pl', 'pt', 'ro', 'ru', 'sv', 'sw', 'ta', 'th', 'tl', 'tr', 'ur', 'vi', 'zh'];
@@ -42,7 +42,6 @@ class NodaWikidataFetcher {
         'lcsh' => 'P244',
         'aat' => 'P1014',
         'iconclass' => 'P1256',
-        'rkd' => 'P650',
         'osm' => 'P402',
         'loc' => 'P244',
         'nomisma' => 'P2950',
@@ -81,6 +80,52 @@ class NodaWikidataFetcher {
     /** @var MDMysqli */
     private MDMysqli $_mysqli_noda;
 
+    /**
+     * Attempts to fetch a Wikidata ID from a provided URL.
+     *
+     * @param string $linkUrl Link to a Wikipedia page.
+     *
+     * @return string
+     */
+    public static function getWikidataIdFromWikipedia(string $linkUrl):string {
+
+        if (!filter_var($linkUrl, FILTER_VALIDATE_URL)) {
+            throw new MDExpectedException("Invalid URL");
+        }
+
+        $wikipedia_cont = MD_STD::runCurl($linkUrl);
+
+        $doc = new DOMDocument();
+        if (!($doc->loadHTML($wikipedia_cont))) {
+            return '';
+        }
+
+        if (!($wikidataLinkLi = $doc->getElementById("t-wikibase"))) {
+            return '';
+        }
+        if (!($wikidataLink = $wikidataLinkLi->firstChild)) {
+            return '';
+        }
+
+        if (!($t_wikibase_href = $wikidataLink->getAttribute('href'))) {
+            return '';
+        }
+        $t_wikibase = (string)$t_wikibase_href;
+
+        if (!empty($t_wikibase)) {
+            if (($wikidata_id_end = strrpos($t_wikibase, '/')) !== false) {
+                $wikidata_id = trim(substr($t_wikibase, $wikidata_id_end + 1), '/');
+
+                if (substr($wikidata_id, 0, 1) === 'Q') {
+                    return $wikidata_id;
+                }
+            }
+        }
+
+        return '';
+
+    }
+
     /**
      * Gets translation source Wikipedia pages from Wikidata.
      *
@@ -1162,9 +1207,9 @@ class NodaWikidataFetcher {
     /**
      * Writes relations to norm data sources to DB.
      *
-     * @param array<array<string>> $nodaLinks   Links to other noda sources.
-     * @param integer              $tag_id      Tag ID.
-     * @param string               $erfasst_von Name of the user to edit this.
+     * @param array<string> $nodaLinks   Links to other noda sources.
+     * @param integer       $tag_id      Tag ID.
+     * @param string        $erfasst_von Name of the user to edit this.
      *
      * @return void
      */