Completed
Push — master ( ad19ae...1ab7e1 )
by Nazar
04:30
created

Assets_processing::html()   A

Complexity

Conditions 2
Paths 1

Size

Total Lines 13
Code Lines 9

Duplication

Lines 0
Ratio 0 %

Code Coverage

Tests 9
CRAP Score 2

Importance

Changes 0
Metric Value
cc 2
eloc 9
nc 1
nop 5
dl 0
loc 13
ccs 9
cts 9
cp 1
crap 2
rs 9.4285
c 0
b 0
f 0
1
<?php
2
/**
3
 * @package   CleverStyle Framework
4
 * @author    Nazar Mokrynskyi <[email protected]>
5
 * @copyright Copyright (c) 2014-2016, Nazar Mokrynskyi
6
 * @license   MIT License, see license.txt
7
 */
8
namespace cs\Page;
9
10
/**
11
 * Class includes few methods used for processing CSS, JS and HTML files before putting into cache
12
 *
13
 * This is because CSS and HTML files may include other CSS, JS files, images, fonts and so on with absolute and relative paths.
14
 * Methods of this class handle all this assets, applies basic minification to CSS and JS files and produce single resulting file (relative paths to
15
 * files that can't be embedded are converted to absolute). This allows to decrease number of HTTP requests on page and avoid breaking of relative paths for
16
 * fonts, images and other assets after putting them into cache directory as well as minimize contents size by removing comments and other redundant stuff.
17
 */
18
class Assets_processing {
19
	/**
20
	 * Do not inline files bigger than 4 KiB
21
	 */
22
	const MAX_EMBEDDING_SIZE = 4096;
23
	protected static $extension_to_mime = [
24
		'jpeg' => 'image/jpg',
25
		'jpe'  => 'image/jpg',
26
		'jpg'  => 'image/jpg',
27
		'gif'  => 'image/gif',
28
		'png'  => 'image/png',
29
		'svg'  => 'image/svg+xml',
30
		'svgz' => 'image/svg+xml',
31
		'woff' => 'application/font-woff',
32
		//'woff2' => 'application/font-woff2'
33
	];
34
	/**
35
	 * Analyses file for images, fonts and css links and include they content into single resulting css file.
36
	 *
37
	 * Supports next file extensions for possible assets:
38
	 * jpeg, jpe, jpg, gif, png, ttf, ttc, svg, svgz, woff, css
39
	 *
40
	 * @param string   $data                   Content of processed file
41
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
42
	 * @param string   $target_directory_path  Target directory for resulting combined files
43
	 * @param string[] $not_embedded_resources Some resources like images and fonts might not be embedded into resulting CSS because of their size
44
	 *
45
	 * @return string    $data
46
	 */
47 10
	public static function css ($data, $file, $target_directory_path = PUBLIC_CACHE, &$not_embedded_resources = []) {
48 10
		$dir = dirname($file);
49
		/**
50
		 * Remove comments, tabs and new lines
51
		 */
52 10
		$data = preg_replace('#(/\*.*?\*/)|\t|\n|\r#s', ' ', $data);
53
		/**
54
		 * Remove unnecessary spaces
55
		 */
56 10
		$data = preg_replace('/\s*([,;>{}(])\s*/', '$1', $data);
57 10
		$data = preg_replace('/\s+/', ' ', $data);
58
		/**
59
		 * Return spaces required in media queries
60
		 */
61 10
		$data = preg_replace('/\s(and|or)\(/', ' $1 (', $data);
62
		/**
63
		 * Duplicated semicolons
64
		 */
65 10
		$data = preg_replace('/;+/m', ';', $data);
66
		/**
67
		 * Minify rgb colors declarations
68
		 */
69 10
		$data = preg_replace_callback(
70 10
			'/rgb\(([0-9,.]+)\)/i',
71
			function ($rgb) {
72 6
				$rgb = explode(',', $rgb[1]);
73
				return
74
					'#'.
75 6
					str_pad(dechex($rgb[0]), 2, 0, STR_PAD_LEFT).
76 6
					str_pad(dechex($rgb[1]), 2, 0, STR_PAD_LEFT).
77 6
					str_pad(dechex($rgb[2]), 2, 0, STR_PAD_LEFT);
78 10
			},
79
			$data
80
		);
81
		/**
82
		 * Minify repeated colors declarations
83
		 */
84 10
		$data = preg_replace('/#([0-9a-f])\1([0-9a-f])\2([0-9a-f])\3/i', '#$1$2$3', $data);
85
		/**
86
		 * Remove unnecessary zeros
87
		 */
88 10
		$data = preg_replace('/(\D)0\.(\d+)/i', '$1.$2', $data);
89
		/**
90
		 * Unnecessary spaces around colons (should have whitespace character after, otherwise might be `.c :disabled` and will be handled incorrectly)
91
		 */
92 10
		$data = preg_replace('/\s*:\s+/', ':', $data);
93
		/**
94
		 * Assets processing
95
		 */
96
		// TODO: replace by loop, track duplicated stuff that are subject to inlining and if they appear more than once, don't inline them
97 10
		$data = preg_replace_callback(
98 10
			'/url\((.*)\)|@import\s*(?:url\()?\s*([\'"].*[\'"])\s*\)??(.*);/U',
99
			function ($match) use ($dir, $target_directory_path, &$not_embedded_resources) {
100 10
				$path_matched = @$match[2] ?: $match[1];
