aboutsummaryrefslogtreecommitdiffhomepage
path: root/src
diff options
context:
space:
mode:
authorSimounet <contact@simounet.net>2018-05-31 23:42:06 +0200
committerSimounet <contact@simounet.net>2018-06-01 13:49:16 +0200
commitc15bb5ad72b1a9692682cf88318a70b7e650d34a (patch)
tree778558ec77177019596ccd2c356d24b1b6074936 /src
parent9707ac46613eea0c536d87f4c141d567fc2ef7d0 (diff)
downloadwallabag-c15bb5ad72b1a9692682cf88318a70b7e650d34a.tar.gz
wallabag-c15bb5ad72b1a9692682cf88318a70b7e650d34a.tar.zst
wallabag-c15bb5ad72b1a9692682cf88318a70b7e650d34a.zip
Fix srcset attribute on images downloaded
Diffstat (limited to 'src')
-rw-r--r--src/Wallabag/CoreBundle/Helper/DownloadImages.php36
1 files changed, 33 insertions, 3 deletions
diff --git a/src/Wallabag/CoreBundle/Helper/DownloadImages.php b/src/Wallabag/CoreBundle/Helper/DownloadImages.php
index 252ba57c..9c9452dd 100644
--- a/src/Wallabag/CoreBundle/Helper/DownloadImages.php
+++ b/src/Wallabag/CoreBundle/Helper/DownloadImages.php
@@ -42,14 +42,17 @@ class DownloadImages
42 public function processHtml($entryId, $html, $url) 42 public function processHtml($entryId, $html, $url)
43 { 43 {
44 $crawler = new Crawler($html); 44 $crawler = new Crawler($html);
45 $result = $crawler 45 $imagesCrawler = $crawler
46 ->filterXpath('//img') 46 ->filterXpath('//img');
47 $imagesUrls = $imagesCrawler
47 ->extract(['src']); 48 ->extract(['src']);
49 $imagesSrcsetUrls = $this->getSrcsetUrls($imagesCrawler);
50 $imagesUrls = array_unique(array_merge($imagesUrls, $imagesSrcsetUrls));
48 51
49 $relativePath = $this->getRelativePath($entryId); 52 $relativePath = $this->getRelativePath($entryId);
50 53
51 // download and save the image to the folder 54 // download and save the image to the folder
52 foreach ($result as $image) { 55 foreach ($imagesUrls as $image) {
53 $imagePath = $this->processSingleImage($entryId, $image, $url, $relativePath); 56 $imagePath = $this->processSingleImage($entryId, $image, $url, $relativePath);
54 57
55 if (false === $imagePath) { 58 if (false === $imagePath) {
@@ -172,6 +175,33 @@ class DownloadImages
172 } 175 }
173 176
174 /** 177 /**
178 * Get images urls from the srcset image attribute.
179 *
180 * @param Crawler $imagesCrawler
181 *
182 * @return array An array of urls
183 */
184 protected function getSrcsetUrls(Crawler $imagesCrawler)
185 {
186 $urls = [];
187 $iterator = $imagesCrawler
188 ->getIterator();
189 while ($iterator->valid()) {
190 $srcsetAttribute = $iterator->current()->getAttribute('srcset');
191 if ('' !== $srcsetAttribute) {
192 $srcset = array_map('trim', explode(',', $srcsetAttribute));
193 $srcsetUrls = array_map(function ($src) {
194 return explode(' ', $src)[0];
195 }, $srcset);
196 $urls = array_merge($srcsetUrls, $urls);
197 }
198 $iterator->next();
199 }
200
201 return $urls;
202 }
203
204 /**
175 * Setup base folder where all images are going to be saved. 205 * Setup base folder where all images are going to be saved.
176 */ 206 */
177 private function setFolder() 207 private function setFolder()