From f39c5a2a702036750b4d7c32d02e7f92955a4eed Mon Sep 17 00:00:00 2001
From: Kevin Decherf <kevin@kdecherf.com>
Date: Sun, 11 Aug 2019 23:51:55 +0200
Subject: Add new Helper to process Ignore Origin rules and RulerZ operator

This commits adds a new helper like RuleBasedTagger for processing
ignore origin rules. It also adds a new custom RulerZ operator for the
'~' pattern matching rule.

Renames 'pattern' with '_all' in IgnoreOriginRule entity.

Signed-off-by: Kevin Decherf <kevin@kdecherf.com>
---
 .../Helper/RuleBasedIgnoreOriginProcessor.php      | 50 ++++++++++++++++++++++
 1 file changed, 50 insertions(+)
 create mode 100644 src/Wallabag/CoreBundle/Helper/RuleBasedIgnoreOriginProcessor.php

(limited to 'src/Wallabag/CoreBundle/Helper')

diff --git a/src/Wallabag/CoreBundle/Helper/RuleBasedIgnoreOriginProcessor.php b/src/Wallabag/CoreBundle/Helper/RuleBasedIgnoreOriginProcessor.php
new file mode 100644
index 00000000..333e5b0a
--- /dev/null
+++ b/src/Wallabag/CoreBundle/Helper/RuleBasedIgnoreOriginProcessor.php
@@ -0,0 +1,50 @@
+<?php
+
+namespace Wallabag\CoreBundle\Helper;
+
+use Psr\Log\LoggerInterface;
+use RulerZ\RulerZ;
+use Wallabag\CoreBundle\Entity\Entry;
+use Wallabag\CoreBundle\Repository\IgnoreOriginInstanceRuleRepository;
+
+class RuleBasedIgnoreOriginProcessor
+{
+    protected $rulerz;
+    protected $logger;
+    protected $ignoreOriginInstanceRuleRepository;
+
+    public function __construct(RulerZ $rulerz, LoggerInterface $logger, IgnoreOriginInstanceRuleRepository $ignoreOriginInstanceRuleRepository)
+    {
+        $this->rulerz = $rulerz;
+        $this->logger = $logger;
+        $this->ignoreOriginInstanceRuleRepository = $ignoreOriginInstanceRuleRepository;
+    }
+
+    /**
+     * @param Entry $entry Entry to process
+     *
+     * @return bool
+     */
+    public function process(Entry $entry)
+    {
+        $url = $entry->getUrl();
+        $userRules = $entry->getUser()->getConfig()->getIgnoreOriginRules()->toArray();
+        $rules = array_merge($this->ignoreOriginInstanceRuleRepository->findAll(), $userRules);
+
+        $parsed_url = parse_url($url);
+        // We add the former url as a new key _all for pattern matching
+        $parsed_url['_all'] = $url;
+
+        foreach ($rules as $rule) {
+            if ($this->rulerz->satisfies($parsed_url, $rule->getRule())) {
+                $this->logger->info('Origin url matching ignore rule.', [
+                    'rule' => $rule->getRule(),
+                ]);
+
+                return true;
+            }
+        }
+
+        return false;
+    }
+}
-- 
cgit v1.2.3


From b22eb276232b5c15a6fbadc9dd10144e709faec3 Mon Sep 17 00:00:00 2001
From: Kevin Decherf <kevin@kdecherf.com>
Date: Sun, 11 Aug 2019 23:55:52 +0200
Subject: ContentProxy: replace ignoreUrl with new
 RuleBasedIgnoreOriginProcessor

Signed-off-by: Kevin Decherf <kevin@kdecherf.com>
---
 src/Wallabag/CoreBundle/Helper/ContentProxy.php | 41 +++----------------------
 1 file changed, 4 insertions(+), 37 deletions(-)

(limited to 'src/Wallabag/CoreBundle/Helper')

diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php
index 9c6fa8db..7e93249d 100644
--- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php
+++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php
@@ -19,6 +19,7 @@ class ContentProxy
 {
     protected $graby;
     protected $tagger;
+    protected $ignoreOriginProcessor;
     protected $validator;
     protected $logger;
     protected $mimeGuesser;
@@ -26,10 +27,11 @@ class ContentProxy
     protected $eventDispatcher;
     protected $storeArticleHeaders;
 
-    public function __construct(Graby $graby, RuleBasedTagger $tagger, ValidatorInterface $validator, LoggerInterface $logger, $fetchingErrorMessage, $storeArticleHeaders = false)
+    public function __construct(Graby $graby, RuleBasedTagger $tagger, RuleBasedIgnoreOriginProcessor $ignoreOriginProcessor, ValidatorInterface $validator, LoggerInterface $logger, $fetchingErrorMessage, $storeArticleHeaders = false)
     {
         $this->graby = $graby;
         $this->tagger = $tagger;
+        $this->ignoreOriginProcessor = $ignoreOriginProcessor;
         $this->validator = $validator;
         $this->logger = $logger;
         $this->mimeGuesser = new MimeTypeExtensionGuesser();
@@ -356,7 +358,7 @@ class ContentProxy
         $diff_keys = array_keys($diff);
         sort($diff_keys);
 
-        if ($this->ignoreUrl($entry->getUrl())) {
+        if ($this->ignoreOriginProcessor->process($entry)) {
             $entry->setUrl($url);
 
             return false;
@@ -395,41 +397,6 @@ class ContentProxy
         }
     }
 
-    /**
-     * Check entry url against an ignore list to replace with content url.
-     *
-     * XXX: move the ignore list in the database to let users handle it
-     *
-     * @param string $url url to test
-     *
-     * @return bool true if url matches ignore list otherwise false
-     */
-    private function ignoreUrl($url)
-    {
-        $ignored_hosts = ['feedproxy.google.com', 'feeds.reuters.com'];
-        $ignored_patterns = ['https?://www\.lemonde\.fr/tiny.*'];
-
-        $parsed_url = parse_url($url);
-
-        $filtered = array_filter($ignored_hosts, function ($var) use ($parsed_url) {
-            return $var === $parsed_url['host'];
-        });
-
-        if ([] !== $filtered) {
-            return true;
-        }
-
-        $filtered = array_filter($ignored_patterns, function ($var) use ($url) {
-            return preg_match("`$var`i", $url);
-        });
-
-        if ([] !== $filtered) {
-            return true;
-        }
-
-        return false;
-    }
-
     /**
      * Validate that the given content has at least a title, an html and a url.
      *
-- 
cgit v1.2.3