101 10
				$path         = trim($path_matched, '\'" ');
102 10
				$link         = explode('?', $path, 2)[0];
103 10
				if (!static::is_relative_path_and_exists($link, $dir)) {
104 4
					return $match[0];
105
				}
106 10
				$extension     = file_extension($link);
107 10
				$absolute_path = static::absolute_path($link, $dir);
108 10
				$content       = file_get_contents($absolute_path);
109 10
				if ($extension == 'css' && @$match[2]) {
110
					/**
111
					 * Only inline CSS imports without media queries, imports with media queries will be placed as separate files
112
					 */
113 6
					if (!trim(@$match[3])) {
114 6
						return static::css($content, $absolute_path, $target_directory_path, $not_embedded_resources);
115
					} else {
116 6
						$filename = static::file_put_contents_with_hash(
117
							$target_directory_path,
118
							$extension,
119 6
							static::css($content, $absolute_path, $target_directory_path)
120
						);
121 6
						return str_replace($path_matched, "'./$filename'", $match[0]);
122
					}
123
				}
124 10
				if (!isset(static::$extension_to_mime[$extension])) {
125
					$filename = static::file_put_contents_with_hash($target_directory_path, $extension, $content);
126
					return str_replace($path_matched, "'./$filename'", $match[0]);
127
				}
128 10
				if (filesize($absolute_path) > static::MAX_EMBEDDING_SIZE) {
129 10
					$filename = static::file_put_contents_with_hash($target_directory_path, $extension, $content);
130 10
					if (strpos($path, '?') === false) {
131 6
						$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
132
					}
133 10
					return str_replace($path_matched, "'./$filename'", $match[0]);
134
				}
135 6
				$mime_type = static::$extension_to_mime[$extension];
136 6
				$content   = base64_encode($content);
137 6
				return str_replace($path, "data:$mime_type;charset=utf-8;base64,$content", $match[0]);
138 10
			},
139
			$data
140
		);
141 10
		return trim($data);
142
	}
143
	/**
144
	 * Put `$content` into `$dir` where filename is `md5($content)` with specified extension
145
	 *
146
	 * @param string $dir
147
	 * @param string $extension
148
	 * @param string $content
149
	 *
150
	 * @return string Filename (without full path)
151
	 */
152 10
	protected static function file_put_contents_with_hash ($dir, $extension, $content) {
153 10
		$hash = md5($content);
154 10
		file_put_contents("$dir/$hash.$extension", $content, LOCK_EX | FILE_BINARY);
155 10
		return "$hash.$extension";
156
	}
157
	/**
158
	 * Simple and fast JS minification
159
	 *
160
	 * @param string $data
161
	 *
162
	 * @return string
163
	 */
164 10
	public static function js ($data) {
165
		/**
166
		 * Split into array of lines
167
		 */
168 10
		$data = explode("\n", $data);
169
		/**
170
		 * Flag that is `true` when inside comment
171
		 */
172 10
		$in_comment              = false;
173 10
		$continue_after_position = -1;
174 10
		foreach ($data as $index => &$current_line) {
175 10
			if ($continue_after_position >= $index) {
176 6
				continue;
177
			}
178 10
			$next_line = isset($data[$index + 1]) ? trim($data[$index + 1]) : '';
179
			/**
180
			 * Remove starting and trailing spaces
181
			 */
182 10
			$current_line = trim($current_line);
183
			/**
184
			 * Remove single-line comments
185
			 */
186 10
			if (mb_strpos($current_line, '//') === 0) {
187 10
				$current_line = '';
188 10
				continue;
189
			}
190
			/**
191
			 * Starts with multi-line comment
192
			 */
193 10
			if (mb_strpos($current_line, '/*') === 0) {
194 10
				$in_comment = true;
195
			}
196 10
			if (!$in_comment) {
197 10
				$backticks_position = strpos($current_line, '`');
198
				/**
199
				 * Handling template strings can be tricky (since they might be multi-line), so let's fast-forward to the last backticks position and continue
200
				 * from there
201
				 */
202 10
				if ($backticks_position !== false) {
203 6
					$last_item_with_backticks = array_keys(
204
						array_filter(
205
							$data,
206
							function ($d) {
207 6
								return strpos($d, '`') !== false;
208 6
							}
209
						)
210
					);
211 6
					$last_item_with_backticks = array_pop($last_item_with_backticks);
212 6
					if ($last_item_with_backticks > $index) {
213 6
						$continue_after_position = $last_item_with_backticks;
214 6
						continue;
215
					}
216
				}
217
				/**
218
				 * Add new line at the end if only needed
219
				 */
220 10
