From 49766ab01f24d1be5743c353b9a6d4d09f55f5b8 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@fakecake.org>
Date: Thu, 27 Mar 2025 22:10:16 +0300
Subject: filter test dialog improvements:

 - properly return results for filter rules matching specific feeds or
   categories
 - fix test results never returned for Uncategorized
 - show tooltip with specific word match and matched rule on resulting
   item hover
---
 js/CommonFilters.js | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

(limited to 'js')

diff --git a/js/CommonFilters.js b/js/CommonFilters.js
index b7bb69a15..83bc4cae9 100644
--- a/js/CommonFilters.js
+++ b/js/CommonFilters.js
@@ -19,6 +19,9 @@ const	Filters = {
 			ACTION_REMOVE_TAG: 10,
 			PARAM_ACTIONS: [4, 6, 7, 9, 10],
 			filter_info: {},
+			formatMatchedRules: function(rules) {
+				return rules.map((r) => r.regexp_matches[0] + ' - ' + r.reg_exp).join('\n');
+			},
 			test: function() {
 				const test_dialog = new fox.SingleUseDialog({
 					title: "Test Filter",
@@ -43,8 +46,8 @@ const	Filters = {
 												.replace("%f", test_dialog.results)
 												.replace("%d", offset);
 
-										results_list.innerHTML += result.items.reduce((current, item) => current + `<li><span class='title'>${item.title}</span><br>
-												<span class='feed'>${item.feed_title}</span>, <span class='date'>${item.date}</span>
+										results_list.innerHTML += result.items.reduce((current, item) => current + `<li title="${App.escapeHtml(dialog.formatMatchedRules(item.matched_rules))}"><span class='title'>${item.title}</span>
+												&mdash; <span class='feed'>${item.feed_title}</span>, <span class='date'>${item.date}</span>
 												<div class='preview text-muted'>${item.content_preview}</div></li>`, '');
 
 										// get the next batch if there may be more available and testing limits haven't been reached
-- 
cgit v1.2.3-54-g00ecf


From 8986a3e7ee4118b7fae9d93c09c870e72de0e49d Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@fakecake.org>
Date: Fri, 28 Mar 2025 07:59:46 +0300
Subject: add limited highlighting of filter test results based on matched
 rules

---
 classes/Pref_Filters.php | 36 ++++++++++++++++++--
 classes/Sanitizer.php    | 89 +++++++++++++++++++++++++++++++++---------------
 js/CommonFilters.js      |  2 +-
 phpstan.neon             |  1 +
 4 files changed, 97 insertions(+), 31 deletions(-)

(limited to 'js')

diff --git a/classes/Pref_Filters.php b/classes/Pref_Filters.php
index 1ca542285..a7528865a 100644
--- a/classes/Pref_Filters.php
+++ b/classes/Pref_Filters.php
@@ -176,11 +176,43 @@ class Pref_Filters extends Handler_Protected {
 					},
 					$entry, $excerpt_length);
 
+				$matches = [];
+
+				$content_preview = $entry["content_preview"];
+				$content_title = $entry["title"];
+
+				// is it even possible to have multiple matched rules here?
+				foreach ($matched_rules as $rule) {
+					$can_highlight_content = false;
+					$can_highlight_title = false;
+
+					$matches[] = $rule['regexp_matches'][0];
+
+					switch ($rule['type']) {
+						case "both":
+							$can_highlight_title = true;
+							$can_highlight_content = true;
+							break;
+						case "title":
+							$can_highlight_title = true;
+							break;
+						case "content":
+							$can_highlight_content = true;
+							break;
+					}
+
+					if ($can_highlight_content)
+						$content_preview = Sanitizer::highlight_words_str($content_preview, $matches);
+
+					if ($can_highlight_title)
+						$content_title = Sanitizer::highlight_words_str($content_title, $matches);
+				}
+
 				$rv['items'][] = [
-					'title' => $entry['title'],
+					'title' => $content_title,
 					'feed_title' => $entry['feed_title'],
 					'date' => mb_substr($entry['date_entered'], 0, 16),
-					'content_preview' => $entry['content_preview'],
+					'content_preview' => $content_preview,
 					'matched_rules' => $matched_rules,
 				];
 			}
diff --git a/classes/Sanitizer.php b/classes/Sanitizer.php
index 94d6fe621..2ae07d8d3 100644
--- a/classes/Sanitizer.php
+++ b/classes/Sanitizer.php
@@ -59,6 +59,65 @@ class Sanitizer {
 		return parse_url(Config::get_self_url(), PHP_URL_SCHEME) == 'https';
 	}
 
