Merge pull request #4844 from wallabag/feature/add-clean-downloaded-images

Add a command to clean downloaded images

phpstan.neon

@@ -7,8 +7,7 @@ parameters:
     symfony:
         container_xml_path: %rootDir%/../../../var/cache/test/appTestDebugProjectContainer.xml
 
-    # https://github.com/phpstan/phpstan/issues/694#issuecomment-350724288
+    bootstrapFiles:
-    autoload_files:
         - vendor/bin/.phpunit/phpunit-8.3-0/vendor/autoload.php
 
     inferPrivatePropertyTypeFromConstructor: true

src/Wallabag/CoreBundle/Command/CleanDownloadedImagesCommand.php

@@ -0,0 +1,101 @@
+<?php
+
+namespace Wallabag\CoreBundle\Command;
+
+use Symfony\Bundle\FrameworkBundle\Command\ContainerAwareCommand;
+use Symfony\Component\Console\Input\InputInterface;
+use Symfony\Component\Console\Input\InputOption;
+use Symfony\Component\Console\Output\OutputInterface;
+use Symfony\Component\Console\Style\SymfonyStyle;
+use Symfony\Component\Finder\Finder;
+
+class CleanDownloadedImagesCommand extends ContainerAwareCommand
+{
+    protected function configure()
+    {
+        $this
+            ->setName('wallabag:clean-downloaded-images')
+            ->setDescription('Cleans downloaded images which are no more associated to an entry')
+            ->addOption(
+               'dry-run',
+               null,
+               InputOption::VALUE_NONE,
+               'Do not remove images, just dump counters'
+            );
+    }
+
+    protected function execute(InputInterface $input, OutputInterface $output)
+    {
+        $io = new SymfonyStyle($input, $output);
+
+        $dryRun = (bool) $input->getOption('dry-run');
+
+        if ($dryRun) {
+            $io->text('Dry run mode <info>enabled</info> (no images will be removed)');
+        }
+
+        $downloadImages = $this->getContainer()->get('wallabag_core.entry.download_images');
+        $baseFolder = $downloadImages->getBaseFolder();
+
+        $io->text('Retrieve existing images');
+
+        // retrieve _existing_ folders in the image folder
+        $finder = new Finder();
+        $finder
+            ->directories()
+            ->ignoreDotFiles(true)
+            ->depth(2)
+            ->in($baseFolder);
+
+        $existingPaths = [];
+        foreach ($finder as $file) {
+            $existingPaths[] = $file->getFilename();
+        }
+
+        $io->text(sprintf('  -> <info>%d</info> images found', \count($existingPaths)));
+
+        $io->text('Retrieve valid folders attached to a user');
+
+        $entries = $this->getContainer()->get('wallabag_core.entry_repository')->findAllEntriesIdByUserId();
+
+        // retrieve _valid_ folders from existing entries
+        $validPaths = [];
+        foreach ($entries as $entry) {
+            $path = $downloadImages->getRelativePath($entry['id']);
+
+            if (!file_exists($baseFolder . '/' . $path)) {
+                continue;
+            }
+
+            // only store the hash, not the full path
+            $validPaths[] = explode('/', $path)[2];
+        }
+
+        $io->text(sprintf('  -> <info>%d</info> folders found', \count($validPaths)));
+
+        $deletedCount = 0;
+
+        $io->text('Remove images');
+
+        // check if existing path are valid, if not, remove all images and the folder
+        foreach ($existingPaths as $existingPath) {
+            if (!\in_array($existingPath, $validPaths, true)) {
+                $fullPath = $baseFolder . '/' . $existingPath[0] . '/' . $existingPath[1] . '/' . $existingPath;
+                $files = glob($fullPath . '/*.*');
+
+                if (!$dryRun) {
+                    array_map('unlink', $files);
+                    rmdir($fullPath);
+                }
+
+                $deletedCount += \count($files);
+
+                $io->text(sprintf('Deleted images in <info>%s</info>: <info>%d</info>', $existingPath, \count($files)));
+            }
+        }
+
+        $io->success(sprintf('Finished cleaning. %d deleted images', $deletedCount));
+
+        return 0;
+    }
+}

src/Wallabag/CoreBundle/Helper/DownloadImages.php

@@ -37,6 +37,11 @@ class DownloadImages
         $this->setFolder();
     }
 
+    public function getBaseFolder()
+    {
+        return $this->baseFolder;
+    }
+
     /**
      * Process the html and extract images URLs from it.
      *
@@ -99,7 +104,7 @@ class DownloadImages
      * @param string $url          Url from where the image were found
      * @param string $relativePath Relative local path to saved the image
      *
-     * @return string Relative url to access the image from the web
+     * @return string|false Relative url to access the image from the web
      */
     public function processSingleImage($entryId, $imagePath, $url, $relativePath = null)
     {
@@ -210,6 +215,29 @@ class DownloadImages
         @rmdir($folderPath);
     }
 
+    /**
+     * Generate the folder where we are going to save images based on the entry url.
+     *
+     * @param int  $entryId      ID of the entry
+     * @param bool $createFolder Should we create the folder for the given id?
+     *
+     * @return string
+     */
+    public function getRelativePath($entryId, $createFolder = true)
+    {
+        $hashId = hash('crc32', $entryId);
+        $relativePath = $hashId[0] . '/' . $hashId[1] . '/' . $hashId;
+        $folderPath = $this->baseFolder . '/' . $relativePath;
+
+        if (!file_exists($folderPath) && $createFolder) {
+            mkdir($folderPath, 0777, true);
+        }
+
+        $this->logger->debug('DownloadImages: Folder used for that Entry id', ['folder' => $folderPath, 'entryId' => $entryId]);
+
+        return $relativePath;
+    }
+
     /**
      * Get images urls from the srcset image attribute.
      *
@@ -254,28 +282,6 @@ class DownloadImages
         }
     }
 
-    /**
-     * Generate the folder where we are going to save images based on the entry url.
-     *
-     * @param int $entryId ID of the entry
-     *
-     * @return string
-     */
-    private function getRelativePath($entryId)
-    {
-        $hashId = hash('crc32', $entryId);
-        $relativePath = $hashId[0] . '/' . $hashId[1] . '/' . $hashId;
-        $folderPath = $this->baseFolder . '/' . $relativePath;
-
-        if (!file_exists($folderPath)) {
-            mkdir($folderPath, 0777, true);
-        }
-
-        $this->logger->debug('DownloadImages: Folder used for that Entry id', ['folder' => $folderPath, 'entryId' => $entryId]);
-
-        return $relativePath;
-    }
-
     /**
      * Make an $url absolute based on the $base.
      *