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>

5 files changed, 88 insertions, 4 deletions

diff --git a/src/Wallabag/CoreBundle/Entity/IgnoreOriginInstanceRule.php b/src/Wallabag/CoreBundle/Entity/IgnoreOriginInstanceRule.php
index 34aed50c..ce3b6e7d 100644
--- a/src/Wallabag/CoreBundle/Entity/IgnoreOriginInstanceRule.php
+++ b/src/Wallabag/CoreBundle/Entity/IgnoreOriginInstanceRule.php
@@ -11,7 +11,6 @@ use Symfony\Component\Validator\Constraints as Assert;
  *
  * @ORM\Entity(repositoryClass="Wallabag\CoreBundle\Repository\IgnoreOriginInstanceRuleRepository")
  * @ORM\Table(name="`ignore_origin_instance_rule`")
- * @ORM\Entity
  */
 class IgnoreOriginInstanceRule implements IgnoreOriginRuleInterface, RuleInterface
 {
@@ -30,7 +29,7 @@ class IgnoreOriginInstanceRule implements IgnoreOriginRuleInterface, RuleInterfa
      * @Assert\NotBlank()
      * @Assert\Length(max=255)
      * @RulerZAssert\ValidRule(
-     *  allowed_variables={"host","pattern"},
+     *  allowed_variables={"host","_all"},
      *  allowed_operators={"=","~"}
      * )
      * @ORM\Column(name="rule", type="string", nullable=false)
diff --git a/src/Wallabag/CoreBundle/Entity/IgnoreOriginUserRule.php b/src/Wallabag/CoreBundle/Entity/IgnoreOriginUserRule.php
index befd6090..0b6f318d 100644
--- a/src/Wallabag/CoreBundle/Entity/IgnoreOriginUserRule.php
+++ b/src/Wallabag/CoreBundle/Entity/IgnoreOriginUserRule.php
@@ -11,7 +11,6 @@ use Symfony\Component\Validator\Constraints as Assert;
  *
  * @ORM\Entity(repositoryClass="Wallabag\CoreBundle\Repository\IgnoreOriginUserRuleRepository")
  * @ORM\Table(name="`ignore_origin_user_rule`")
- * @ORM\Entity
  */
 class IgnoreOriginUserRule implements IgnoreOriginRuleInterface, RuleInterface
 {
@@ -30,7 +29,7 @@ class IgnoreOriginUserRule implements IgnoreOriginRuleInterface, RuleInterface
      * @Assert\NotBlank()
      * @Assert\Length(max=255)
      * @RulerZAssert\ValidRule(
-     *  allowed_variables={"host","pattern"},
+     *  allowed_variables={"host","_all"},
      *  allowed_operators={"=","~"}
      * )
      * @ORM\Column(name="rule", type="string", nullable=false)
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' => $rule->getRule(),
+                ]);
+
+                return true;
+            }
+        }
+
+        return false;
+    }
+}
diff --git a/src/Wallabag/CoreBundle/Operator/PHP/PatternMatches.php b/src/Wallabag/CoreBundle/Operator/PHP/PatternMatches.php
new file mode 100644
index 00000000..532e2bb3
--- /dev/null
+++ b/src/Wallabag/CoreBundle/Operator/PHP/PatternMatches.php
@@ -0,0 +1,23 @@
+<?php
+
+namespace Wallabag\CoreBundle\Operator\PHP;
+
+/**
+ * Provides a "~" operator used for ignore origin rules.
+ *
+ * It asserts that a subject matches a given regexp pattern, in a
+ * case-insensitive way.
+ *
+ * This operator will be used to compile ignore origin rules in PHP, usable
+ * directly on Entry objects for instance.
+ * It's registered in RulerZ using a service (wallabag.operator.array.pattern_matches);
+ */
+class PatternMatches
+{
+    public function __invoke($subject, $pattern)
+    {
+        $count = preg_match("`$pattern`i", $subject);
+
+        return \is_int($count) && $count > 0;
+    }
+}
diff --git a/src/Wallabag/CoreBundle/Resources/config/services.yml b/src/Wallabag/CoreBundle/Resources/config/services.yml
index 4ece046a..8417ac35 100644
--- a/src/Wallabag/CoreBundle/Resources/config/services.yml
+++ b/src/Wallabag/CoreBundle/Resources/config/services.yml
@@ -92,6 +92,7 @@ services:
         arguments:
             - "@wallabag_core.graby"
             - "@wallabag_core.rule_based_tagger"
+            - "@wallabag_core.rule_based_ignore_origin_processor"
             - "@validator"
             - "@logger"
             - '%wallabag_core.fetching_error_message%'
@@ -110,6 +111,13 @@ services:
             - "@wallabag_core.entry_repository"
             - "@logger"
 
+    wallabag_core.rule_based_ignore_origin_processor:
+        class: Wallabag\CoreBundle\Helper\RuleBasedIgnoreOriginProcessor
+        arguments:
+            - "@rulerz"
+            - "@logger"
+            - "@wallabag_core.ignore_origin_instance_rule_repository"
+
     # repository as a service
     wallabag_core.entry_repository:
         class: Wallabag\CoreBundle\Repository\EntryRepository
@@ -164,6 +172,11 @@ services:
         tags:
             - { name: rulerz.operator, target: doctrine, operator: notmatches, inline: true }
 
+    wallabag.operator.array.pattern_matches:
+        class: Wallabag\CoreBundle\Operator\PHP\PatternMatches
+        tags:
+            - { name: rulerz.operator, target: native, operator: "~" }
+
     wallabag_core.helper.redirect:
         class: Wallabag\CoreBundle\Helper\Redirect
         arguments: