From 1a552bd60eab4a4b940d3896376b599e155d7da0 Mon Sep 17 00:00:00 2001
From: Alexandre Alapetite <alexandre@alapetite.fr>
Date: Fri, 6 Sep 2024 09:35:58 +0200
Subject: Regex search (#6706)

* Regex search
fix https://github.com/FreshRSS/FreshRSS/issues/3549

* Fix PHPStan

* Fix escape

* Fix ungreedy

* Initial support for regex search in PostgreSQL and MySQL

* Improvements, support MySQL

* Fix multiline

* Add support for SQLite

* A few tests

* Added author: and inurl: support, documentation

* author example

* Remove \b for now

* Disable regex sanitization for now

* Fix getInurlRegex

* getNotInurlRegex

* Quotes for inurl:

* Fix test

* Fix quoted tags + regex for tags
https://github.com/FreshRSS/FreshRSS/issues/6761

* Fix wrong regex detection

* Add MariaDB

* Fix logic

* Increase requirements for MySQL and MariaDB
Check support for multiline mode in MySQL

* Remove sanitizeRegexes()

* Allow searching HTML code
Allow searching for instance `/<pre>/`
Fix https://github.com/FreshRSS/FreshRSS/issues/6775#issuecomment-2331769883

* Doc regex search HTML

* Fix Doctype
---
 app/Models/DatabaseDAO.php    |  27 +++++-
 app/Models/Entry.php          |  66 ++++++++++++++
 app/Models/EntryDAO.php       | 116 ++++++++++++++++++++++-
 app/Models/EntryDAOPGSQL.php  |  26 ++++++
 app/Models/EntryDAOSQLite.php |  21 +++++
 app/Models/Search.php         | 207 +++++++++++++++++++++++++++++++++++-------
 6 files changed, 425 insertions(+), 38 deletions(-)

(limited to 'app')

diff --git a/app/Models/DatabaseDAO.php b/app/Models/DatabaseDAO.php
index 363225cdb..ba0ee3e79 100644
--- a/app/Models/DatabaseDAO.php
+++ b/app/Models/DatabaseDAO.php
@@ -185,6 +185,30 @@ class FreshRSS_DatabaseDAO extends Minz_ModelPdo {
 		return $list;
 	}
 
+	private static ?string $staticVersion = null;
+	/**
+	 * To override the database version. Useful for testing.
+	 */
+	public static function setStaticVersion(?string $version): void {
+		self::$staticVersion = $version;
+	}
+
+	public function version(): string {
+		if (self::$staticVersion !== null) {
+			return self::$staticVersion;
+		}
+		static $version = null;
+		if ($version === null) {
+			$version = $this->fetchValue('SELECT version()') ?? '';
+		}
+		return $version;
+	}
+
+	final public function isMariaDB(): bool {
+		// MariaDB includes its name in version, but not MySQL
+		return str_contains($this->version(), 'MariaDB');
+	}
+
 	public function size(bool $all = false): int {
 		$db = FreshRSS_Context::systemConf()->db;
 
@@ -237,8 +261,7 @@ SQL;
 			$isMariaDB = false;
 
 			if ($this->pdo->dbType() === 'mysql') {
-				$dbVersion = $this->fetchValue('SELECT version()') ?? '';
-				$isMariaDB = stripos($dbVersion, 'MariaDB') !== false;	// MariaDB includes its name in version, but not MySQL
+				$isMariaDB = $this->isMariaDB();
 				if (!$isMariaDB) {
 					// MySQL does not support `DROP INDEX IF EXISTS` yet https://dev.mysql.com/doc/refman/8.3/en/drop-index.html
 					// but MariaDB does https://mariadb.com/kb/en/drop-index/
diff --git a/app/Models/Entry.php b/app/Models/Entry.php
index 4b331419b..415bc0235 100644
--- a/app/Models/Entry.php
+++ b/app/Models/Entry.php
@@ -631,27 +631,60 @@ HTML;
 						$ok &= stripos(implode(';', $this->authors), $author) !== false;
 					}
 				}
+				if ($ok && $filter->getAuthorRegex()) {
+					foreach ($filter->getAuthorRegex() as $author) {
+						$ok &= preg_match($author, implode("\n", $this->authors)) === 1;
+					}
+				}
 				if ($ok && $filter->getNotAuthor()) {
 					foreach ($filter->getNotAuthor() as $author) {
 						$ok &= stripos(implode(';', $this->authors), $author) === false;
 					}
 				}
+				if ($ok && $filter->getNotAuthorRegex()) {
+					foreach ($filter->getNotAuthorRegex() as $author) {
+						$ok &= preg_match($author, implode("\n", $this->authors)) === 0;
+					}
+				}
 				if ($ok && $filter->getIntitle()) {
 					foreach ($filter->getIntitle() as $title) {
 						$ok &= stripos($this->title, $title) !== false;
 					}
 				}
+				if ($ok && $filter->getIntitleRegex()) {
+					foreach ($filter->getIntitleRegex() as $title) {
+						$ok &= preg_match($title, $this->title) === 1;
+					}
+				}
 				if ($ok && $filter->getNotIntitle()) {
 					foreach ($filter->getNotIntitle() as $title) {
 						$ok &= stripos($this->title, $title) === false;
 					}
 				}
+				if ($ok && $filter->getNotIntitleRegex()) {
+					foreach ($filter->getNotIntitleRegex() as $title) {
+						$ok &= preg_match($title, $this->title) === 0;
+					}
+				}
 				if ($ok && $filter->getTags()) {
 					foreach ($filter->getTags() as $tag2) {
 						$found = false;
 						foreach ($this->tags as $tag1) {
 							if (strcasecmp($tag1, $tag2) === 0) {
 								$found = true;
+								break;
+							}
+						}
+						$ok &= $found;
+					}
+				}
+				if ($ok && $filter->getTagsRegex()) {
+					foreach ($filter->getTagsRegex() as $tag2) {
+						$found = false;
+						foreach ($this->tags as $tag1) {
+							if (preg_match($tag2, $tag1) === 1) {
+								$found = true;
+								break;
 							}
 						}
 						$ok &= $found;
@@ -663,6 +696,19 @@ HTML;
 						foreach ($this->tags as $tag1) {
 							if (strcasecmp($tag1, $tag2) === 0) {
 								$found = true;
+								break;
+							}
+						}
+						$ok &= !$found;
+					}
+				}
+				if ($ok && $filter->getNotTagsRegex()) {
+					foreach ($filter->getNotTagsRegex() as $tag2) {
+						$found = false;
+						foreach ($this->tags as $tag1) {
+							if (preg_match($tag2, $tag1) === 1) {
+								$found = true;
+								break;
 							}
 						}
 						$ok &= !$found;
@@ -673,11 +719,21 @@ HTML;
 						$ok &= stripos($this->link, $url) !== false;
 					}
 				}
+				if ($ok && $filter->getInurlRegex()) {
+					foreach ($filter->getInurlRegex() as $url) {
+						$ok &= preg_match($url, $this->link) === 1;
+					}
+				}
 				if ($ok && $filter->getNotInurl()) {
 					foreach ($filter->getNotInurl() as $url) {
 						$ok &= stripos($this->link, $url) === false;
 					}
 				}
+				if ($ok && $filter->getNotInurlRegex()) {
+					foreach ($filter->getNotInurlRegex() as $url) {
+						$ok &= preg_match($url, $this->link) === 0;
+					}
+				}
 				if ($ok && $filter->getSearch()) {
 					foreach ($filter->getSearch() as $needle) {
 						$ok &= (stripos($this->title, $needle) !== false || stripos($this->content, $needle) !== false);
@@ -688,6 +744,16 @@ HTML;
 						$ok &= (stripos($this->title, $needle) === false && stripos($this->content, $needle) === false);
 					}
 				}
+				if ($ok && $filter->getSearchRegex()) {
+					foreach ($filter->getSearchRegex() as $needle) {
+						$ok &= (preg_match($needle, $this->title) === 1 || preg_match($needle, $this->content) === 1);
+					}
+				}
+				if ($ok && $filter->getNotSearchRegex()) {
+					foreach ($filter->getNotSearchRegex() as $needle) {
+						$ok &= (preg_match($needle, $this->title) === 0 && preg_match($needle, $this->content) === 0);
+					}
+				}
 				if ($ok) {
 					return true;
 				}
diff --git a/app/Models/EntryDAO.php b/app/Models/EntryDAO.php
index 175df15c3..6a00e2108 100644
--- a/app/Models/EntryDAO.php
+++ b/app/Models/EntryDAO.php
@@ -27,6 +27,55 @@ class FreshRSS_EntryDAO extends Minz_ModelPdo {
 		return str_replace('INSERT INTO ', 'INSERT IGNORE INTO ', $sql);
 	}
 
