Completed
Push — master ( bf102e...fafcea )
by Nazar
06:46
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) 2014-2017, 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, nested files are also copied
15
 * to target directory and processed if needed.
16
 */
17
class Assets_processing {
18
	/**
19
	 * Analyses file for images, fonts and css links and include they content into single resulting css file.
20
	 *
21
	 * Supports next file extensions for possible assets:
22
	 * jpeg, jpe, jpg, gif, png, ttf, ttc, svg, svgz, woff, css
23
	 *
24
	 * @param string   $data                   Content of processed file
25
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
26
	 * @param string   $target_directory_path  Target directory for resulting combined files
27
	 * @param string[] $not_embedded_resources Some resources like images and fonts might not be embedded into resulting CSS because of their size
28
	 *
29
	 * @return string    $data
30
	 */
31 10
	public static function css ($data, $file, $target_directory_path = PUBLIC_CACHE, &$not_embedded_resources = []) {
32 10
		$dir = dirname($file);
33
		/**
34
		 * Remove comments, tabs and new lines
35
		 */
36 10
		$data = preg_replace('#(/\*.*?\*/)|\t|\n|\r#s', ' ', $data);
37
		/**
38
		 * Remove unnecessary spaces
39
		 */
40 10
		$data = preg_replace('/\s*([,;>{}(])\s*/', '$1', $data);
41 10
		$data = preg_replace('/\s+/', ' ', $data);
42
		/**
43
		 * Return spaces required in media queries
44
		 */
45 10
		$data = preg_replace('/\s(and|or)\(/', ' $1 (', $data);
46
		/**
47
		 * Duplicated semicolons
48
		 */
49 10
		$data = preg_replace('/;+/m', ';', $data);
50
		/**
51
		 * Minify rgb colors declarations
52
		 */
53 10
		$data = preg_replace_callback(
54 10
			'/rgb\(([0-9,.]+)\)/i',
55
			function ($rgb) {
56 6
				$rgb = explode(',', $rgb[1]);
57
				return
58
					'#'.
59 6
					str_pad(dechex($rgb[0]), 2, 0, STR_PAD_LEFT).
60 6
					str_pad(dechex($rgb[1]), 2, 0, STR_PAD_LEFT).
61 6
					str_pad(dechex($rgb[2]), 2, 0, STR_PAD_LEFT);
62 10
			},
63 10
			$data
64
		);
65
		/**
66
		 * Minify repeated colors declarations
67
		 */
68 10
		$data = preg_replace('/#([0-9a-f])\1([0-9a-f])\2([0-9a-f])\3/i', '#$1$2$3', $data);
69
		/**
70
		 * Remove unnecessary zeros
71
		 */
72 10
		$data = preg_replace('/(\D)0\.(\d+)/i', '$1.$2', $data);
73
		/**
74
		 * Unnecessary spaces around colons (should have whitespace character after, otherwise `.class :disabled` will be handled incorrectly)
75
		 */
76 10
		$data = preg_replace('/\s*:\s+/', ':', $data);
77
		/**
78
		 * Assets processing
79
		 */
80
		// TODO: replace by loop, track duplicated stuff that are subject to inlining and if they appear more than once, don't inline them
81 10
		$data = preg_replace_callback(
82 10
			'/url\((.*)\)|@import\s*(?:url\()?\s*([\'"].*[\'"])\s*\)??.*;/U',
83
			function ($match) use ($dir, $target_directory_path, &$not_embedded_resources) {
84 10
				$path_matched = $match[2] ?? $match[1];
85 10
				$path         = trim($path_matched, '\'" ');
86 10
				$link         = explode('?', $path, 2)[0];
87 10
				if (!static::is_relative_path_and_exists($link, $dir)) {
88 4
					return $match[0];
89
				}
90 10
				$extension     = file_extension($link);
91 10
				$absolute_path = static::absolute_path($link, $dir);
92 10
				$content       = file_get_contents($absolute_path);
93
				/**
94
				 * Process recursively CSS imports, but ignore non embedded resources there
95
				 */
96
				/** @noinspection UsageOfSilenceOperatorInspection */
97 10
				if ($extension == 'css' && @$match[2]) {
98 6
					$content = static::css($content, $absolute_path, $target_directory_path);
99
				}
100 10
				$filename = static::file_put_contents_with_hash($target_directory_path, $extension, $content);
101
				/** @noinspection UsageOfSilenceOperatorInspection */
102
				/**
103
				 * We ignore files with query parameters as well as CSS imports with media queries
