Completed
Push — master ( 1ab7e1...3c0bb6 )
by Nazar
04:30
created

Assets_processing::absolute_path()   A

Complexity

Conditions 2
Paths 2

Size

Total Lines 6
Code Lines 4

Duplication

Lines 0
Ratio 0 %

Code Coverage

Tests 4
CRAP Score 2

Importance

Changes 0
Metric Value
cc 2
eloc 4
nc 2
nop 2
dl 0
loc 6
ccs 4
cts 4
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
					$filename = md5_file($absolute_path).'.'.$extension;
130 10
					copy($absolute_path, "$target_directory_path/$filename");
131 10
					if (strpos($path, '?') === false) {
132 6
						$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
133
					}
134 10
					return str_replace($path_matched, "'./$filename'", $match[0]);
135
				}
136 6
				$mime_type = static::$extension_to_mime[$extension];
137 6
				$content   = base64_encode($content);
138 6
				return str_replace($path, "data:$mime_type;charset=utf-8;base64,$content", $match[0]);
139 10
			},
140
			$data
141
		);
142 10
		return trim($data);
143
	}
144
	/**
145
	 * Put `$content` into `$dir` where filename is `md5($content)` with specified extension
146
	 *
147
	 * @param string $dir
148
	 * @param string $extension
149
	 * @param string $content
150
	 *
151
	 * @return string Filename (without full path)
152
	 */
153 6
	protected static function file_put_contents_with_hash ($dir, $extension, $content) {
154 6
		$hash = md5($content);
155 6
		file_put_contents("$dir/$hash.$extension", $content, LOCK_EX | FILE_BINARY);
156 6
		return "$hash.$extension";
157
	}
158
	/**
159
	 * Simple and fast JS minification
160
	 *
161
	 * @param string $data
162
	 *
163
	 * @return string
164
	 */
165 10
	public static function js ($data) {
166
		/**
167
		 * Split into array of lines
168
		 */
169 10
		$data = explode("\n", $data);
170
		/**
171
		 * Flag that is `true` when inside comment
172
		 */
173 10
		$in_comment              = false;
174 10
		$continue_after_position = -1;
175 10
		foreach ($data as $index => &$current_line) {
176 10
			if ($continue_after_position >= $index) {
177 6
				continue;
178
			}
179 10
			$next_line = isset($data[$index + 1]) ? trim($data[$index + 1]) : '';
180
			/**
181
			 * Remove starting and trailing spaces
182
			 */
183 10
			$current_line = trim($current_line);
184
			/**
185
			 * Remove single-line comments
186
			 */
187 10
			if (mb_strpos($current_line, '//') === 0) {
188 10
				$current_line = '';
189 10
				continue;
190
			}
191
			/**
192
			 * Starts with multi-line comment
193
			 */
194 10
			if (mb_strpos($current_line, '/*') === 0) {
195 10
				$in_comment = true;
196
			}
197 10
			if (!$in_comment) {
198 10
				$backticks_position = strpos($current_line, '`');
199
				/**
200
				 * Handling template strings can be tricky (since they might be multi-line), so let's fast-forward to the last backticks position and continue
201
				 * from there
202
				 */
203 10
				if ($backticks_position !== false) {
204 6
					$last_item_with_backticks = array_keys(
205
						array_filter(
206
							$data,
207
							function ($d) {
208 6
								return strpos($d, '`') !== false;
209 6
							}
210
						)
211
					);
212 6
					$last_item_with_backticks = array_pop($last_item_with_backticks);
213 6
					if ($last_item_with_backticks > $index) {
214 6
						$continue_after_position = $last_item_with_backticks;
215 6
						continue;
216
					}
217
				}
218
				/**
219
				 * Add new line at the end if only needed
220
				 */
221 10
				if (static::new_line_needed($current_line, $next_line)) {
222 10
					$current_line .= "\n";
223
				}
224
				/**
225
				 * Single-line comment
226
				 */
227 10
				$current_line = preg_replace('#^\s*//[^\'"]+$#', '', $current_line);
228
				/**
229
				 * If we are not sure - just add new line afterwards
230
				 */
231 10
				$current_line = preg_replace('#//.*$#', "\\0\n", $current_line);
232
			} else {
233
				/**
234
				 * End of multi-line comment
235
				 */
236 10
				if (strpos($current_line, '*/') !== false) {
237 10
					$current_line = explode('*/', $current_line)[1];
238 10
					$in_comment   = false;
239
				} else {
240 10
					$current_line = '';
241