+	/** @return array{pattern?:string,matchType?:string} */
+	protected static function regexToSql(string $regex): array {
+		if (preg_match('#^/(?P<pattern>.*)/(?P<matchType>[im]*)$#', $regex, $matches)) {
+			return $matches;
+		}
+		return [];
+	}
+
+	/** @param array<int|string> $values */
+	protected static function sqlRegex(string $expression, string $regex, array &$values): string {
+		// The implementation of this function is solely for MySQL and MariaDB
+		static $databaseDAOMySQL = null;
+		if ($databaseDAOMySQL === null) {
+			$databaseDAOMySQL = new FreshRSS_DatabaseDAO();
+		}
+
+		$matches = static::regexToSql($regex);
+		if (isset($matches['pattern'])) {
+			$matchType = $matches['matchType'] ?? '';
+			if ($databaseDAOMySQL->isMariaDB()) {
+				if (str_contains($matchType, 'm')) {
+					// multiline mode
+					$matches['pattern'] = '(?m)' . $matches['pattern'];
+				}
+				if (str_contains($matchType, 'i')) {
+					// case-insensitive match
+					$matches['pattern'] = '(?i)' . $matches['pattern'];
+				} else {
+					$matches['pattern'] = '(?-i)' . $matches['pattern'];
+				}
+				$values[] = $matches['pattern'];
+				return "{$expression} REGEXP ?";
+			} else {	// MySQL
+				if (!str_contains($matchType, 'i')) {
+					// Case-sensitive matching
+					$matchType .= 'c';
+				}
+				$values[] = $matches['pattern'];
+				return "REGEXP_LIKE({$expression},?,'{$matchType}')";
+			}
+		}
+		return '';
+	}
+
+	/** Register any needed SQL function for the query, e.g. application-defined functions for SQLite */
+	protected function registerSqlFunctions(string $sql): void {
+		// Nothing to do for MySQL
+	}
+
 	private function updateToMediumBlob(): bool {
 		if ($this->pdo->dbType() !== 'mysql') {
 			return false;
@@ -910,24 +959,44 @@ SQL;
 					$values[] = "%{$author}%";
 				}
 			}
+			if ($filter->getAuthorRegex() !== null) {
+				foreach ($filter->getAuthorRegex() as $author) {
+					$sub_search .= 'AND ' . static::sqlRegex("REPLACE({$alias}author, ';', '\n')", $author, $values) . ' ';
+				}
+			}
 			if ($filter->getIntitle() !== null) {
 				foreach ($filter->getIntitle() as $title) {
 					$sub_search .= 'AND ' . $alias . 'title LIKE ? ';
 					$values[] = "%{$title}%";
 				}
 			}
+			if ($filter->getIntitleRegex() !== null) {
+				foreach ($filter->getIntitleRegex() as $title) {
+					$sub_search .= 'AND ' . static::sqlRegex($alias . 'title', $title, $values) . ' ';
+				}
+			}
 			if ($filter->getTags() !== null) {
 				foreach ($filter->getTags() as $tag) {
 					$sub_search .= 'AND ' . static::sqlConcat('TRIM(' . $alias . 'tags) ', " ' #'") . ' LIKE ? ';
 					$values[] = "%{$tag} #%";
 				}
 			}
+			if ($filter->getTagsRegex() !== null) {
+				foreach ($filter->getTagsRegex() as $tag) {
+					$sub_search .= 'AND ' . static::sqlRegex("REPLACE(REPLACE({$alias}tags, ' #', '#'), '#', '\n')", $tag, $values) . ' ';
+				}
+			}
 			if ($filter->getInurl() !== null) {
 				foreach ($filter->getInurl() as $url) {
 					$sub_search .= 'AND ' . $alias . 'link LIKE ? ';
 					$values[] = "%{$url}%";
 				}
 			}
+			if ($filter->getInurlRegex() !== null) {
+				foreach ($filter->getInurlRegex() as $url) {
+					$sub_search .= 'AND ' . static::sqlRegex($alias . 'link', $url, $values) . ' ';
+				}
+			}
 
 			if ($filter->getNotAuthor() !== null) {
 				foreach ($filter->getNotAuthor() as $author) {
@@ -935,29 +1004,49 @@ SQL;
 					$values[] = "%{$author}%";
 				}
 			}
+			if ($filter->getNotAuthorRegex() !== null) {
+				foreach ($filter->getNotAuthorRegex() as $author) {
+					$sub_search .= 'AND NOT ' . static::sqlRegex("REPLACE({$alias}author, ';', '\n')", $author, $values) . ' ';
+				}
+			}
 			if ($filter->getNotIntitle() !== null) {
 				foreach ($filter->getNotIntitle() as $title) {
 					$sub_search .= 'AND ' . $alias . 'title NOT LIKE ? ';
 					$values[] = "%{$title}%";
 				}
 			}
+			if ($filter->getNotIntitleRegex() !== null) {
+				foreach ($filter->getNotIntitleRegex() as $title) {
+					$sub_search .= 'AND NOT ' . static::sqlRegex($alias . 'title', $title, $values) . ' ';
+				}
+			}
 			if ($filter->getNotTags() !== null) {
 				foreach ($filter->getNotTags() as $tag) {
 					$sub_search .= 'AND ' . static::sqlConcat('TRIM(' . $alias . 'tags) ', " ' #'") . ' NOT LIKE ? ';
 					$values[] = "%{$tag} #%";
 				}
 			}
+			if ($filter->getNotTagsRegex() !== null) {
+				foreach ($filter->getNotTagsRegex() as $tag) {
+					$sub_search .= 'AND NOT ' . static::sqlRegex("REPLACE(REPLACE({$alias}tags, ' #', '#'), '#', '\n')", $tag, $values) . ' ';
+				}
+			}
 			if ($filter->getNotInurl() !== null) {
 				foreach ($filter->getNotInurl() as $url) {
 					$sub_search .= 'AND ' . $alias . 'link NOT LIKE ? ';
 					$values[] = "%{$url}%";
 				}
 			}
+			if ($filter->getNotInurlRegex() !== null) {
+				foreach ($filter->getNotInurlRegex() as $url) {
+					$sub_search .= 'AND NOT ' . static::sqlRegex($alias . 'link', $url, $values) . ' ';
+				}
+			}
 