104
				 */
105 10
				if (strpos($path, '?') === false && !trim(@$match[3])) {
106 6
					$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
107
				}
108 10
				return str_replace($path_matched, "'./$filename'", $match[0]);
109 10
			},
110 10
			$data
111
		);
112 10
		return trim($data);
113
	}
114
	/**
115
	 * Put `$content` into `$dir` where filename is `md5($content)` with specified extension
116
	 *
117
	 * @param string $dir
118
	 * @param string $extension
119
	 * @param string $content
120
	 *
121
	 * @return string Filename (without full path)
122
	 */
123 10
	protected static function file_put_contents_with_hash ($dir, $extension, $content) {
124 10
		$hash = md5($content);
125 10
		file_put_contents("$dir/$hash.$extension", $content, LOCK_EX | FILE_BINARY);
126 10
		return "$hash.$extension";
127
	}
128
	/**
129
	 * Simple and fast JS minification
130
	 *
131
	 * @param string $data
132
	 *
133
	 * @return string
134
	 */
135 10
	public static function js ($data) {
136
		/**
137
		 * Split into array of lines
138
		 */
139 10
		$data = explode("\n", $data);
140
		/**
141
		 * Flag that is `true` when inside comment
142
		 */
143 10
		$in_comment              = false;
144 10
		$continue_after_position = -1;
145 10
		foreach ($data as $index => &$current_line) {
146 10
			if ($continue_after_position >= $index) {
147 2
				continue;
148
			}
149 10
			$next_line = isset($data[$index + 1]) ? trim($data[$index + 1]) : '';
150
			/**
151
			 * Remove starting and trailing spaces
152
			 */
153 10
			$current_line = trim($current_line);
154
			/**
155
			 * Remove single-line comments
156
			 */
157 10
			if (mb_strpos($current_line, '//') === 0) {
158 10
				$current_line = '';
159 10
				continue;
160
			}
161
			/**
162
			 * Starts with multi-line comment
163
			 */
164 10
			if (mb_strpos($current_line, '/*') === 0) {
165 10
				$in_comment = true;
166
			}
167 10
			if (!$in_comment) {
168 10
				$backticks_position = strpos($current_line, '`');
169
				/**
170
				 * Handling template strings can be tricky (since they might be multi-line), so let's fast-forward to the last backticks position and continue
171
				 * from there
172
				 */
173 10
				if ($backticks_position !== false) {
174 6
					$last_item_with_backticks = array_keys(
175 6
						array_filter(
176 6
							$data,
177
							function ($d) {
178 6
								return strpos($d, '`') !== false;
179 6
							}
180
						)
181
					);
182 6
					$last_item_with_backticks = array_pop($last_item_with_backticks);
183 6
					if ($last_item_with_backticks > $index) {
184 2
						$continue_after_position = $last_item_with_backticks;
185 2
						continue;
186
					}
187
				}
188
				/**
189
				 * Add new line at the end if only needed
190
				 */
191 10
				if (static::new_line_needed($current_line, $next_line)) {
192 10
					$current_line .= "\n";
193
				}
194
				/**
195
				 * Single-line comment
196
				 */
197 10
				$current_line = preg_replace('#^\s*//[^\'"]+$#', '', $current_line);
198
				/**
199
				 * If we are not sure - just add new line afterwards
200
				 */
201 10
				$current_line = preg_replace('#//.*$#', "\\0\n", $current_line);
202
			} else {
203
				/**
204
				 * End of multi-line comment
205
				 */
206 10
				if (strpos($current_line, '*/') !== false) {
207 10
					$current_line = explode('*/', $current_line)[1];
208 10
					$in_comment   = false;
209
				} else {
210 10
					$current_line = '';
211
				}
212
			}
213
		}
214 10
		$data = implode('', $data);
215 10
		$data = str_replace('</script>', '<\/script>', $data);
216 10
		return trim($data, ';').';';
217
	}
218
	/**
219
	 * @param string $current_line
220
	 * @param string $next_line
221
	 *
222
	 * @return bool
223
	 */
224 10
	protected static function new_line_needed ($current_line, $next_line) {
225
		/**
226
		 * Set of symbols that are safe to be concatenated without new line with anything else