+	/** @param array<string> $words */
+	public static function highlight_words_str(string $str, array $words) : string {
+		$doc = new DOMDocument();
+
+		if ($doc->loadHTML('<?xml encoding="UTF-8"><span>' . $str . '</span>')) {
+			$xpath = new DOMXPath($doc);
+
+			if (self::highlight_words($doc, $xpath, $words)) {
+				$res = $doc->saveHTML();
+
+				/* strip everything outside of <body>...</body> */
+				$res_frag = array();
+
+				if (preg_match('/<body>(.*)<\/body>/is', $res, $res_frag)) {
+					return $res_frag[1];
+				} else {
+					return $res;
+				}
+			}
+		}
+
+		return $str;
+	}
+
+	/** @param array<string> $words */
+	public static function highlight_words(DOMDocument &$doc, DOMXPath $xpath, array $words) : bool {
+		$rv = false;
+
+		foreach ($words as $word) {
+
+			// http://stackoverflow.com/questions/4081372/highlight-keywords-in-a-paragraph
+			$elements = $xpath->query("//*/text()");
+
+			foreach ($elements as $child) {
+
+				$fragment = $doc->createDocumentFragment();
+				$text = $child->textContent;
+
+				while (($pos = mb_stripos($text, $word)) !== false) {
+					$fragment->appendChild(new DOMText(mb_substr($text, 0, (int)$pos)));
+					$word = mb_substr($text, (int)$pos, mb_strlen($word));
+					$highlight = $doc->createElement('span');
+					$highlight->appendChild(new DOMText($word));
+					$highlight->setAttribute('class', 'highlight');
+					$fragment->appendChild($highlight);
+					$text = mb_substr($text, $pos + mb_strlen($word));
+				}
+
+				if (!empty($text)) $fragment->appendChild(new DOMText($text));
+
+				$child->parentNode->replaceChild($fragment, $child);
+
+				$rv = true;
+			}
+		}
+
+		return $rv;
+	}
+
 	/**
 	 * @param array<int, string>|null $highlight_words Words to highlight in the HTML output.
 	 *
@@ -197,34 +256,8 @@ class Sanitizer {
 			$div->appendChild($entry);
 		}
 
-		if (is_array($highlight_words)) {
-			foreach ($highlight_words as $word) {
-
-				// http://stackoverflow.com/questions/4081372/highlight-keywords-in-a-paragraph
-
-				$elements = $xpath->query("//*/text()");
-
-				foreach ($elements as $child) {
-
-					$fragment = $doc->createDocumentFragment();
-					$text = $child->textContent;
-
-					while (($pos = mb_stripos($text, $word)) !== false) {
-						$fragment->appendChild(new DOMText(mb_substr($text, 0, (int)$pos)));
-						$word = mb_substr($text, (int)$pos, mb_strlen($word));
-						$highlight = $doc->createElement('span');
-						$highlight->appendChild(new DOMText($word));
-						$highlight->setAttribute('class', 'highlight');
-						$fragment->appendChild($highlight);
-						$text = mb_substr($text, $pos + mb_strlen($word));
-					}
-
-					if (!empty($text)) $fragment->appendChild(new DOMText($text));
-
-					$child->parentNode->replaceChild($fragment, $child);
-				}
-			}
-		}
+		if (is_array($highlight_words))
+			self::highlight_words($doc, $xpath, $highlight_words);
 
 		$res = $doc->saveHTML();
 