 			if ($filter->getSearch() !== null) {
 				foreach ($filter->getSearch() as $search_value) {
 					if (static::isCompressed()) {	// MySQL-only
-						$sub_search .= 'AND CONCAT(' . $alias . 'title, UNCOMPRESS(' . $alias . 'content_bin)) LIKE ? ';
+						$sub_search .= "AND CONCAT({$alias}title, '\\n', UNCOMPRESS({$alias}content_bin)) LIKE ? ";
 						$values[] = "%{$search_value}%";
 					} else {
 						$sub_search .= 'AND (' . $alias . 'title LIKE ? OR ' . $alias . 'content LIKE ?) ';
@@ -966,10 +1055,21 @@ SQL;
 					}
 				}
 			}
+			if ($filter->getSearchRegex() !== null) {
+				foreach ($filter->getSearchRegex() as $search_value) {
+					if (static::isCompressed()) {	// MySQL-only
+						$sub_search .= 'AND (' . static::sqlRegex($alias . 'title', $search_value, $values) .
+							' OR ' . static::sqlRegex("UNCOMPRESS({$alias}content_bin)", $search_value, $values) . ') ';
+					} else {
+						$sub_search .= 'AND (' . static::sqlRegex($alias . 'title', $search_value, $values) .
+							' OR ' . static::sqlRegex($alias . 'content', $search_value, $values) . ') ';
+					}
+				}
+			}
 			if ($filter->getNotSearch() !== null) {
 				foreach ($filter->getNotSearch() as $search_value) {
 					if (static::isCompressed()) {	// MySQL-only
-						$sub_search .= 'AND CONCAT(' . $alias . 'title, UNCOMPRESS(' . $alias . 'content_bin)) NOT LIKE ? ';
+						$sub_search .= "AND CONCAT({$alias}title, '\\n', UNCOMPRESS({$alias}content_bin)) NOT LIKE ? ";
 						$values[] = "%{$search_value}%";
 					} else {
 						$sub_search .= 'AND ' . $alias . 'title NOT LIKE ? AND ' . $alias . 'content NOT LIKE ? ';
@@ -978,6 +1078,17 @@ SQL;
 					}
 				}
 			}
+			if ($filter->getNotSearchRegex() !== null) {
+				foreach ($filter->getNotSearchRegex() as $search_value) {
+					if (static::isCompressed()) {	// MySQL-only
+						$sub_search .= 'AND NOT ' . static::sqlRegex($alias . 'title', $search_value, $values) .
+							' ANT NOT ' . static::sqlRegex("UNCOMPRESS({$alias}content_bin)", $search_value, $values) . ' ';
+					} else {
+						$sub_search .= 'AND NOT ' . static::sqlRegex($alias . 'title', $search_value, $values) .
+							' AND NOT ' . static::sqlRegex($alias . 'content', $search_value, $values) . ' ';
+					}
+				}
+			}
 
 			if ($sub_search != '') {
 				if ($isOpen) {
@@ -1039,6 +1150,7 @@ SQL;
 			if ($filterSearch !== '') {
 				$search .= 'AND (' . $filterSearch . ') ';
 				$values = array_merge($values, $filterValues);
+				$this->registerSqlFunctions($search);
 			}
 		}
 		return [$values, $search];
diff --git a/app/Models/EntryDAOPGSQL.php b/app/Models/EntryDAOPGSQL.php
index 8adeffe9e..fe157308c 100644
--- a/app/Models/EntryDAOPGSQL.php
+++ b/app/Models/EntryDAOPGSQL.php
@@ -23,6 +23,32 @@ class FreshRSS_EntryDAOPGSQL extends FreshRSS_EntryDAOSQLite {
 		return rtrim($sql, ' ;') . ' ON CONFLICT DO NOTHING';
 	}
 
+	#[\Override]
+	protected static function sqlRegex(string $expression, string $regex, array &$values): string {
+		$matches = static::regexToSql($regex);
+		if (isset($matches['pattern'])) {
+			$matchType = $matches['matchType'] ?? '';
+			if (str_contains($matchType, 'm')) {
+				// newline-sensitive matching
+				$matches['pattern'] = '(?m)' . $matches['pattern'];
+			}
+			$values[] = $matches['pattern'];
+			if (str_contains($matchType, 'i')) {
+				// case-insensitive matching
+				return "{$expression} ~* ?";
+			} else {
+				// case-sensitive matching
+				return "{$expression} ~ ?";
+			}
+		}
+		return '';
+	}
+
+	#[\Override]
+	protected function registerSqlFunctions(string $sql): void {
+		// Nothing to do for PostgreSQL
+	}
+
 	/** @param array<string|int> $errorInfo */
 	#[\Override]
 	protected function autoUpdateDb(array $errorInfo): bool {
diff --git a/app/Models/EntryDAOSQLite.php b/app/Models/EntryDAOSQLite.php
index 9c2b37623..6d604f25a 100644
--- a/app/Models/EntryDAOSQLite.php
+++ b/app/Models/EntryDAOSQLite.php
@@ -28,6 +28,27 @@ class FreshRSS_EntryDAOSQLite extends FreshRSS_EntryDAO {
 		return str_replace('INSERT INTO ', 'INSERT OR IGNORE INTO ', $sql);
 	}
 