				}
242
			}
243
		}
244 10
		$data = implode('', $data);
245 10
		$data = str_replace('</script>', '<\/script>', $data);
246 10
		return trim($data, ';').';';
247
	}
248
	/**
249
	 * @param string $current_line
250
	 * @param string $next_line
251
	 *
252
	 * @return bool
253
	 */
254 10
	protected static function new_line_needed ($current_line, $next_line) {
255
		/**
256
		 * Set of symbols that are safe to be concatenated without new line with anything else
257
		 */
258
		$regexp = /** @lang PhpRegExp */
259 10
			'[:;,.+\-*/{}?><^\'"\[\]=&(]';
260
		return
261 10
			$current_line &&
262 10
			$next_line &&
263 10
			!preg_match("#$regexp\$#", $current_line) &&
264 10
			!preg_match("#^$regexp#", $next_line);
265
	}
266
	/**
267
	 * Analyses file for scripts and styles, combines them into resulting files in order to optimize loading process
268
	 * (files with combined scripts and styles will be created)
269
	 *
270
	 * @param string   $data                   Content of processed file
271
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
272
	 * @param string   $target_directory_path  Target directory for resulting combined files
273
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
274
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
275
	 *
276
	 * @return string
277
	 */
278 8
	public static function html ($data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources = []) {
279 8
		static::html_process_links_and_styles($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
280 8
		static::html_process_scripts($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
281
		// Removing HTML comments (those that are mostly likely comments, to avoid problems)
282 8
		$data = preg_replace_callback(
283 8
			'/^\s*<!--([^>-].*[^-])?-->/Ums',
284 8
			function ($matches) {
285 8
				return mb_strpos('--', $matches[1]) === false ? '' : $matches[0];
286 8
			},
287
			$data
288
		);
289 8
		return preg_replace("/\n+/", "\n", $data);
290
	}
291
	/**
292
	 * @param string   $data                   Content of processed file
293
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
294
	 * @param string   $target_directory_path  Target directory for resulting combined files
295
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
296
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
297
	 */
298 8
	protected static function html_process_scripts (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
299 8
		if (!preg_match_all('/<script(.*)<\/script>/Uims', $data, $scripts)) {
300 8
			return;
301
		}
302 8
		$scripts_content    = '';
303 8
		$scripts_to_replace = [];
304 8
		$dir                = dirname($file);
305 8
		foreach ($scripts[1] as $index => $script) {
306 8
			$script = explode('>', $script, 2);
307 8
			if (preg_match('/src\s*=\s*[\'"](.*)[\'"]/Uims', $script[0], $url)) {
308 8
				$url = $url[1];
309 8
				if (!static::is_relative_path_and_exists($url, $dir)) {
310 4
					continue;
311
				}
312 8
				$scripts_to_replace[] = $scripts[0][$index];
313 8
				$scripts_content .= file_get_contents("$dir/$url").";\n";
314
			} else {
315 8
				$scripts_to_replace[] = $scripts[0][$index];
316 8
				$scripts_content .= "$script[1];\n";
317
			}
318
		}
319 8
		$scripts_content = static::js($scripts_content);
320 8
		if (!$scripts_to_replace) {
321 4
			return;
322
		}
323
		// Remove all scripts
324 8
		$data = str_replace($scripts_to_replace, '', $data);
325
		/**
326
		 * If vulcanization is not used - put contents into separate file, and put link to it, otherwise put minified content back
327
		 */
328 8
		if (!$vulcanization) {
329
			// TODO: Remove in 7.x; For backward compatibility, since some modules might use this by specifying file path
330 2
			if (!is_dir($target_directory_path)) {
331
				$target_directory_path = dirname($target_directory_path);
332
			}
333 2
			$filename = static::file_put_contents_with_hash($target_directory_path, 'js', $scripts_content);
334
			// Add script with combined content file to the end
335 2
			$data .= "<script src=\"./$filename\"></script>";
336 2
			$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