diff --git a/js/CommonFilters.js b/js/CommonFilters.js
index 83bc4cae9..f1f0e9e0c 100644
--- a/js/CommonFilters.js
+++ b/js/CommonFilters.js
@@ -20,7 +20,7 @@ const	Filters = {
 			PARAM_ACTIONS: [4, 6, 7, 9, 10],
 			filter_info: {},
 			formatMatchedRules: function(rules) {
-				return rules.map((r) => r.regexp_matches[0] + ' - ' + r.reg_exp).join('\n');
+				return rules.map((r) => r.regexp_matches[0] + ' - ' + r.reg_exp + ' (' + r.type + ')').join('\n');
 			},
 			test: function() {
 				const test_dialog = new fox.SingleUseDialog({
diff --git a/phpstan.neon b/phpstan.neon
index 74daee0ce..5bd18c58d 100644
--- a/phpstan.neon
+++ b/phpstan.neon
@@ -32,6 +32,7 @@ parameters:
       - plugins.local/**/tests/*
       - plugins.local/*/vendor/intervention/*
       - plugins.local/*/vendor/psr/log/*
+      - plugins.local/af_readability/*
       - plugins.local/cache_s3/vendor/*
       - plugins/**/test/*
       - plugins/**/Test/*
-- 
cgit v1.2.3-54-g00ecf


From b4962b670d02f7c5d4da3cafe99cffffe059e8a1 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@fakecake.org>
Date: Sun, 30 Mar 2025 20:21:06 +0300
Subject: stop sending matched content twice for the tooltip, use smaller
 objects containing only regular expression and rule type

---
 classes/Pref_Filters.php | 8 +++++++-
 js/CommonFilters.js      | 4 ++--
 2 files changed, 9 insertions(+), 3 deletions(-)

(limited to 'js')

diff --git a/classes/Pref_Filters.php b/classes/Pref_Filters.php
index e30a1468c..d5844b902 100644
--- a/classes/Pref_Filters.php
+++ b/classes/Pref_Filters.php
@@ -171,6 +171,7 @@ class Pref_Filters extends Handler_Protected {
 				$content_preview = "";
 
 				$matches = [];
+				$rules = [];
 
 				$entry_title = $entry["title"];
 
@@ -184,6 +185,11 @@ class Pref_Filters extends Handler_Protected {
 
 					$matches[] = $rule_regexp_match;
 
+					$rules[] = [
+						'reg_exp' => $rule['reg_exp'],
+						'type' => $rule['type']
+					];
+
 					if (in_array($rule['type'], ['content', 'both'])) {
 						// also stripping [\r\n\t] to match what's done for content in RSSUtils#get_article_filters()
 						$entry_content_text = strip_tags(preg_replace("/[\r\n\t]/", "", $entry["content"]));
@@ -232,7 +238,7 @@ class Pref_Filters extends Handler_Protected {
 					'feed_title' => $entry['feed_title'],
 					'date' => mb_substr($entry['date_entered'], 0, 16),
 					'content_preview' => $content_preview,
-					'matched_rules' => $matched_rules, // TODO shorten object passed to frontend
+					'rules' => $rules
 				];
 			}
 		}
diff --git a/js/CommonFilters.js b/js/CommonFilters.js
index f1f0e9e0c..ce871a2af 100644
--- a/js/CommonFilters.js
+++ b/js/CommonFilters.js
@@ -20,7 +20,7 @@ const	Filters = {
 			PARAM_ACTIONS: [4, 6, 7, 9, 10],
 			filter_info: {},
 			formatMatchedRules: function(rules) {
-				return rules.map((r) => r.regexp_matches[0] + ' - ' + r.reg_exp + ' (' + r.type + ')').join('\n');
+				return rules.map((r) => r.reg_exp + ' (' + r.type + ')').join('\n');
 			},
 			test: function() {
 				const test_dialog = new fox.SingleUseDialog({
@@ -46,7 +46,7 @@ const	Filters = {
 												.replace("%f", test_dialog.results)
 												.replace("%d", offset);
 
-										results_list.innerHTML += result.items.reduce((current, item) => current + `<li title="${App.escapeHtml(dialog.formatMatchedRules(item.matched_rules))}"><span class='title'>${item.title}</span>
+										results_list.innerHTML += result.items.reduce((current, item) => current + `<li title="${App.escapeHtml(dialog.formatMatchedRules(item.rules))}"><span class='title'>${item.title}</span>
 												&mdash; <span class='feed'>${item.feed_title}</span>, <span class='date'>${item.date}</span>
 												<div class='preview text-muted'>${item.content_preview}</div></li>`, '');
 
-- 
cgit v1.2.3-54-g00ecf


From 2d041f7d28ce555bffaf1b99eea5d0199e8a3c29 Mon Sep 17 00:00:00 2001
From: Andrew Dolgov <fox@fakecake.org>
Date: Sun, 30 Mar 2025 20:41:50 +0300
Subject: use server-side localized formatting for matching rule to display as
 a tooltip (for now)