				if (static::new_line_needed($current_line, $next_line)) {
221 10
					$current_line .= "\n";
222
				}
223
				/**
224
				 * Single-line comment
225
				 */
226 10
				$current_line = preg_replace('#^\s*//[^\'"]+$#', '', $current_line);
227
				/**
228
				 * If we are not sure - just add new line afterwards
229
				 */
230 10
				$current_line = preg_replace('#//.*$#', "\\0\n", $current_line);
231
			} else {
232
				/**
233
				 * End of multi-line comment
234
				 */
235 10
				if (strpos($current_line, '*/') !== false) {
236 10
					$current_line = explode('*/', $current_line)[1];
237 10
					$in_comment   = false;
238
				} else {
239 10
					$current_line = '';
240
				}
241
			}
242
		}
243 10
		$data = implode('', $data);
244 10
		$data = str_replace('</script>', '<\/script>', $data);
245 10
		return trim($data, ';').';';
246
	}
247
	/**
248
	 * @param string $current_line
249
	 * @param string $next_line
250
	 *
251
	 * @return bool
252
	 */
253 10
	protected static function new_line_needed ($current_line, $next_line) {
254
		/**
255
		 * Set of symbols that are safe to be concatenated without new line with anything else
256
		 */
257
		$regexp = /** @lang PhpRegExp */
258 10
			'[:;,.+\-*/{}?><^\'"\[\]=&(]';
259
		return
260 10
			$current_line &&
261 10
			$next_line &&
262 10
			!preg_match("#$regexp\$#", $current_line) &&
263 10
			!preg_match("#^$regexp#", $next_line);
264
	}
265
	/**
266
	 * Analyses file for scripts and styles, combines them into resulting files in order to optimize loading process
267
	 * (files with combined scripts and styles will be created)
268
	 *
269
	 * @param string   $data                   Content of processed file
270
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
271
	 * @param string   $target_directory_path  Target directory for resulting combined files
272
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
273
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
274
	 *
275
	 * @return string
276
	 */
277 8
	public static function html ($data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources = []) {
278 8
		static::html_process_links_and_styles($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
279 8
		static::html_process_scripts($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
280
		// Removing HTML comments (those that are mostly likely comments, to avoid problems)
281 8
		$data = preg_replace_callback(
282 8
			'/^\s*<!--([^>-].*[^-])?-->/Ums',
283 8
			function ($matches) {
284 8
				return mb_strpos('--', $matches[1]) === false ? '' : $matches[0];
285 8
			},
286
			$data
287
		);
288 8
		return preg_replace("/\n+/", "\n", $data);
289
	}
290
	/**
291
	 * @param string   $data                   Content of processed file
292
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
293
	 * @param string   $target_directory_path  Target directory for resulting combined files
294
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
295
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
296
	 */
297 8
	protected static function html_process_scripts (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
298 8
		if (!preg_match_all('/<script(.*)<\/script>/Uims', $data, $scripts)) {
299 8
			return;
300
		}
301 8
		$scripts_content    = '';
302 8
		$scripts_to_replace = [];
303 8
		$dir                = dirname($file);
304 8
		foreach ($scripts[1] as $index => $script) {
305 8
			$script = explode('>', $script, 2);
306 8
			if (preg_match('/src\s*=\s*[\'"](.*)[\'"]/Uims', $script[0], $url)) {
307 8
				$url = $url[1];
308 8
				if (!static::is_relative_path_and_exists($url, $dir)) {
309 4
					continue;
310
				}
311 8
				$scripts_to_replace[] = $scripts[0][$index];
312 8
				$scripts_content .= file_get_contents("$dir/$url").";\n";
313
			} else {
314 8
				$scripts_to_replace[] = $scripts[0][$index];
315 8
				$scripts_content .= "$script[1];\n";
316
			}
317
		}
318 8
		$scripts_content = static::js($scripts_content);
319 8
		if (!$scripts_to_replace) {
320 4
			return;
321
		}
322
		// Remove all scripts
323 8
		$data = str_replace($scripts_to_replace, '', $data);
324
		/**
325
		 * If vulcanization is not used - put contents into separate file, and put link to it, otherwise put minified content back
326
		 */
327 8
		if (!$vulcanization) {
328
			// TODO: Remove in 7.x; For backward compatibility, since some modules might use this by specifying file path
329 2
			if (!is_dir($target_directory_path)) {
330
