From 2fe79719f8b83ebc4701a1c0fb4ca232de9b2aea Mon Sep 17 00:00:00 2001
From: Demian Katz <demian.katz@villanova.edu>
Date: Wed, 23 Jan 2013 09:42:58 -0500
Subject: [PATCH] Resolving VUFIND-739 (Wikipedia circular redirect causes
infinite loop).
---
.../src/VuFind/Recommend/AuthorInfo.php | 62 +++++++++++++++----
1 file changed, 51 insertions(+), 11 deletions(-)
diff --git a/module/VuFind/src/VuFind/Recommend/AuthorInfo.php b/module/VuFind/src/VuFind/Recommend/AuthorInfo.php
index de0e938429c..e0ee8f14543 100644
--- a/module/VuFind/src/VuFind/Recommend/AuthorInfo.php
+++ b/module/VuFind/src/VuFind/Recommend/AuthorInfo.php
@@ -86,6 +86,13 @@ class AuthorInfo implements RecommendInterface, TranslatorAwareInterface
*/
protected $useViaf = false;
+ /**
+ * Log of Wikipedia pages already retrieved
+ *
+ * @var array
+ */
+ protected $pagesRetrieved = array();
+
/**
* Constructor
*
@@ -179,6 +186,23 @@ class AuthorInfo implements RecommendInterface, TranslatorAwareInterface
$this->searchObject = $results;
}
+ /**
+ * Check if a page has already been retrieved; if it hasn't, flag it as
+ * retrieved for future reference.
+ *
+ * @param string $author Author being retrieved
+ *
+ * @return bool
+ */
+ protected function alreadyRetrieved($author)
+ {
+ if (isset($this->pagesRetrieved[$author])) {
+ return true;
+ }
+ $this->pagesRetrieved[$author] = true;
+ return false;
+ }
+
/**
* Returns info from Wikipedia to the view
*
@@ -218,6 +242,12 @@ class AuthorInfo implements RecommendInterface, TranslatorAwareInterface
*/
protected function getWikipedia($author)
{
+ // Don't retrieve the same page multiple times; this indicates a loop
+ // that needs to be broken!
+ if ($this->alreadyRetrieved($author)) {
+ return array();
+ }
+
// Get information from Wikipedia API
$uri = 'http://' . $this->lang . '.wikipedia.org/w/api.php' .
'?action=query&prop=revisions&rvprop=content&format=php' .
@@ -248,18 +278,28 @@ class AuthorInfo implements RecommendInterface, TranslatorAwareInterface
return null;
}
- // Get the default page
- $body = array_shift($body['query']['pages']);
- $info = array('name' => $body['title'], 'wiki_lang' => $this->lang);
-
- // Get the latest revision
- $body = array_shift($body['revisions']);
- // Check for redirection
- $as_lines = explode("\n", $body['*']);
- if (stristr($as_lines[0], '#REDIRECT')) {
- preg_match('/\[\[(.*)\]\]/', $as_lines[0], $matches);
- return $this->getWikipedia($matches[1]);
+ // Loop through the pages and find the first that isn't a redirect:
+ foreach ($body['query']['pages'] as $page) {
+ $info['name'] = $page['title'];
+
+ // Get the latest revision
+ $page = array_shift($page['revisions']);
+ // Check for redirection
+ $as_lines = explode("\n", $page['*']);
+ if (stristr($as_lines[0], '#REDIRECT')) {
+ preg_match('/\[\[(.*)\]\]/', $as_lines[0], $matches);
+ $redirectTo = $matches[1];
+ } else {
+ $redirectTo = false;
+ break;
+ }
+ }
+
+ // Recurse if we only found redirects:
+ if ($redirectTo) {
+ return $this->getWikipedia($redirectTo);
}
+ $body = $page;
/* Infobox */
--
GitLab