337
		} else {
338
			// Add combined content inline script to the end
339 6
			$data .= "<script>$scripts_content</script>";
340
		}
341 8
	}
342
	/**
343
	 * @param string   $data                   Content of processed file
344
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
345
	 * @param string   $target_directory_path  Target directory for resulting combined files
346
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
347
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
348
	 */
349 8
	protected static function html_process_links_and_styles (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
350
		// Drop Polymer inclusion, since it is already present
351 8
		$data = str_replace('<link rel="import" href="../polymer/polymer.html">', '', $data);
352 8
		if (!preg_match_all('/<link(.*)>|<style(.*)<\/style>/Uims', $data, $links_and_styles)) {
353 8
			return;
354
		}
355 8
		$dir = dirname($file);
356 8
		foreach ($links_and_styles[1] as $index => $link) {
357
			/**
358
			 * Check for custom styles `is="custom-style"` or styles includes `include=".."` - we'll skip them
359
			 * Or if content is plain CSS
360
			 */
361
			if (
362 8
				preg_match('/^[^>]*(is="custom-style"|include=)[^>]*>/Uim', $links_and_styles[2][$index]) ||
363 8
				mb_strpos($links_and_styles[0][$index], '</style>') > 0
364
			) {
365 8
				$content = explode('>', $links_and_styles[2][$index], 2)[1];
366 8
				$data    = str_replace(
367
					$content,
368 8
					static::css($content, $file, $target_directory_path, $not_embedded_resources),
369
					$data
370
				);
371 8
				continue;
372
			}
373 8
			if (!static::has_relative_href($link, $url, $dir)) {
374 4
				continue;
375
			}
376 8
			$import = preg_match('/rel\s*=\s*[\'"]import[\'"]/Uim', $link);
377
			/**
378
			 * CSS imports are available in Polymer alongside with HTML imports
379
			 */
380 8
			$css_import = $import && preg_match('/type\s*=\s*[\'"]css[\'"]/Uim', $link);
381 8
			$stylesheet = preg_match('/rel\s*=\s*[\'"]stylesheet[\'"]/Uim', $link);
382
			// TODO: Polymer only supports `style[is=custom-style]`, but no `link`-based counterpart, so we can't provide CSP-compatibility for CSS anyway
383 8
			if ($css_import || $stylesheet) {
384
				/**
385
				 * If content is link to CSS file
386
				 */
387 8
				$css  = static::css(
388 8
					file_get_contents("$dir/$url"),
389 8
					"$dir/$url",
390
					$target_directory_path,
391
					$not_embedded_resources
392
				);
393 8
				$data = preg_replace(
394 8
					'/'.$links_and_styles[0][$index].'.*<template>/Uims',
395 8
					"<template><style>$css</style>",
396
					$data
397
				);
398 4
			} elseif ($import) {
399
				/**
400
				 * If content is HTML import
401
				 */
402 4
				$data = str_replace(
403 4
					$links_and_styles[0][$index],
404 4
					static::html(
405 4
						file_get_contents("$dir/$url"),
406 8
						"$dir/$url",
407
						$target_directory_path,
408
						$vulcanization,
409
						$not_embedded_resources
410
					),
411
					$data
412
				);
413
			}
414
		}
415 8
	}
416
	/**
417
	 * @param string $link
418
	 * @param string $url
419
	 * @param string $dir
420
	 *
421
	 * @return bool
422
	 */
423 8
	protected static function has_relative_href ($link, &$url, $dir) {
424
		$result =
425 8
			$link &&
426 8
			preg_match('/href\s*=\s*[\'"](.*)[\'"]/Uims', $link, $url);
427 8
		if ($result && static::is_relative_path_and_exists($url[1], $dir)) {
428 8
			$url = $url[1];
429 8
			return true;
430
		}
431 4
		return false;
432
	}
433
	/**
434
	 * @param string $path
435
	 * @param string $dir
436
	 *
437
	 * @return bool
438
	 */
439 10
	protected static function is_relative_path_and_exists ($path, $dir) {
440 10
		return $dir && !preg_match('#^https?://#i', $path) && file_exists(static::absolute_path($path, $dir));
441
	}
442
	/**
443
	 * @param string $path
444
	 * @param string $dir
445
	 *
446
	 * @return string
447
	 */
448 10
	protected static function absolute_path ($path, $dir) {
449 10
		if (strpos($path, '/') === 0) {
450 6
			return realpath(getcwd().$path);
451
		}
452 10
		return realpath("$dir/$path");
453
	}
454
}
455