From dbb4f92ba06537cdbaac6b3282a31527d02d40bd Mon Sep 17 00:00:00 2001
From: Demian Katz <demian.katz@villanova.edu>
Date: Fri, 8 Nov 2013 08:10:08 -0500
Subject: [PATCH] More refactoring/simplification.
---
.../src/VuFind/Recommend/AuthorInfo.php | 87 ++++++++++---------
1 file changed, 45 insertions(+), 42 deletions(-)
diff --git a/module/VuFind/src/VuFind/Recommend/AuthorInfo.php b/module/VuFind/src/VuFind/Recommend/AuthorInfo.php
index d8f6a23243d..0bd6caecfda 100644
--- a/module/VuFind/src/VuFind/Recommend/AuthorInfo.php
+++ b/module/VuFind/src/VuFind/Recommend/AuthorInfo.php
@@ -341,6 +341,45 @@ class AuthorInfo implements RecommendInterface, TranslatorAwareInterface
return array($imageName, $imageCaption);
}
+ /**
+ * Support method for sanitizeWikipediaBody -- strip image/file links.
+ *
+ * @param string $body The Wikipedia response to sanitize
+ *
+ * @return string
+ */
+ protected function stripImageAndFileLinks($body)
+ {
+ // Remove unwanted image/file links
+ // Nested brackets make this annoying: We can't add 'File' or 'Image' as
+ // mandatory because the recursion fails, or as optional because then
+ // normal links get hit.
+ // ... unless there's a better pattern? TODO
+ // eg. [[File:Johann Sebastian Bach.jpg|thumb|Bach in a 1748 portrait by
+ // [[Elias Gottlob Haussmann|Haussmann]]]]
+ $open = "\\[";
+ $close = "\\]";
+ $content = "(?>[^\\[\\]]+)"; // Anything but [ or ]
+ // We can either find content or recursive brackets:
+ $recursive_match = "($content|(?R))*";
+ $body .= "[[file:bad]]";
+ preg_match_all("/".$open.$recursive_match.$close."/Us", $body, $new_matches);
+ // Loop through every match (link) we found
+ if (is_array($new_matches)) {
+ foreach ($new_matches as $nm) {
+ foreach ((array)$nm as $n) {
+ // If it's a file link get rid of it
+ if (strtolower(substr($n, 0, 7)) == "[[file:"
+ || strtolower(substr($n, 0, 8)) == "[[image:"
+ ) {
+ $body = str_replace($n, "", $body);
+ }
+ }
+ }
+ }
+ return $body;
+ }
+
/**
* Support method for parseWikipedia - fix up details in the body
*
@@ -350,6 +389,12 @@ class AuthorInfo implements RecommendInterface, TranslatorAwareInterface
*/
protected function sanitizeWikipediaBody($body)
{
+ // Cull our content back to everything before the first heading
+ $body = trim(substr($body, 0, strpos($body, "==")));
+
+ // Strip out links
+ $body = $this->stripImageAndFileLinks($body);
+
// Initialize arrays of processing instructions
$pattern = array();
$replacement = array();
@@ -472,48 +517,6 @@ class AuthorInfo implements RecommendInterface, TranslatorAwareInterface
// No infobox -- use whole thing:
$body = $body['*'];
}
- // Find the first heading
- $end = strpos($body, "==");
- // Now cull our content back to everything before the first heading
- $body = trim(substr($body, 0, $end));
-
- // Remove unwanted image/file links
- // Nested brackets make this annoying: We can't add 'File' or 'Image' as
- // mandatory because the recursion fails, or as optional because then
- // normal links get hit.
- // ... unless there's a better pattern? TODO
- // eg. [[File:Johann Sebastian Bach.jpg|thumb|Bach in a 1748 portrait by
- // [[Elias Gottlob Haussmann|Haussmann]]]]
- $open = "\\[";
- $close = "\\]";
- $content = "(?>[^\\[\\]]+)"; // Anything but [ or ]
- // We can either find content or recursive brackets:
- $recursive_match = "($content|(?R))*";
- preg_match_all("/".$open.$recursive_match.$close."/Us", $body, $new_matches);
- // Loop through every match (link) we found
- if (is_array($new_matches)) {
- foreach ($new_matches as $nm) {
- // Might be an array of arrays
- if (is_array($nm)) {
- foreach ($nm as $n) {
- // If it's a file link get rid of it
- if (strtolower(substr($n, 0, 7)) == "[[file:"
- || strtolower(substr($n, 0, 8)) == "[[image:"
- ) {
- $body = str_replace($n, "", $body);
- }
- }
- } else {
- // Or just a normal array...
- // If it's a file link get rid of it
- if (strtolower(substr($nm, 0, 7)) == "[[file:"
- || strtolower(substr($nm, 0, 8)) == "[[image:"
- ) {
- $body = str_replace($nm, "", $body);
- }
- }
- }
- }
if (isset($imageUrl) && $imageUrl != false) {
$info['image'] = $imageUrl;
--
GitLab