+	#[\Override]
+	protected static function sqlRegex(string $expression, string $regex, array &$values): string {
+		$values[] = $regex;
+		return "{$expression} REGEXP ?";
+	}
+
+	#[\Override]
+	protected function registerSqlFunctions(string $sql): void {
+		if (!str_contains($sql, ' REGEXP ')) {
+			return;
+		}
+		// https://php.net/pdo.sqlitecreatefunction
+		// https://www.sqlite.org/lang_expr.html#the_like_glob_regexp_match_and_extract_operators
+		$this->pdo->sqliteCreateFunction('regexp',
+			function (string $pattern, string $text): bool {
+				return preg_match($pattern, $text) === 1;
+			},
+			2
+		);
+	}
+
 	/** @param array<string|int> $errorInfo */
 	#[\Override]
 	protected function autoUpdateDb(array $errorInfo): bool {
diff --git a/app/Models/Search.php b/app/Models/Search.php
index 7eaf741c3..755cf6b59 100644
--- a/app/Models/Search.php
+++ b/app/Models/Search.php
@@ -27,6 +27,8 @@ class FreshRSS_Search {
 	private ?array $label_names = null;
 	/** @var array<string>|null */
 	private ?array $intitle = null;
+	/** @var array<string>|null */
+	private ?array $intitle_regex = null;
 	/** @var int|false|null */
 	private $min_date = null;
 	/** @var int|false|null */
@@ -38,11 +40,19 @@ class FreshRSS_Search {
 	/** @var array<string>|null */
 	private ?array $inurl = null;
 	/** @var array<string>|null */
+	private ?array $inurl_regex = null;
+	/** @var array<string>|null */
 	private ?array $author = null;
 	/** @var array<string>|null */
+	private ?array $author_regex = null;
+	/** @var array<string>|null */
 	private ?array $tags = null;
 	/** @var array<string>|null */
+	private ?array $tags_regex = null;
+	/** @var array<string>|null */
 	private ?array $search = null;
+	/** @var array<string>|null */
+	private ?array $search_regex = null;
 
 	/** @var array<string>|null */
 	private ?array $not_entry_ids = null;
@@ -54,6 +64,8 @@ class FreshRSS_Search {
 	private ?array $not_label_names = null;
 	/** @var array<string>|null */
 	private ?array $not_intitle = null;
+	/** @var array<string>|null */
+	private ?array $not_intitle_regex = null;
 	/** @var int|false|null */
 	private $not_min_date = null;
 	/** @var int|false|null */
@@ -65,11 +77,19 @@ class FreshRSS_Search {
 	/** @var array<string>|null */
 	private ?array $not_inurl = null;
 	/** @var array<string>|null */
+	private ?array $not_inurl_regex = null;
+	/** @var array<string>|null */
 	private ?array $not_author = null;
 	/** @var array<string>|null */
+	private ?array $not_author_regex = null;
+	/** @var array<string>|null */
 	private ?array $not_tags = null;
 	/** @var array<string>|null */
+	private ?array $not_tags_regex = null;
+	/** @var array<string>|null */
 	private ?array $not_search = null;
+	/** @var array<string>|null */
+	private ?array $not_search_regex = null;
 
 	public function __construct(string $input) {
 		$input = self::cleanSearch($input);
@@ -156,9 +176,17 @@ class FreshRSS_Search {
 		return $this->intitle;
 	}
 	/** @return array<string>|null */
+	public function getIntitleRegex(): ?array {
+		return $this->intitle_regex;
+	}
+	/** @return array<string>|null */
 	public function getNotIntitle(): ?array {
 		return $this->not_intitle;
 	}
+	/** @return array<string>|null */
+	public function getNotIntitleRegex(): ?array {
+		return $this->not_intitle_regex;
+	}
 
 	public function getMinDate(): ?int {
 		return $this->min_date ?: null;
@@ -199,36 +227,68 @@ class FreshRSS_Search {
 		return $this->inurl;
 	}
 	/** @return array<string>|null */
+	public function getInurlRegex(): ?array {
+		return $this->inurl_regex;
+	}
+	/** @return array<string>|null */
 	public function getNotInurl(): ?array {
 		return $this->not_inurl;
 	}
+	/** @return array<string>|null */
+	public function getNotInurlRegex(): ?array {
+		return $this->not_inurl_regex;
+	}
 
 	/** @return array<string>|null */
 	public function getAuthor(): ?array {
 		return $this->author;
 	}
 	/** @return array<string>|null */
+	public function getAuthorRegex(): ?array {
+		return $this->author_regex;
+	}
+	/** @return array<string>|null */
 	public function getNotAuthor(): ?array {
 		return $this->not_author;
 	}
+	/** @return array<string>|null */
+	public function getNotAuthorRegex(): ?array {
+		return $this->not_author_regex;
+	}
 