---
 classes/Pref_Filters.php | 20 ++++++++++----------
 js/CommonFilters.js      |  5 +----
 2 files changed, 11 insertions(+), 14 deletions(-)

(limited to 'js')

diff --git a/classes/Pref_Filters.php b/classes/Pref_Filters.php
index d5844b902..a6063d898 100644
--- a/classes/Pref_Filters.php
+++ b/classes/Pref_Filters.php
@@ -95,7 +95,6 @@ class Pref_Filters extends Handler_Protected {
 
 			if (is_array($rule)) {
 				$rule['type'] = $filter_types[$rule['filter_type']];
-				unset($rule['filter_type']);
 				array_push($filter['rules'], $rule);
 
 				$scope_inner_qparts = [];
@@ -185,10 +184,7 @@ class Pref_Filters extends Handler_Protected {
 
 					$matches[] = $rule_regexp_match;
 
-					$rules[] = [
-						'reg_exp' => $rule['reg_exp'],
-						'type' => $rule['type']
-					];
+					$rules[] = self::_get_rule_name($rule, '');
 
 					if (in_array($rule['type'], ['content', 'both'])) {
 						// also stripping [\r\n\t] to match what's done for content in RSSUtils#get_article_filters()
@@ -452,7 +448,7 @@ class Pref_Filters extends Handler_Protected {
 	/**
 	 * @param array<string, mixed>|null $rule
 	 */
-	private function _get_rule_name(?array $rule = null): string {
+	private function _get_rule_name(?array $rule = null, string $format = 'html'): string {
 		if (!$rule) $rule = json_decode(clean($_REQUEST["rule"]), true);
 
 		$feeds = $rule["feed_id"];
@@ -487,10 +483,14 @@ class Pref_Filters extends Handler_Protected {
 
 		$inverse = isset($rule["inverse"]) ? "inverse" : "";
 
-		return "<span class='filterRule $inverse'>" .
-			T_sprintf("%s on %s in %s %s", htmlspecialchars($rule["reg_exp"]),
-			"<span class='field'>$filter_type</span>", "<span class='feed'>$feed</span>", isset($rule["inverse"]) ? __("(inverse)") : "") . "</span>";
-	}
+		if ($format === 'html')
+			return "<span class='filterRule $inverse'>" .
+				T_sprintf("%s on %s in %s %s", htmlspecialchars($rule["reg_exp"]),
+				"<span class='field'>$filter_type</span>", "<span class='feed'>$feed</span>", isset($rule["inverse"]) ? __("(inverse)") : "") . "</span>";
+		else
+			return T_sprintf("%s on %s in %s %s", $rule["reg_exp"],
+				$filter_type, $feed, isset($rule["inverse"]) ? __("(inverse)") : "");
+		}
 
 	function printRuleName(): void {
 		print $this->_get_rule_name(json_decode(clean($_REQUEST["rule"]), true));
diff --git a/js/CommonFilters.js b/js/CommonFilters.js
index ce871a2af..9c0fc5cce 100644
--- a/js/CommonFilters.js
+++ b/js/CommonFilters.js
@@ -19,9 +19,6 @@ const	Filters = {
 			ACTION_REMOVE_TAG: 10,
 			PARAM_ACTIONS: [4, 6, 7, 9, 10],
 			filter_info: {},
-			formatMatchedRules: function(rules) {
-				return rules.map((r) => r.reg_exp + ' (' + r.type + ')').join('\n');
-			},
 			test: function() {
 				const test_dialog = new fox.SingleUseDialog({
 					title: "Test Filter",
@@ -46,7 +43,7 @@ const	Filters = {
 												.replace("%f", test_dialog.results)
 												.replace("%d", offset);
 
-										results_list.innerHTML += result.items.reduce((current, item) => current + `<li title="${App.escapeHtml(dialog.formatMatchedRules(item.rules))}"><span class='title'>${item.title}</span>
+										results_list.innerHTML += result.items.reduce((current, item) => current + `<li title="${App.escapeHtml(item.rules.join('\n'))}"><span class='title'>${item.title}</span>
 												&mdash; <span class='feed'>${item.feed_title}</span>, <span class='date'>${item.date}</span>
 												<div class='preview text-muted'>${item.content_preview}</div></li>`, '');
 
-- 
cgit v1.2.3-54-g00ecf