wp-optimize / webp / class-wpo-webp-alter-html.php

class-wpo-webp-alter-html.php in WP-Optimize – Cache, Compress images, Minify & Clean database to boost page speed & performance 4.5.4, at webp/class-wpo-webp-alter-html.php

197 lines 5.5 KB

Raw Download Zip

1	<?php
2
3	if (!defined('ABSPATH')) die('No direct access allowed');
4
5	if (!class_exists('WPO_WebP_Alter_HTML')) :
6
7	class WPO_WebP_Alter_HTML {
8
9	/**
10	* HTML tags that may contain image-related attributes eligible for WebP replacement.
11	*/
12	const SUPPORTED_TAGS = array('img', 'source', 'input', 'iframe', 'div', 'li', 'link', 'a', 'section', 'video');
13
14	/**
15	* Regex pattern to identify image-related attribute names such as src, srcset,
16	* poster, and common data-* attributes used for lazy loading or responsive images.
17	*/
18	const IMAGE_ATTRIBUTE_PATTERN = '#^(src\|srcset\|poster\|(data-[^=](lazy\|small\|slide\|img\|large\|src\|thumb\|source\|set\|bg-url)[^=]))$#i';
19
20	/**
21	* Constructor — private to enforce singleton pattern.
22	*/
23	private function __construct() {
24	}
25
26	/**
27	* Returns singleton instance.
28	*
29	* @return self
30	*/
31	public static function get_instance(): self {
32	static $_instance = null;
33	if (null === $_instance) {
34	$_instance = new self();
35	}
36	return $_instance;
37	}
38
39	/**
40	* Alter HTML to replace image-related attributes with WebP URLs where available.
41	*
42	* @param string $html HTML document as string.
43	* @return string
44	*/
45	public function alter_html(string $html): string {
46	if (apply_filters('wpo_disable_webp_alter_html', false)) {
47	return $html;
48	}
49
50	// MAX_FILE_SIZE is defined in simple_html_dom — ensure it has a safe default.
51	defined('MAX_FILE_SIZE') \|\| define('MAX_FILE_SIZE', 600000);
52
53	$dom = WP_Optimize_Utils::get_simple_html_dom_object($html);
54
55	if (false === $dom) {
56	return $this->get_skipped_html($html);
57	}
58
59	$this->replace_image_attributes($dom);
60
61	return $dom->save();
62	}
63
64	/**
65	* Walks all supported tags in the DOM and replaces image attribute values
66	* with their WebP counterparts when available.
67	*
68	* @param simplehtmldom\HtmlDocument $dom A simple_html_dom instance.
69	*
70	* @return void
71	*/
72	private function replace_image_attributes($dom): void {
73	foreach (self::SUPPORTED_TAGS as $tag) {
74	foreach ($dom->find($tag) as $elem) {
75	$attributes = $elem->getAllAttributes();
76	foreach ($attributes as $attr_name => $attr_value) {
77	if ($this->is_image_attribute($attr_name)) {
78	$elem->setAttribute($attr_name, $this->handle_attribute($attr_value));
79	}
80	}
81	}
82	}
83	}
84
85	/**
86	* Returns the original HTML annotated with a comment explaining why
87	* the WebP alteration was skipped.
88	*
89	* @param string $html The original HTML document.
90	* @return string
91	*/
92	private function get_skipped_html($html): string {
93	if (strlen($html) > MAX_FILE_SIZE) {
94	return $html . "\n" . "<!-- Alter HTML was skipped because the HTML is too big to process! " .
95	"(limit is set to " . MAX_FILE_SIZE . " bytes) -->";
96	}
97	return $html . "\n" . "<!-- Alter HTML was skipped because the helper library refused to process the html -->";
98	}
99
100	/**
101	* Returns the URL with a .webp extension appended if a WebP version exists on disk.
102	*
103	* @param string $url URL of the original image.
104	* @return string
105	*/
106	private function maybe_replace_url($url): string {
107	if ($this->is_webp_version_available($url)) {
108	$url .= '.webp';
109	}
110	return $url;
111	}
112
113	/**
114	* Processes a srcset attribute value, replacing each image URL with its
115	* WebP counterpart when available.
116	*
117	* A srcset value is a comma-separated list of entries such as
118	* "image.jpg 520w" or "image.jpg 2x" or just "image.jpg".
119	*
120	* @param string $attr_value The raw srcset attribute value.
121	* @return string
122	*/
123	private function handle_srcset($attr_value): string {
124	// $attr_value is ie: <img data-x="1.jpg 1000w, 2.jpg">
125	$srcset_arr = explode(',', $attr_value);
126	foreach ($srcset_arr as $i => $srcset_entry) {
127	// $srcset_entry is ie "image.jpg 520w", but can also lack width, ie just "image.jpg"
128	// it can also be ie "image.jpg 2x"
129	$srcset_entry = trim($srcset_entry);
130	$entry_parts = preg_split('/\s+/', $srcset_entry, 2);
131
132	if (is_array($entry_parts) && 2 === count($entry_parts)) {
133	list($src, $descriptors) = $entry_parts;
134	} else {
135	$src = $srcset_entry;
136	$descriptors = null;
137	}
138
139	$url = $this->maybe_replace_url($src);
140	$srcset_arr[$i] = $url . (isset($descriptors) ? ' ' . $descriptors : '');
141	}
142	return implode(', ', $srcset_arr);
143	}
144
145	/**
146	* Determines whether an attribute value resembles a srcset declaration
147	* (contains a width or pixel-density descriptor).
148	*
149	* @param string $value The attribute value to inspect.
150	* @return bool
151	*/
152	private function looks_like_srcset($value): bool {
153	return (bool) preg_match('#\s\d*(w\|x)#', $value);
154	}
155
156	/**
157	* Dispatches attribute value handling based on whether it looks like a
158	* srcset or a plain image source URL.
159	*
160	* @param string $value The attribute value.
161	* @return string
162	*/
163	private function handle_attribute($value): string {
164	if ($this->looks_like_srcset($value)) {
165	return $this->handle_srcset($value);
166	}
167	return $this->maybe_replace_url($value);
168	}
169
170	/**
171	* Determines whether an attribute name is image-related and should be
172	* considered for WebP replacement.
173	*
174	* @param string $attr_name The attribute name.
175	* @return bool
176	*/
177	private function is_image_attribute($attr_name): bool {
178	return (bool) preg_match(self::IMAGE_ATTRIBUTE_PATTERN, $attr_name);
179	}
180
181	/**
182	* Checks whether a WebP version of the given image URL exists on disk.
183	*
184	* @param string $url The original image URL.
185	* @return bool
186	*/
187	private function is_webp_version_available($url): bool {
188	$filename = WP_Optimize_Utils::get_file_path($url);
189	if (empty($filename)) {
190	return false;
191	}
192	return file_exists($filename . '.webp');
193	}
194	}
195
196	endif;
197