 	/** @return array<string>|null */
 	public function getTags(): ?array {
 		return $this->tags;
 	}
 	/** @return array<string>|null */
+	public function getTagsRegex(): ?array {
+		return $this->tags_regex;
+	}
+	/** @return array<string>|null */
 	public function getNotTags(): ?array {
 		return $this->not_tags;
 	}
+	/** @return array<string>|null */
+	public function getNotTagsRegex(): ?array {
+		return $this->not_tags_regex;
+	}
 
 	/** @return array<string>|null */
 	public function getSearch(): ?array {
 		return $this->search;
 	}
 	/** @return array<string>|null */
+	public function getSearchRegex(): ?array {
+		return $this->search_regex;
+	}
+	/** @return array<string>|null */
 	public function getNotSearch(): ?array {
 		return $this->not_search;
 	}
+	/** @return array<string>|null */
+	public function getNotSearchRegex(): ?array {
+		return $this->not_search_regex;
+	}
 
 	/**
 	 * @param array<string>|null $anArray
@@ -253,11 +313,19 @@ class FreshRSS_Search {
 		return $value;
 	}
 
+	/**
+	 * @param array<string> $strings
+	 * @return array<string>
+	 */
+	private static function htmlspecialchars_decodes(array $strings): array {
+		return array_map(static fn(string $s) => htmlspecialchars_decode($s, ENT_QUOTES), $strings);
+	}
+
 	/**
 	 * Parse the search string to find entry (article) IDs.
 	 */
 	private function parseEntryIds(string $input): string {
-		if (preg_match_all('/\be:(?P<search>[0-9,]*)/', $input, $matches)) {
+		if (preg_match_all('/\\be:(?P<search>[0-9,]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$ids_lists = $matches['search'];
 			$this->entry_ids = [];
@@ -273,7 +341,7 @@ class FreshRSS_Search {
 	}
 
 	private function parseNotEntryIds(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]e:(?P<search>[0-9,]*)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]e:(?P<search>[0-9,]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$ids_lists = $matches['search'];
 			$this->not_entry_ids = [];
@@ -289,7 +357,7 @@ class FreshRSS_Search {
 	}
 
 	private function parseFeedIds(string $input): string {
-		if (preg_match_all('/\bf:(?P<search>[0-9,]*)/', $input, $matches)) {
+		if (preg_match_all('/\\bf:(?P<search>[0-9,]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$ids_lists = $matches['search'];
 			$this->feed_ids = [];
@@ -307,7 +375,7 @@ class FreshRSS_Search {
 	}
 
 	private function parseNotFeedIds(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]f:(?P<search>[0-9,]*)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]f:(?P<search>[0-9,]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$ids_lists = $matches['search'];
 			$this->not_feed_ids = [];
@@ -328,7 +396,7 @@ class FreshRSS_Search {
 	 * Parse the search string to find tags (labels) IDs.
 	 */
 	private function parseLabelIds(string $input): string {
-		if (preg_match_all('/\b[lL]:(?P<search>[0-9,]+|[*])/', $input, $matches)) {
+		if (preg_match_all('/\\b[lL]:(?P<search>[0-9,]+|[*])/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$ids_lists = $matches['search'];
 			$this->label_ids = [];
@@ -350,7 +418,7 @@ class FreshRSS_Search {
 	}
 
 	private function parseNotLabelIds(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-][lL]:(?P<search>[0-9,]+|[*])/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-][lL]:(?P<search>[0-9,]+|[*])/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$ids_lists = $matches['search'];
 			$this->not_label_ids = [];
@@ -376,11 +444,11 @@ class FreshRSS_Search {
 	 */
 	private function parseLabelNames(string $input): string {
 		$names_lists = [];
-		if (preg_match_all('/\blabels?:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+		if (preg_match_all('/\\blabels?:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$names_lists = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
 		}
-		if (preg_match_all('/\blabels?:(?P<search>[^\s"]*)/', $input, $matches)) {
+		if (preg_match_all('/\\blabels?:(?P<search>[^\s"]*)/', $input, $matches)) {
 			$names_lists = array_merge($names_lists, $matches['search']);
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -402,11 +470,11 @@ class FreshRSS_Search {
 	 */
 	private function parseNotLabelNames(string $input): string {
 		$names_lists = [];
-		if (preg_match_all('/(?<=\s|^)[!-]labels?:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]labels?:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$names_lists = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
 		}
-		if (preg_match_all('/(?<=\s|^)[!-]labels?:(?P<search>[^\s"]*)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]labels?:(?P<search>[^\\s"]*)/', $input, $matches)) {
 			$names_lists = array_merge($names_lists, $matches['search']);
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -428,11 +496,15 @@ class FreshRSS_Search {
 	 * The search is the first word following the keyword.
 	 */
 	private function parseIntitleSearch(string $input): string {
-		if (preg_match_all('/\bintitle:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+		if (preg_match_all('#\\bintitle:(?P<search>/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->intitle_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/\\bintitle:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->intitle = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
 		}
-		if (preg_match_all('/\bintitle:(?P<search>[^\s"]*)/', $input, $matches)) {
+		if (preg_match_all('/\\bintitle:(?P<search>[^\s"]*)/', $input, $matches)) {
 			$this->intitle = array_merge($this->intitle ?: [], $matches['search']);
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -444,11 +516,15 @@ class FreshRSS_Search {
 	}
 