227
		 */
228
		$regexp = /** @lang PhpRegExp */
229 10
			'[:;,.+\-*/{}?><^\'"\[\]=&(]';
230
		return
231 10
			$current_line &&
232 10
			$next_line &&
233 10
			!preg_match("#$regexp\$#", $current_line) &&
234 10
			!preg_match("#^$regexp#", $next_line);
235
	}
236
	/**
237
	 * Analyses file for scripts and styles, combines them into resulting files in order to optimize loading process
238
	 * (files with combined scripts and styles will be created)
239
	 *
240
	 * @param string   $data                   Content of processed file
241
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
242
	 * @param string   $target_directory_path  Target directory for resulting combined files
243
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
244
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
245
	 *
246
	 * @return string
247
	 */
248 8
	public static function html ($data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources = []) {
249 8
		static::html_process_links_and_styles($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
250 8
		static::html_process_scripts($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
251
		// Removing HTML comments (those that are mostly likely comments, to avoid problems)
252 8
		$data = preg_replace_callback(
253 8
			'/^\s*<!--([^>-].*[^-])?-->/Ums',
254
			function ($matches) {
255 8
				return mb_strpos('--', $matches[1]) === false ? '' : $matches[0];
256 8
			},
257 8
			$data
258
		);
259 8
		return preg_replace("/\n+/", "\n", $data);
260
	}
261
	/**
262
	 * @param string   $data                   Content of processed file
263
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
264
	 * @param string   $target_directory_path  Target directory for resulting combined files
265
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
266
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
267
	 */
268 8
	protected static function html_process_scripts (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
269 8
		if (!preg_match_all('/<script(.*)<\/script>/Uims', $data, $scripts)) {
270 8
			return;
271
		}
272 8
		$scripts_content    = '';
273 8
		$scripts_to_replace = [];
274 8
		$dir                = dirname($file);
275 8
		foreach ($scripts[1] as $index => $script) {
276 8
			$script = explode('>', $script, 2);
277 8
			if (preg_match('/src\s*=\s*[\'"](.*)[\'"]/Uims', $script[0], $url)) {
278 8
				$url = $url[1];
279 8
				if (!static::is_relative_path_and_exists($url, $dir)) {
280 4
					continue;
281
				}
282 8
				$scripts_to_replace[] = $scripts[0][$index];
283 8
				$scripts_content      .= file_get_contents("$dir/$url").";\n";
284
			} else {
285 8
				$scripts_to_replace[] = $scripts[0][$index];
286 8
				$scripts_content      .= "$script[1];\n";
287
			}
288
		}
289 8
		$scripts_content = static::js($scripts_content);
290 8
		if (!$scripts_to_replace) {
291 4
			return;
292
		}
293
		// Remove all scripts
294 8
		$data = str_replace($scripts_to_replace, '', $data);
295
		/**
296
		 * If vulcanization is not used - put contents into separate file, and put link to it, otherwise put minified content back
297
		 */
298 8
		if (!$vulcanization) {
299 2
			$filename = static::file_put_contents_with_hash($target_directory_path, 'js', $scripts_content);
300
			// Add script with combined content file to the end
301 2
			$data                     .= "<script src=\"./$filename\"></script>";
302 2
			$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
303
		} else {
304
			// Add combined content inline script to the end
305 6
			$data .= "<script>$scripts_content</script>";
306
		}
307 8
	}
308
	/**
309
	 * @param string   $data                   Content of processed file
310
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
311
	 * @param string   $target_directory_path  Target directory for resulting combined files
312
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
