From 285f5c02a294654b013e15464239a1bb4d73cec6 Mon Sep 17 00:00:00 2001
From: Demian Katz <demian.katz@villanova.edu>
Date: Wed, 3 Sep 2014 09:25:17 -0400
Subject: [PATCH] Improved image extraction for some non-English languages.
---
.../src/VuFind/Connection/Wikipedia.php | 36 ++++++++++++++-----
1 file changed, 27 insertions(+), 9 deletions(-)
diff --git a/module/VuFind/src/VuFind/Connection/Wikipedia.php b/module/VuFind/src/VuFind/Connection/Wikipedia.php
index 8de59b6fe71..c00364f36ea 100644
--- a/module/VuFind/src/VuFind/Connection/Wikipedia.php
+++ b/module/VuFind/src/VuFind/Connection/Wikipedia.php
@@ -110,7 +110,7 @@ class Wikipedia implements TranslatorAwareInterface
*/
public function setLanguage($lang)
{
- $this->lang = $lang;
+ $this->lang = substr($lang, 0, 2); // strip off regional suffixes
}
/**
@@ -172,7 +172,8 @@ class Wikipedia implements TranslatorAwareInterface
$imageName = $imageCaption = null;
// Get rid of the last pair of braces and split
- $infobox = explode("\n|", substr($infoboxStr, 2, -2));
+ $infobox = explode("\n|", preg_replace('/^\s+|/m', '', substr($infoboxStr, 2, -2)));
+
// Look through every row of the infobox
foreach ($infobox as $row) {
$data = explode("=", $row);
@@ -185,11 +186,16 @@ class Wikipedia implements TranslatorAwareInterface
case "image":
case "image:":
case "image_name":
+ case "imagem":
+ case 'imagen':
+ case 'immagine':
$imageName = str_replace(' ', '_', $value);
break;
case "caption":
case "img_capt":
case "image_caption":
+ case "legenda":
+ case 'textoimagen':
$imageCaption = $value;
break;
default:
@@ -213,11 +219,17 @@ class Wikipedia implements TranslatorAwareInterface
// We are looking for the infobox inside "{{...}}"
// It may contain nested blocks too, thus the recursion
preg_match_all('/\{([^{}]++|(?R))*\}/s', $body['*'], $matches);
+
foreach ($matches[1] as $m) {
- // If this is the Infobox
- if (substr($m, 0, 8) == "{Infobox") {
- // Keep the string for later, we need the body block that follows it
- return "{".$m."}";
+ // Check if this is the Infobox; name may vary by language
+ $infoboxTags = array(
+ 'Bio', 'Ficha de escritor', 'Infobox', 'Info/Biografia'
+ );
+ foreach ($infoboxTags as $tag) {
+ if (substr($m, 0, strlen($tag) + 1) == '{' . $tag) {
+ // We found an infobox!!
+ return "{".$m."}";
+ }
}
}
@@ -234,10 +246,16 @@ class Wikipedia implements TranslatorAwareInterface
protected function extractImageFromBody($body)
{
$imageName = $imageCaption = null;
- $pattern = '/(\x5b\x5b)Image:([^\x5d]*)(\x5d\x5d)/U';
+ // The tag marking image files will vary depending on API language:
+ $tags = array(
+ 'Archivo', 'Bestand', 'Datei', 'Ficheiro', 'Fichier', 'File', 'Image'
+ );
+ $pattern = '/(\x5b\x5b)('
+ . implode('|', $tags)
+ . '):([^\x5d]*\.jpg[^\x5d]*)(\x5d\x5d)/U';
preg_match_all($pattern, $body['*'], $matches);
- if (isset($matches[2][0])) {
- $parts = explode('|', $matches[2][0]);
+ if (isset($matches[3][0])) {
+ $parts = explode('|', $matches[3][0]);
$imageName = str_replace(' ', '_', $parts[0]);
if (count($parts) > 1) {
$imageCaption = strip_tags(
--
GitLab