 	private function parseNotIntitleSearch(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]intitle:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+		if (preg_match_all('#(?<=\\s|^)[!-]intitle:(?P<search>/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->not_intitle_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/(?<=\\s|^)[!-]intitle:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->not_intitle = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
 		}
-		if (preg_match_all('/(?<=\s|^)[!-]intitle:(?P<search>[^\s"]*)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]intitle:(?P<search>[^\s"]*)/', $input, $matches)) {
 			$this->not_intitle = array_merge($this->not_intitle ?: [], $matches['search']);
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -465,11 +541,15 @@ class FreshRSS_Search {
 	 * a delimiter. Supported delimiters are single quote (') and double quotes (").
 	 */
 	private function parseAuthorSearch(string $input): string {
-		if (preg_match_all('/\bauthor:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+		if (preg_match_all('#\\bauthor:(?P<search>/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->author_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/\\bauthor:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->author = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
 		}
-		if (preg_match_all('/\bauthor:(?P<search>[^\s"]*)/', $input, $matches)) {
+		if (preg_match_all('/\\bauthor:(?P<search>[^\s"]*)/', $input, $matches)) {
 			$this->author = array_merge($this->author ?: [], $matches['search']);
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -481,11 +561,15 @@ class FreshRSS_Search {
 	}
 
 	private function parseNotAuthorSearch(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]author:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+		if (preg_match_all('#(?<=\\s|^)[!-]author:(?P<search>/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->not_author_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/(?<=\\s|^)[!-]author:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->not_author = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
 		}
-		if (preg_match_all('/(?<=\s|^)[!-]author:(?P<search>[^\s"]*)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]author:(?P<search>[^\s"]*)/', $input, $matches)) {
 			$this->not_author = array_merge($this->not_author ?: [], $matches['search']);
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -501,19 +585,41 @@ class FreshRSS_Search {
 	 * The search is the first word following the keyword.
 	 */
 	private function parseInurlSearch(string $input): string {
-		if (preg_match_all('/\binurl:(?P<search>[^\s]*)/', $input, $matches)) {
+		if (preg_match_all('#\\binurl:(?P<search>/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->inurl_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/\\binurl:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->inurl = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
-			$this->inurl = self::removeEmptyValues($this->inurl);
+		}
+		if (preg_match_all('/\\binurl:(?P<search>[^\\s]*)/', $input, $matches)) {
+			$this->inurl = $matches['search'];
+			$input = str_replace($matches[0], '', $input);
+		}
+		$this->inurl = self::removeEmptyValues($this->inurl);
+		if (empty($this->inurl)) {
+			$this->inurl = null;
 		}
 		return $input;
 	}
 
 	private function parseNotInurlSearch(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]inurl:(?P<search>[^\s]*)/', $input, $matches)) {
+		if (preg_match_all('#(?<=\\s|^)[!-]inurl:(?P<search>/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->not_inurl_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/(?<=\\s|^)[!-]inurl:(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->not_inurl = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
-			$this->not_inurl = self::removeEmptyValues($this->not_inurl);
+		}
+		if (preg_match_all('/(?<=\\s|^)[!-]inurl:(?P<search>[^\\s]*)/', $input, $matches)) {
+			$this->not_inurl = $matches['search'];
+			$input = str_replace($matches[0], '', $input);
+		}
+		$this->not_inurl = self::removeEmptyValues($this->not_inurl);
+		if (empty($this->not_inurl)) {
+			$this->not_inurl = null;
 		}
 		return $input;
 	}
@@ -523,7 +629,7 @@ class FreshRSS_Search {
 	 * The search is the first word following the keyword.
 	 */
 	private function parseDateSearch(string $input): string {
-		if (preg_match_all('/\bdate:(?P<search>[^\s]*)/', $input, $matches)) {
+		if (preg_match_all('/\\bdate:(?P<search>[^\\s]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$dates = self::removeEmptyValues($matches['search']);
 			if (!empty($dates[0])) {
@@ -534,7 +640,7 @@ class FreshRSS_Search {
 	}
 
 	private function parseNotDateSearch(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]date:(?P<search>[^\s]*)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]date:(?P<search>[^\\s]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$dates = self::removeEmptyValues($matches['search']);
 			if (!empty($dates[0])) {
@@ -550,7 +656,7 @@ class FreshRSS_Search {
 	 * The search is the first word following the keyword.
 	 */
 	private function parsePubdateSearch(string $input): string {
-		if (preg_match_all('/\bpubdate:(?P<search>[^\s]*)/', $input, $matches)) {
+		if (preg_match_all('/\\bpubdate:(?P<search>[^\\s]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$dates = self::removeEmptyValues($matches['search']);
 			if (!empty($dates[0])) {
@@ -561,7 +667,7 @@ class FreshRSS_Search {
 	}
 