				$target_directory_path = dirname($target_directory_path);
331
			}
332 2
			$filename = static::file_put_contents_with_hash($target_directory_path, 'js', $scripts_content);
333
			// Add script with combined content file to the end
334 2
			$data .= "<script src=\"./$filename\"></script>";
335 2
			$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
336
		} else {
337
			// Add combined content inline script to the end
338 6
			$data .= "<script>$scripts_content</script>";
339
		}
340 8
	}
341
	/**
342
	 * @param string   $data                   Content of processed file
343
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
344
	 * @param string   $target_directory_path  Target directory for resulting combined files
345
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
346
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
347
	 */
348 8
	protected static function html_process_links_and_styles (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
349
		// Drop Polymer inclusion, since it is already present
350 8
		$data = str_replace('<link rel="import" href="../polymer/polymer.html">', '', $data);
351 8
		if (!preg_match_all('/<link(.*)>|<style(.*)<\/style>/Uims', $data, $links_and_styles)) {
352 8
			return;
353
		}
354 8
		$dir = dirname($file);
355 8
		foreach ($links_and_styles[1] as $index => $link) {
356
			/**
357
			 * Check for custom styles `is="custom-style"` or styles includes `include=".."` - we'll skip them
358
			 * Or if content is plain CSS
359
			 */
360
			if (
361 8
				preg_match('/^[^>]*(is="custom-style"|include=)[^>]*>/Uim', $links_and_styles[2][$index]) ||
362 8
				mb_strpos($links_and_styles[0][$index], '</style>') > 0
363
			) {
364 8
				$content = explode('>', $links_and_styles[2][$index], 2)[1];
365 8
				$data    = str_replace(
366
					$content,
367 8
					static::css($content, $file, $target_directory_path, $not_embedded_resources),
368
					$data
369
				);
370 8
				continue;
371
			}
372 8
			if (!static::has_relative_href($link, $url, $dir)) {
373 4
				continue;
374
			}
375 8
			$import = preg_match('/rel\s*=\s*[\'"]import[\'"]/Uim', $link);
376
			/**
377
			 * CSS imports are available in Polymer alongside with HTML imports
378
			 */
379 8
			$css_import = $import && preg_match('/type\s*=\s*[\'"]css[\'"]/Uim', $link);
380 8
			$stylesheet = preg_match('/rel\s*=\s*[\'"]stylesheet[\'"]/Uim', $link);
381
			// TODO: Polymer only supports `style[is=custom-style]`, but no `link`-based counterpart, so we can't provide CSP-compatibility for CSS anyway
382 8
			if ($css_import || $stylesheet) {
383
				/**
384
				 * If content is link to CSS file
385
				 */
386 8
				$css  = static::css(
387 8
					file_get_contents("$dir/$url"),
388 8
					"$dir/$url",
389
					$target_directory_path,
390
					$not_embedded_resources
391
				);
392 8
				$data = preg_replace(
393 8
					'/'.$links_and_styles[0][$index].'.*<template>/Uims',
394 8
					"<template><style>$css</style>",
395
					$data
396
				);
397 4
			} elseif ($import) {
398
				/**
399
				 * If content is HTML import
400
				 */
401 4
				$data = str_replace(
402 4
					$links_and_styles[0][$index],
403 4
					static::html(
404 4
						file_get_contents("$dir/$url"),
405 8
						"$dir/$url",
406
						$target_directory_path,
407
						$vulcanization,
408
						$not_embedded_resources
409
					),
410
					$data
411
				);
412
			}
413
		}
414 8
	}
415
	/**
416
	 * @param string $link
417
	 * @param string $url
418
	 * @param string $dir
419
	 *
420
	 * @return bool
421
	 */
422 8
	protected static function has_relative_href ($link, &$url, $dir) {
423
		$result =
424 8
			$link &&
425 8
			preg_match('/href\s*=\s*[\'"](.*)[\'"]/Uims', $link, $url);
426 8
		if ($result && static::is_relative_path_and_exists($url[1], $dir)) {
427 8
			$url = $url[1];
428 8
			return true;
429
		}
430 4
		return false;
431
	}
432
	/**
433
	 * @param string $path
434
	 * @param string $dir
435
	 *
436
	 * @return bool
437
	 */
438 10
	protected static function is_relative_path_and_exists ($path, $dir) {
439 10
		return $dir && !preg_match('#^https?://#i', $path) && file_exists(static::absolute_path($path, $dir));
440
	}
441
	/**
442
	 * @param string $path
443
	 * @param string $dir
444
	 *
445
	 * @return string
446
	 */
447 10
	protected static function absolute_path ($path, $dir) {
448 10
		if (strpos($path, '/') === 0) {
449 6
			return realpath(getcwd().$path);
450
		}
451 10
		return realpath("$dir/$path");
452
	}
453
}
454