313
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
314
	 */
315 8
	protected static function html_process_links_and_styles (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
316 8
		if (!preg_match_all('/<link(.*)>|<style(.*)<\/style>/Uims', $data, $links_and_styles)) {
317 8
			return;
318
		}
319 8
		$dir = dirname($file);
320
		// With reverse order we can add inlined CSS imports directly to the beginning of `<template>` element
321 8
		$links_and_styles = array_map(
322 8
			function ($item) {
323 8
				return array_reverse($item, true);
324 8
			},
325 8
			$links_and_styles
326
		);
327 8
		foreach ($links_and_styles[1] as $index => $link) {
328
			/**
329
			 * For plain styles we do not do anything fancy besides minifying its sources (no rearrangement or anything like that)
330
			 */
331 8
			if (mb_strpos($links_and_styles[0][$index], '</style>') > 0) {
332 8
				$content = explode('>', $links_and_styles[2][$index], 2)[1];
333 8
				$data    = str_replace(
334 8
					$content,
335 8
					static::css($content, $file, $target_directory_path, $not_embedded_resources),
336 8
					$data
337
				);
338 8
				continue;
339
			}
340 8
			if (!static::has_relative_href($link, $url, $dir)) {
341 4
				continue;
342
			}
343 8
			$import = preg_match('/rel\s*=\s*[\'"]import[\'"]/Uim', $link);
344
			/**
345
			 * CSS imports are available in Polymer alongside with HTML imports
346
			 */
347 8
			$css_import = $import && preg_match('/type\s*=\s*[\'"]css[\'"]/Uim', $link);
348 8
			$stylesheet = preg_match('/rel\s*=\s*[\'"]stylesheet[\'"]/Uim', $link);
349
			/**
350
			 * TODO: Polymer only supports `custom-style > style`, but no `link`-based counterpart, so we can't provide CSP-compatibility in general,
351
			 * thus always inlining styles into HTML
352
			 */
353 8
			if ($css_import || $stylesheet) {
354
				/**
355
				 * If content is link to CSS file
356
				 */
357 8
				$css  = static::css(
358 8
					file_get_contents("$dir/$url"),
359 8
					"$dir/$url",
360 8
					$target_directory_path,
361 8
					$not_embedded_resources
362
				);
363 8
				$data = preg_replace(
364 8
					'/'.preg_quote($links_and_styles[0][$index], '/').'(.*)<template>/Uims',
365 8
					"$1<template><style>$css</style>",
366 8
					$data
367
				);
368 8
				$data = preg_replace(
369 8
					'/(<template><style>[^<]+)<\/style><style>/Uim',
370 8
					'$1',
371 8
					$data
372
				);
373 4
			} elseif ($import) {
374
				/**
375
				 * If content is HTML import
376
				 */
377 4
				$data = str_replace(
378 4
					$links_and_styles[0][$index],
379 4
					static::html(
380 4
						file_get_contents("$dir/$url"),
381 4
						"$dir/$url",
382 4
						$target_directory_path,
383 4
						$vulcanization,
384 4
						$not_embedded_resources
385
					),
386 8
					$data
387
				);
388
			}
389
		}
390 8
	}
391
	/**
392
	 * @param string $link
393
	 * @param string $url
394
	 * @param string $dir
395
	 *
396
	 * @return bool
397
	 */
398 8
	protected static function has_relative_href ($link, &$url, $dir) {
399
		$result =
400 8
			$link &&
401 8
			preg_match('/href\s*=\s*[\'"](.*)[\'"]/Uims', $link, $url);
402 8
		if ($result && static::is_relative_path_and_exists($url[1], $dir)) {
403 8
			$url = $url[1];
404 8
			return true;
405
		}
406 4
		return false;
407
	}
408
	/**
409
	 * @param string $path
410
	 * @param string $dir
411
	 *
412
	 * @return bool
413
	 */
414 10
	protected static function is_relative_path_and_exists ($path, $dir) {
415 10
		return $dir && !preg_match('#^https?://#i', $path) && file_exists(static::absolute_path($path, $dir));
416
	}
417
	/**
418
	 * @param string $path
419
	 * @param string $dir
420
	 *
421
	 * @return string
422
	 */
423 10
	protected static function absolute_path ($path, $dir) {
424 10
		if (strpos($path, '/') === 0) {
425 6
			return realpath(getcwd().$path);
426
		}
427 10
		return realpath("$dir/$path");
428
	}
429
}
430