 	private function parseNotPubdateSearch(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]pubdate:(?P<search>[^\s]*)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-]pubdate:(?P<search>[^\\s]*)/', $input, $matches)) {
 			$input = str_replace($matches[0], '', $input);
 			$dates = self::removeEmptyValues($matches['search']);
 			if (!empty($dates[0])) {
@@ -577,20 +683,44 @@ class FreshRSS_Search {
 	 * The search is the first word following the #.
 	 */
 	private function parseTagsSearch(string $input): string {
-		if (preg_match_all('/#(?P<search>[^\s]+)/', $input, $matches)) {
+		if (preg_match_all('%#(?P<search>/.*?(?<!\\\\)/[im]*)%', $input, $matches)) {
+			$this->tags_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/#(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->tags = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
-			$this->tags = self::removeEmptyValues($this->tags);
+		}
+		if (preg_match_all('/#(?P<search>[^\\s]+)/', $input, $matches)) {
+			$this->tags = $matches['search'];
+			$input = str_replace($matches[0], '', $input);
+		}
+		$this->tags = self::removeEmptyValues($this->tags);
+		if (empty($this->tags)) {
+			$this->tags = null;
+		} else {
 			$this->tags = self::decodeSpaces($this->tags);
 		}
 		return $input;
 	}
 
 	private function parseNotTagsSearch(string $input): string {
-		if (preg_match_all('/(?<=\s|^)[!-]#(?P<search>[^\s]+)/', $input, $matches)) {
+		if (preg_match_all('%(?<=\\s|^)[!-]#(?P<search>/.*?(?<!\\\\)/[im]*)%', $input, $matches)) {
+			$this->not_tags_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/(?<=\\s|^)[!-]#(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+			$this->not_tags = $matches['search'];
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/(?<=\\s|^)[!-]#(?P<search>[^\\s]+)/', $input, $matches)) {
 			$this->not_tags = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
-			$this->not_tags = self::removeEmptyValues($this->not_tags);
+		}
+		$this->not_tags = self::removeEmptyValues($this->not_tags);
+		if (empty($this->not_tags)) {
+			$this->not_tags = null;
+		} else {
 			$this->not_tags = self::decodeSpaces($this->not_tags);
 		}
 		return $input;
@@ -599,13 +729,18 @@ class FreshRSS_Search {
 	/**
 	 * Parse the search string to find search values.
 	 * Every word is a distinct search value using a delimiter.
-	 * Supported delimiters are single quote (') and double quotes (").
+	 * Supported delimiters are single quote (') and double quotes (") and regex (/).
 	 */
 	private function parseQuotedSearch(string $input): string {
 		$input = self::cleanSearch($input);
 		if ($input === '') {
 			return '';
 		}
+		if (preg_match_all('#(?<=\\s|^)(?<![!-\\\\])(?P<search>/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->search_regex = self::htmlspecialchars_decodes($matches['search']);
+			//TODO: Replace all those str_replace with PREG_OFFSET_CAPTURE
+			$input = str_replace($matches[0], '', $input);
+		}
 		if (preg_match_all('/(?<![!-])(?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->search = $matches['search'];
 			//TODO: Replace all those str_replace with PREG_OFFSET_CAPTURE
@@ -636,7 +771,11 @@ class FreshRSS_Search {
 		if ($input === '') {
 			return '';
 		}
-		if (preg_match_all('/(?<=\s|^)[!-](?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
+		if (preg_match_all('#(?<=\\s|^)[!-](?P<search>(?<!\\\\)/.*?(?<!\\\\)/[im]*)#', $input, $matches)) {
+			$this->not_search_regex = self::htmlspecialchars_decodes($matches['search']);
+			$input = str_replace($matches[0], '', $input);
+		}
+		if (preg_match_all('/(?<=\\s|^)[!-](?P<delim>[\'"])(?P<search>.*)(?P=delim)/U', $input, $matches)) {
 			$this->not_search = $matches['search'];
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -644,7 +783,7 @@ class FreshRSS_Search {
 		if ($input === '') {
 			return '';
 		}
-		if (preg_match_all('/(?<=\s|^)[!-](?P<search>[^\s]+)/', $input, $matches)) {
+		if (preg_match_all('/(?<=\\s|^)[!-](?P<search>[^\\s]+)/', $input, $matches)) {
 			$this->not_search = array_merge(is_array($this->not_search) ? $this->not_search : [], $matches['search']);
 			$input = str_replace($matches[0], '', $input);
 		}
@@ -656,7 +795,7 @@ class FreshRSS_Search {
 	 * Remove all unnecessary spaces in the search
 	 */
 	private static function cleanSearch(string $input): string {
-		$input = preg_replace('/\s+/', ' ', $input);
+		$input = preg_replace('/\\s+/', ' ', $input);
 		if (!is_string($input)) {
 			return '';
 		}
-- 
cgit v1.2.3